COLLABORATIVE PROJECT DELIVERY PRACTICES, GOAL ALIGNMENT, AND PERFORMANCE IN ARCHITECTURE, ENGINEERING, AND CONSTRUCTION PROJECT TEAMS By Anthony E lijah Sparkling A DISSERTATION Submitted to Michigan State University in partial fulfi llment of the requirements f or the degree of Planning, Design, and Construction Doctor of Philosophy 2018 ABSTRACT COLLABORATIVE DELIVERY PRACTICES, GOAL ALIGNMENT, AND PERFORMANCE IN ARCHITECTURE, ENGINEERING, AND CONSTRUCTION PROJECT T EAMS By Anthony E lijah Sparkling The Architecture, Engineering, and Construction (AEC) industry is increasingly challenged with improving the efficacy of project team performance through collaborative working arrangements. Collaborative working arrangem ents such as integrated project delivery, design - build, and project partnering are all comprised of interorganizational project teams. These teams, according to relational governance theory, generally function with flexibility, solidarity, mutual respect, and openly share information. Recent research shows that collaborative and cohesive teams are perpetuated by strategies to facilitate team integration methods. Efficient knowledge sharing and processing systems, also called transactive memory systems (TMS) , are integral to cohesive project teams and their tasks coordination. Although the AEC literature is widespread on the importance of team integration and cohesion, little emphasis is placed on the effects of goal alignment practices and its relationship t o performance outcomes. Thus, this research aims to explore this relationship along with the moderating effects of TMS in the context of partnered - projects by investigating interorganizational AEC project teams. Some goal alignment characteristics of part nered - projects are generally in the form of partnering workshops, establishing clear goals and objectives, and the early involvement of key stakeholders (e.g., owner, designer, contractor, subcontracts). The link between partnering practices and project su ccess dominates AEC literature, yet the elements of partnering practice should be examined separately. This research asserts collaborative project delivery practices affect goal alignment and performance in AEC project teams. Furthermore, this research poi nts out how behavioral attributes (i.e., transactive memory systems) of partnered - project teams are important to successful project delivery on AEC projects. To achieve the aim of this study, data were collected from six case study projects and 125 partic ipants using web - based surveys. P roject information was accessed via partnering documents collected from key project stakeholders. A mixed methods approach was followed where 1) Qualitative data w ere analyzed using content analysis and case study tactics i ncluding pattern - matching and cross - case synthesis ; and, 2) Quantitative data w ere analyzed using confirmatory factor analysis and multivariate regression analysis. The theoretical contribution resulting from this research help explain the variation in in terorganizational project team performance by examining key behavioral attributes emanating from organizational theory. Researchers have alluded to cognitive behaviors and social norms as potential moderators between performance outcomes and collaborative project delivery approaches. This study takes a step further positing clear metrics to understand goal alignment and team dynamics via transactive memory systems. It espouses relational governance theory as explanatory for the unique dynamics underpinning team integration. Findings show maintaining g oal alignment becomes problematic as the number of performance measures increases or when competing messages are sent. This is exacerbated when performance measures such disincentives are codified in contracts without subsequent incentive or rewards. Other results show transactive memory systems has a positive effect on individual performance. Copyright by ANTHONY ELIJAH SPARKLING 2018 v ACKNOWLEDGMENTS The author would like to acknowledge the multitude of individuals that have played a key role in bringing this dissertation to fruition. The journey has been very enjoyable despite all the challenges and sacrifices required for a monumental task as this. Thus, I would be remiss if homage was not given to those faculty, friends, and family who helped support this effort to which I am forever grateful. Dr. Sinem Mollaoglu, my success is a testament to all the guidance, insights, inspiration, and trust you have given me. With your wisdom, my pat h is firmly established, while you have equipped me with all the tools requisite for a bright career in academia. I am truly grateful for your leadership, encouragements, and owe you many thanks. To my doctoral committee, though our engagement has always been measured, your knowledge, recommendations, and careful insights are spot on. It is because of you that my research is filled with richness and rigor that is necessary to advance the knowledge base. The way each of you ha s challenged my thinking is gre atly appreciated. I will continue to refer to those conversations and tips as I develop my voice both as a Professor and mentor. For this, I kindly say thank you to Dr. Matt Syal, Dr. Richard DeShon, Dr. Ahmet Kirca, and Dr. George Berghorn . Next, I would like to thank my research cohort and friends that know all too well the hard work that has gone into th is milestone. Several recent doctoral recipients, Dr. Angelo Garciacortes, Dr. George Berghorn, and Dr. Daniel Duah have all been instrumental in my vi suc cess during graduate studies. And, of course, Faizan Shafique, who soon became part of th e core network as he and I studied together for one of our courses. We will forever share a bond formulated while studying and collaborating in the original fourth - flo or lab. This was our second home where we would share our common challenges while encouraging one another. All the best my friends. To the crew and their families, Santinio Jones, Anthony Eccelston, Dr. Eugene Blair, Dr. Chiron Graves, Nigel Fields , and D r. Steve Fletcher . Your kind words of support provided the much - needed motivation to complete this journey. This includes the Jarjoura Family, Stemple Family, Burn - Laliberte Family , Frye Family, Grantham Family, and specifically my golfing buddies Chris, G eorge, and Ali. It was especially important to have your lend an ear to my loving wife who has remained by my side to see th is finality . This is even more special as she was the impetus setting in motion my insatiable passion for knowledge. Things have not always been easy especially when the sacrifice s appear ed greater than the reward, but this core group continued to remind us otherwise. I am indebted to your families and say thank you from the heart. Last but certainly not least, I give my all to those nearest and dear to my heart: Pamela, Brooklyn, Treasure, and Buddy (Family Dog). I love you all so much. We have come so far since I in 2007 and subsequently a Ph.D. I pray you have begun to see the evidence illustrating how this path will continue to benefit you for years to come. We have endured many changes that only you know but each of you has remained strong along the way. It is my hope that you will be blessed by the resiliency, fortitude , and tenacity we have learned in the years to come. I also thank my parents: Larry Sparkling, James vii and Bobbie Overstreet who unknowingly paved the way for my success. As the youngest of four siblings, I have always maintained a keen eye and ear as to learn from those around me. It is this wisdom that somehow has brought me to this place. I offer a tremendous thank you to all Too many to name but I must acknowledge, my mother - in - law , Carrie Allen who has ste adily cheered us on feverishly without end . A final note of gratitude to many other educators, mentors, friends, collaborators, and consultants that are not mention ed by name from The International Partnering Institute (IPI) and its partnering facilitato rs, Eastern Michigan University (EMU) , Ronald E. McNair Scholars Program (EMU), Michigan State University (MSU) , Center for Statistical Training and Consultin g (MSU) , and Granger Construction Company. None of this would have be en possible without your supp ort so thank s . I am s ure to have missed some so thank you all in advance! viii This material is based on work supported by the National Science Foundation Graduate Research Fellowship Program under Grant No. DGE1424871. Any opinions, findings, and concl usions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the National Science Foundation. in put and collaboration efforts. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the International Partnering Institute. ix TABLE OF CONTENTS LIST OF TABLES ................................ ................................ ................................ ............................... xii LIST OF FIGURES ................................ ................................ ................................ ............................. xv KEY ABBREVIATIONS ................................ ................................ ................................ .................... xvii CHAPTER 1 INTRODUCTION ................................ ................................ ................................ ............ 1 1.1 Background ................................ ................................ ................................ ....................... 1 1.2 Problem Statement ................................ ................................ ................................ .......... 6 1.3 Research Goal and Objectives ................................ ................................ .......................... 9 1.4 Research Scope ................................ ................................ ................................ .............. 11 1.5 Research Questions ................................ ................................ ................................ ........ 13 1.6 Research Design and Approach ................................ ................................ ...................... 14 1.7 Expected Deliverables ................................ ................................ ................................ .... 17 1.8 Dissertation Organization ................................ ................................ .............................. 18 CHAPTER 2 LITERATURE REVIEW ................................ ................................ ................................ .. 19 2.1 Background ................................ ................................ ................................ ..................... 19 2.2 Project Partnering ................................ ................................ ................................ .......... 23 2.2.1 Project Risk Factors ................................ ................................ ................................ . 24 2.2.2 Collaborative Practices in Partner ing ................................ ................................ ...... 28 2.3 Partnered Project Delivery Framework ................................ ................................ ......... 32 2.4 Relational Governance, Teams, and Feedback ................................ .............................. 39 2.4.1 Relational Governance ................................ ................................ ............................ 39 2.4.2 Teams of Teams ................................ ................................ ................................ ...... 42 2.4.3 Feedback Processes and Theory ................................ ................................ ............. 45 2.5 Transactive Memory Systems ................................ ................................ ........................ 48 2.6 Project Teams in the AEC Industry ................................ ................................ ................. 49 2.6.1 Team Integration ................................ ................................ ................................ .... 50 2.6.2 Goal Alignment ................................ ................................ ................................ ....... 55 2.6.3 Collaborative Working Arrangements ................................ ................................ .... 57 2.7 Study Propositions and Theoretical Framework ................................ ............................ 61 2.8 Study Hypotheses and Theoretical Framework ................................ ............................. 62 2.9 Summary ................................ ................................ ................................ ........................ 63 CHAPTER 3 METHODOLOGY ................................ ................................ ................................ ......... 65 3.1 Introduction ................................ ................................ ................................ .................... 65 3.2 Research Goals and Objectives ................................ ................................ ...................... 66 3.3 Selecting the Research Strategy ................................ ................................ ..................... 67 3.4 Research Process ................................ ................................ ................................ ............ 69 x 3.5 Study Population ................................ ................................ ................................ ............ 70 3.6 Study Metrics ................................ ................................ ................................ ................. 71 3.6.1 Study Metrics for Qualitati ve Data ................................ ................................ ......... 73 3.6.2 Study Metrics for Quantitative Data ................................ ................................ ....... 82 3.7 Data Collection Procedure ................................ ................................ ............................. 87 3.7.1 Qualitative Data Collection ................................ ................................ ..................... 88 3.7.2 Quantitative Data Collection ................................ ................................ .................. 88 3.8 Qualitative Data Analysi s Methods ................................ ................................ ................ 89 3.8.1 Data Quality in Case Studies ................................ ................................ ................... 90 3.8.2 Case Study Data Analysis ................................ ................................ ........................ 92 3.1 Quantitative Survey Data Analysis Methods ................................ ................................ .. 94 3.1.1 Data Quality for Survey Measures ................................ ................................ .......... 95 3.1.2 Model Vali dation using Confirmatory Factor Analysis ................................ ........... 96 3.1.3 Statistical Analysis using Multilevel Modeling ................................ ........................ 97 3.2 Summary ................................ ................................ ................................ ...................... 102 CHAPTER 4 QUALITATIVE ANALYSIS ................................ ................................ ........................... 103 4.1 Summary of Case Study Projects ................................ ................................ .................. 103 4 .1.1 Case Study Project #1 ................................ ................................ ........................... 103 4.1.2 Case Study Project #2 ................................ ................................ ........................... 106 4.1.3 Case Study Project #3 ................................ ................................ ........................... 108 4.1.4 Case Study Project #4 ................................ ................................ ........................... 111 4.1.5 Case Study Project #5 ................................ ................................ ........................... 115 4.1.6 Case Study Project #6 ................................ ................................ ........................... 117 4.2 Pattern - matching of Case Study Projects ................................ ................................ .... 120 4.2.1 Characteristics of Case Study Projects ................................ ................................ .. 120 4.2.2 Cost and Schedule Growth Comparison ................................ ............................... 122 4.3 Cross - Case Synthesis of Case Study Projects ................................ ............................... 126 4.3.1 Initial Patterns and Trends for Case Studies ................................ ......................... 127 4.3.2 Proposition Testing using Case Study Analysis ................................ ..................... 134 4.4 Summary ................................ ................................ ................................ ...................... 157 CHAPTER 5 QUANTITATIVE ANALYSIS ................................ ................................ ......................... 158 5.1 Case Study Data Demographics ................................ ................................ ................... 159 5.2 Confirmatory Factor Analysis ................................ ................................ ....................... 161 5.3 Structural Equation Modeling (SEM) ................................ ................................ ........... 169 5.4 Multiple Regression/Correlation An alysis ................................ ................................ .... 169 5.5 Summary ................................ ................................ ................................ ...................... 177 CHAPTER 6 DISCUSSIONS ................................ ................................ ................................ ............ 178 6.1.1 Quali tative Findings ................................ ................................ .............................. 178 6.1.2 Quantitative Findings ................................ ................................ ............................ 180 6.1.3 Study Framework Finding Summarized ................................ ................................ 182 xi 6.2 Utility of Study Metrics ................................ ................................ ................................ 183 6.3 - ................................ ........ 186 6.4 Theoretical Contributions ................................ ................................ ............................ 188 6.5 Summary ................................ ................................ ................................ ...................... 191 CHAPTER 7 CONCLUSIONS ................................ ................................ ................................ .......... 192 7.1 Summary of Research Goals and Objectives ................................ ................................ 192 7.2 Summary of Research Methods ................................ ................................ ................... 193 7.3 Summary of Findings ................................ ................................ ................................ .... 194 7.4 Contributions to the Body of Knowledge ................................ ................................ ..... 196 7.5 Limitations and Suggestions for Future Research ................................ ........................ 198 APPENDICES ................................ ................................ ................................ ................................ 200 APPENDIX A: Partnered - project Case Study Survey Instrument ................................ ............. 201 APPENDIX B: Structured Interview Questions ................................ ................................ ........ 209 APPENDIX C: Full Mplus Code and Output for Confirmatory Factor Analysis of Study Measurement Model ................................ ................................ ................................ .............. 21 5 APPENDIX D: Full Mplus Code and Output for Confirmatory Factor Analysis of Study Latent Variables ................................ ................................ ................................ ................................ .. 228 APPENDIX E: Full Mplus Code and Output for SEM Reliability of Latent Variabl es ................ 250 APPENDIX F: Full SAS Code and Results for Nonlinearity SEM Reliability of Latent Variables 274 REFERENCES ................................ ................................ ................................ ................................ 305 xii LIST OF TABLES Table 2 - 1: Analysis of Partnered - Project Delivery Framework Category: Drivers during delivery ( Sparkling et al., 2016) ................................ ................................ ................................ ........... 35 Table 2 - 2: Analysis of Partnered - Project Delivery Framework Category: Team Characteristics ( Sparkling et al., 2016) ................................ ................................ ................................ ........... 37 Table 2 - 3: Analysis of Partnered - Pr oject Delivery Framework Category: Project Performance Outcomes ( Sparkling et al., 2016) ................................ ................................ ......................... 38 Table 3 - 1: Measures used to assess common project risk factors (Gransberg et al., 1999; IPI, 2016). ................................ ................................ ................................ ................................ ..... 74 Table 3 - 2: Full description of collaborative project delivery practices construct, metrics and measures used in this study (Chan et al., 2004; IPI, 2016; Mollaoglu & Sparkling, 2015). ... 77 Table 3 - 3: Full description of the project performance constructs, metrics and measures. ....... 80 Table 3 - 4: Full description of goa l alignment construct and measures used in this study. ......... 83 Table 3 - 5: Full description of transactive memory system construct, metrics and measures used in this study (Lewis, 2003). ................................ ................................ ................................ .... 84 Table 3 - 6: Individual/team performance construct, metrics and measures used in this study. . 86 Table 4 - 1: Case Study #1 Project Charte r Goals and Performance Metrics ............................... 105 Table 4 - 2: Case Study #2 Project Charter Goals and Performance Metrics ............................... 107 Table 4 - 3: Case Study #3 Project Charter Goals and Performance Metrics ............................... 110 Table 4 - 4: Case Study #4 Project Charter Goals and Performance Metrics ............................... 113 Table 4 - 5: Case Study #5 Project Charter Goals and Performance Metrics ............................... 116 Table 4 - 6: Case Study # 6 Project Charter Goals and Performance Metrics ............................... 119 Table 4 - 7: Characteristics of Case Study Projects ................................ ................................ ...... 121 Table 4 - 8: Case Study Partnering cost as a percentage of original contract value. ................... 122 xiii Table 4 - 9: Overall Project Performance Ranking ................................ ................................ ....... 126 Table 4 - 10: Project Risk Factors and Overall Project Performance Rank for Case Studies. ....... 129 Table 4 - 11: Comparing Project Risk Factors among Overall Project Performance Rank ........... 131 Table 4 - 12: Comparison of Highest and Low est Overall Project Performing Case Study. ......... 132 Table 4 - 13: Comparison of Project Type and High/Low Overall Project Performing Case Study. ................................ ................................ ................................ ................................ ............. 133 Table 4 - 14: Case Study Scores used in Proposition Testing sorted by Case Study Number ...... 138 Table 4 - 15: Results from Structured Interviews and Project Scorecards sorted by Proje ct Risk Scores ................................ ................................ ................................ ................................ .. 139 Table 4 - 16: Project Risk Factors and Overall Risk Scores ................................ ........................... 140 Table 4 - 17: Collaborative Project Delivery Pra ctices sorted by Overall Score ........................... 143 Table 4 - 18: Results from Structured Interviews and Project Scorecards sorted by Collaborative Project Delivery Practice Scores ................................ ................................ .......................... 145 Table 4 - 19: Results from High/Low Score Analysis sorted by Case Study ................................ . 147 Table 4 - 20: Results from Structured Interviews and Project Scorecards sorted by Goal Alignment Scores ................................ ................................ ................................ ................. 149 Table 4 - 21: Goal Alignment Metrics used in Case Studies sorted by Project Performance ...... 152 Table 4 - 22: Goal Alignment and Congruence items from Partnering Documents .................... 155 Table 5 - 1: Case Study Sample Demographics ................................ ................................ ............ 160 Table 5 - 2 : Summary of Study Sample and Responses ................................ ............................... 160 Table 5 - 3: Respondent Demographics based on Project Role ................................ ................... 161 Table 5 - 4: Factor Str ucture and Factor Loadings for Goal Alignment ................................ ........ 163 Table 5 - 5: Factor Structure and Factor Loadings for Transactive Memory Systems ................. 165 Table 5 - 6: Factor Structure and Factor Loadings for Individual/Team Performance ................ 167 xiv Table 5 - 7: Correlations among Higher - order Latent Variables and Sub - factors ........................ 168 Table 5 - 8: Descriptive statistics for Goal Alignment, TMS, and Individual/Team Performance 173 Table 5 - 9: Descriptive statistics from multivariate regression analysis ................................ ..... 175 Table 6 - 1: Descriptive statistics for the relationship among study variables ............................ 182 Table 6 - 2: Findings from CFA for goal alignment and TMS ................................ ........................ 185 Table 6 - 3: Findings from CFA for individual performance ................................ ......................... 186 xv LIST OF FIGURES Figure 1 - 1: Framework illustrating the relationships between project risk factors, collaborative project delivery practices, goal alignment, TMS, and performance outcomes. ................... 11 Figure 1 - 2: Three primary stages followed in this study as part of the research approach. ....... 17 Figure 2 - 1: Collaborative model: dynamic learning capability amongst multiple stakeholders across organizational levels over the project life cycle, with performance feedback (Manley & Chen, 2015). ................................ ................................ ................................ ....................... 21 Figure 2 - 2: IPI Vertical Construction Project Partnering Scalability matrix for collaborative partnering. ................................ ................................ ................................ ............................. 27 Figure 2 - 3: Partnered project delivery framework adopted from Sparkling et al. 2016 ............. 33 Figure 2 - 4: Relationships and propositions between project risk factors, collaborative project delivery practices, and project performance ................................ ................................ ........ 62 Figure 2 - 5: The relationship and hypotheses between goal alignment, transactive memory systems, and individual/team performance ................................ ................................ ......... 63 Figure 3 - 1 : Research process and steps followed in this study . ................................ ................... 70 Figure 3 - 2: Sample Case Study Project Scorecard, Goal Aligning Actions, and Performance Metric ................................ ................................ ................................ ................................ .... 78 Figure 3 - 3: The Relationship and Hypotheses between Goal alignment, Transactive Mem ory Systems, and Individual/Team performance ................................ ................................ ........ 97 Figure 3 - 4: Multilevel sample selected at random from population ................................ ........... 99 Figure 4 - 1 : Cost Growth for Micro - Small Projects or less than $10M ................................ ....... 123 Figure 4 - 2: Cost Growth for Large - Mega Projects or greater than $25M ................................ .. 124 Figure 4 - 3: Schedule Growth for Case Study Projects based on Workday Durations ................ 125 Figure 4 - 4 : Common Project Risk Factors Evaluation Method ................................ ................... 135 Figure 4 - 5: Collaborative Project Delivery Evaluation Method ................................ .................. 136 Figure 4 - 6: Theoretical Framework for Relationships among Project Risk Factors, Coll aborative Project Delivery Practices, Goal Alignment, and Project Performance. ............................. 137 xvi Figure 4 - 7: Sample Case Study Project Scorecard, Goal Aligning Actions, and Performance Metric ................................ ................................ ................................ ................................ .. 149 Figure 4 - 8: Goal Alignment ratings from Scorecards (Large - Mega) ................................ ........... 153 Figure 5 - 1: Goal Alignment Latent Variable and Factor Indicators ................................ ............ 162 Figure 5 - 2: Factor Structure for the TMS Latent Variable; * Indicators with weak loadings or negative residual variances ................................ ................................ ................................ . 164 Figure 5 - 3: Factor Structure for the Individual/Team Performance Latent Variable; * Indicators with wea k loadings or negative residual variances ................................ ............................. 166 Figure 5 - 4: P - P Plot for individual performance ................................ ................................ ......... 171 Figure 5 - 5: Histogram and distribution curve for individual performance ................................ 172 Figure 5 - 6: Residual plot of individual performance ................................ ................................ .. 172 xvii KEY TO ABBREVIATIONS AEC Architecture Engineering and Construction AIA American Institute of Architects BEA United States Bureau of Economic A nalysis BIM Building Information Modeling CFA Confirmatory Factor Analysis CFI Comparative Fit Index CII Construction Industry Instit ute CM Construction Management CM/GC Construction Management/General Contractor DB Design - build DBB Design Bid Build DBE Disadvantaged Business Enterprise DOT Department of Transportation FA Factor Analysis FAST Focused Action Strategic Team GDP Gross D omestic Product IPD Integrated Project Delivery IPI International Partnering Institute IRB Institutional Review Board JV Joint Venture LBE Local Business Enterprise xviii LPS Last Planner System MOT Maintenance of Traffic MRC Multivariate Regression Correlati on OLS Ordinary Least Squares QA Quality Assurance QC Quality Control QO Quality Oversight RMSEA Root Mean Squared Error of Approximation SAS Statistical Analysis Software SEM Structural Equation Modeling SME Subject Matter Expert SMM Shared Mental Mode l TCE Transaction Cost Economics TMM Team Mental Model TMS Transactive Memory System s US United States 1 CHAPTER 1 INTRODUCTION This chapter broadly covers current practices and dilemmas facing interorganizational architectural, engineering, and c onstruction (AEC) project teams. Particularly, as they strive to align their goals and objectives while maximizing performance outcomes. The problem statement in Section 1.2 is guided by collaborative project delivery practices and associated challenges di scussed next. Construction is often considered a gritty, hard - nosed, hard - hat industry. The industry, despite its image, accounts for approximately four percent (4.3 %) of the United States (U.S.) gross domestic product (GDP) or $ 826.1 billion do llars in expected added value in 2018 (Bureau of Economic Analysis (BEA) , 2018) . It encompasses a host of design and construction professionals such as architects, engineers, steelworkers , electricians, and pl umbers, to name a few. These disciplines come together working towards one common goal, to deliver construction projects for their clients safely, on time, and below budgeted costs. As various organizations come together, they bring their vast knowledge base and expertise. Individuals from respective organizations are expected to deploy their knowledge to advance project objectives while controlling organizational and project risks. But, what happens when they fail to agree on certain aspects of the proje ct such as the price of a completed change order resulting from added design scope? What about other project risks highlighted by the industry such as errors, omissions, cost overruns, and productivity losses? These are common dilemmas which occur on many construction projects and inherently imposes risks on 2 the entire project team . R ahman and Kumaraswamy ( 2002) maintain risks must be dealt with whether by sharing, transferring, managing, accepting, or controlling for it. Many problems in construction projects are exacerbated due to the lack of clear communication, coordination, and early collaboration. Consequently, companies rely on claims, arbitration and/or litigation to solve their problems. This well - known fact has been delineated by industry practitioners and researchers alike, changing the way owners and contractors deliver co nstruction project while managing risks such as cost, schedule, and project uncertainties. Traditional construction project delivery methods such as design - bid - build (DBB), construction management (CM), and design - build (DB) have dominated construction c ontracts in the U.S. for years. These trusted methods, though effective, do not always encourage collaboration and communication across organizations during the early planning stages of the construction process. In fact, many of these approaches put contra ctors in a position where behaviors are focused on transaction costs [i.e., transaction cost economics (TCE) or any activity engaged in to satisfy each party to an exchange values in accord with expectations that are both given and received (Ouchi, 1980) ] and positioning themselves against uncertainty involved in project delivery (Li, Arditi, & Wang, 2013) . More recent relational project delivery methodologies (e.g., Project Partnering, Strategic or Project Alli ancing, and Integrated Project Delivery [IPD]) surfaced in the late 1980s continuing over a fifteen year period bent on increasing levels of collaboration across organizations and to help mitigate risks (Lahdenperä, 2012) . Relational governance theory 3 the enforcement of ob ligations, promises, and expectations occurring th rough social processes that promote while normalizing flexib ility, solidarity, and information exchange (Poppo & Zenger, 2002) found in traditional contracts are relaxed as individuals focus on trust to minimize opportunistic behaviors (Zaheer & Venkatraman, 1995) . These newer collaborative contracting practices came into exi stence to meet the expectations of clients in delivering predictable results and satisfaction through integrated teams ( Baiden, Price, & Dainty, 2003) . An integrated team as defined by Baiden et al. (2003) is : a team of individually distinct groups or teams with functional identities wor king together consciously and in a c ontinuous way to achieve a set objective or target through a system of unrestricted cross - sharing of information. In turn, efficient and effective decision making occurs under competent team leadership with the ability to drive the overall optimum achievem Within this definition , the motivation is to merge multi - disciplinary organizations into one cohesive unit singularly responsible to the client and is culturally joined together. However, removing cultural barriers r equires strategies such as project team member consistency, colocation, and early involvement of all team members (e.g., prime contractors and specialty subcontractors ) in decision - making ( Baiden et al., 2003) . These team integration strategies help project team members share information openly and honestly while tapping into a broad range of knowledge and exper tise early on when decisions are less costly and more effective (Ospina - Alvarado, Cast ro - Lacouture, & Roberts, 4 2016) . These distinct advantages comprised of experiences, mental models [i.e., an organized structure of shared knowledge among the team (Moh ammed , & Dumville, 2001) ], and motivation bring s about goal alignment within project teams (Dietrich, Eskerod, Dalcher, & Sandhawalia, 2010) . Recent research points out existing paradoxes in the field of AEC project management and collaborative approaches as being generalized id eas that, when followed, do not always improve performance (Jacobsson & Roth, 2014) . In particular, partnering research urges us to place greater emphasis on informal attributes and formal me chanisms to add richness to its practice and implemen tation (Bresnen, 2007; Suprapto, Bakker, & Mooi, 2015) . As an example, a Delphi study of industry professionals was used to establish an alliance team integration performance index using quantitative measures (Che Ibrahim, Costello, & Wilkinson, 2015) . Their study provides a useful tool for AEC researchers and practitioners that monitor the strengths and weaknesses of team integration during project delivery. In another instance, Ospina - Alvarado et al. (2016) assembled a ranking of integration attributes based on a survey of AEC professionals to deli neate what team integration means to a project. Both studies broadly addres s the need suggested by Baiden, Price, an d Dainty (2006) to develop team integration measure s , yet are sparse in illustrating the behavioral commitments and understandings required of individuals . The literature lacks clear evidence demonstrating how collaborative practices (e.g., partnering wo rkshops, partnering training sessions) enhance team integration and, more importantly, goal alignment and behavioral systems. Some argue , increasing collaborative practices alone does not improve team integration and performance outcomes ( Cheng & Li, 5 2001; Kumaraswamy et al., 2005) . However, these collaborative practices are still believed to help interorganizationa l AEC project teams align their goals and objectives for the benefit of the individual , team, and project performance (Dietrich et al., 2010) . Although the architecture, engineering, and construction (AEC) literature are widespread on the importance of team integration and cohes ion, little emphasis is placed on the goal alignment aspect and its relationship to performance outcomes . This research aims to explore the relationship between goal alignment and performance outcomes in interorganizational AEC project teams, in the conte x t of partnered - projects. Additionally, it examines t he moderating effects of transactive memory systems (TMS) on this relationship. Understanding this is important to address a gap in the literature, therefore, several objectives are posited for this study . The objectives of this study are to (1) Develop a framework demonstrating the relationship between project risk factors, collaborative project delivery practices, goal alignment, TMS , and performance outcomes ; (2) Test an evaluation metric for goal align ment and utility of TMS metric to investigate AEC collaboration during project delivery; (3) Help facilitate collaborative contracting in construction projects by identifying key characteristics individuals hold in common ; and, (4) Provide theoretical cont ributions in AEC literature understanding collaborative project delivery methodologies as a form of relational governance. T o achieve the study objectives, surveys of industry professionals involved in collaborative project delivery is collecte d from six case studies . This permitted the researcher to investigate the effects of goal alignment, in the context of partnered projects, on individual, team and project performance. It is further ant 6 memory system (i.e ., knowledge processing system for coordination, co mmunication, and specialization ) acts as a moderator of performance outcomes. Process and performance feedback (i.e., monthly partnering meetings, partnering training sessi ons, professional neutral third - p arty facilitator, monthly scorecards, etc.) is given to project team members at defined intervals during project delivery and is inherent to partnering processes. This feedback intends to hel p project team members align their goals, resources, and efforts with those of the project. It also encourages team collaboration and trusting relationships to develop within and between project teams. To summarize, this section described the challenges and continued efforts the AEC industry is pursuing to address tea m collaboration. Despite adversarial mentalities reported in traditional project delivery methods, collaborative project delivery methods and approaches such as IPD and partnering are closing in on key informal attributes and mechanisms. The literature, as evinced above, is still emerging in explaining how collaborative practices manifest within AEC project teams. Thus, the problem statement below articulates the rationale for this dissertation . AEC construction project teams increasingl y seek to improve project performance efficiency (i.e., cost and schedule) and effectiveness (i.e., quality and safety) while also mitigating conflicts without clear guidance or directi on (Suprapto et al., 2015) . Many attribute this dilemma to the fragmentation among discipli nes dur ing project delivery processes (Fellows & Liu , 2012; Roehrich & Lewis, 2010) . The benefits of collaborative working arrangements such as project partnering is typified as a vital component to help address this discontinuity in project teams 7 (Anderson & Polking horn, 2011; Black, Akintoye, & Fitzgerald, 2000; Bubshait, 2001; Gransberg, Dillon, Reynolds, & Boyd, 1999; Hong, Chan, Chan, & Yeung, 2012) . Anderson and Polkinghorn (2011) report several key benefits of project partnering such as improved team collabora tion, improved conflict resolution strategies to manage project risks and uncertainty , stronger relationships and increased trust among project teams. Few of these studies, until recently, have examined how team related factors are related to project perfo rmances in the context of AEC project teams (Comu, Iorio, Taylor, & Dossick, 2013; Franz, L eicht, Molenaar, & Messner, 2016) . A gap exists in the AEC litera ture to advance this topic, in particular, the role goal alignment and TMS have in the project delivery process. For AEC industry researchers and practitioners, it is important to advance project - based collaboration as a means to improve the certainty of project performance outcomes . Especially considering challenges from the temporary nature of construction projects and its ever - changing teams. In fact, Suprapto et al. (2015) suggest continuous attention must be applied by managers to enhance the benefits of collaborative practices (e.g., project partnering) on successful project delivery. This information is of great value to both owners and contractors in an industry often associated with trust and communication issues resulting in construction disputes and adversarial mentalities (Bresnen, 2007; Drexler Jr. & Larson, 2000; Ng, Rose, Mak, & Chen, 2002) . The challenge inherent in AEC projects is to gain willin g cooperation among interorganizational project teams . These teams, while transient across projects, pick up habits and practices through experience that can have a positive or negative effect on performance . Not only are there project teams, but teams of teams at various level s that form over the life 8 cycle of a project . For example, AEC projects are comprised of individual team members (i.e., individual members of each organization) and interorganizational sub - team, design and enginee ring team, contractor team, and subcontracting teams). Collaborative project delivery approaches attempt to bring these teams together early and often during project delivery. Reason being, one can expect optimized results and efficiencies through all phas es of design, fabrication, and construction (AIA, 2007) . Despite noticeable bene fits, some contend collaborative working arrangements are grounded within interpersonal dynamics and relationships which explain the effect on project performance (Bresnen, 2009; Bygballe, Jahre, & Swärd, 2010 ; Cacamis & El Asmar, 2014; Doloi, 2009; Yeung, Chan, & Chan, 2012) . In other words, AEC projects are people - oriented practices which require varied organizations to coalesce around shared goals and objectives. Many of these may conflict with those of the individual or organization. Hence, several levels of analyses are present with disparate objectives for individuals versus interorganizational sub - teams working on an AEC project. Research shows how behavioral commitments involve team cognitive processes (e.g., team mental models, transactive memory systems) and impacts process and project performance for team members involved in information system development projects ( Hsu, Chang, Klein, & Jiang, 2011; J. S. Hsu, Liang, Wu, Klein, & Jiang, 2011) . Hsu et al. (2011b) assert continuous team building activities increases team mental models and ultimately facilitates bett er problem solving, goal alignment, learning, and improves performance outcomes. practices on goal alignment among individuals. 9 The pri mary aim s of this research are to investigate how goal alignment affect s performance in AEC project teams when collaborative project delivery practices are followed An interesting phenomenon is present where teams of teams working on projects attempt to align individual goals with those of interorganizational sub - teams. A gap exists in the AEC literature, demonstrating a need to systematically identify the underlying attributes of collaborative project delivery approaches which result in better performan ce. Focusing on partnered - projects as a type and subset of collaborative AEC project delivery approaches, the specific objectives of this study are to: 1. Develop a framework demonstrating the relationship between project risk factors, collaborative project d elivery practices, goal alignment, TMS , and performance outcomes by; a. Qualitatively examining the following at partnered - project level: i. The links among project risk factors , collaborative project delivery practices , and project performance ; and, b. Quantitativ ely examining the following at individual - level in interorganizational AEC project teams: i. The relationship between individual /team performance , goal alignment , and TMS . 10 2. Test an evaluation metric for goal alignment and utility of TMS metric to investigate AEC collaboratio n during project delivery; 3. Help facilitate collaborative contracting in construction projects by identifying key characteristics individuals in common. 4. Provide theoretical contributions to AEC literature understanding co llaborative project delivery methodologies as a form of relational governance. Based on the goals and objectives of this research, a conceptual framework in Figure 1.1 is proffered illustrating the moderating role of team behaviors during project delivery. The framework is used as a guide for this study to investigate the relationships among project risk factors , collaborative project delivery practices , goal alignment , TMS , and performance outcomes . In addition, several metrics are tested which are made av ailable to researchers in AEC literature for future research involving AEC project teams. T hus, t hese metrics which are generally found within organizational research are advanced to AEC literature. 11 Figure 1 - 1 : Framework illustrating the relationships between project risk factors, collaborative project delivery practices, goal alignment, TMS, and performance outcomes. This study examine s collaborative project delivery through the lens of partnered - projects . Partnering in the context of AEC industry is more commonly defined as - term commitment between two or more organizations for the purpose of achieving specific business (Construction Industry Institute [CII], 1989) . A purposeful pool of partnered - projects is used to collect both individual - level data from AEC project teams and project information representing a small subset of the AEC construction industry. The final study participant sample (n) size was 125 potential survey respondents from six case study projects. Collaborative Project Delivery Practices Individual/Team Performance Project Performance Goal Alignment Project Level Data / Qualitative Investigation TMS Individual Level Data / Quantitative Investigation Risk Factors 12 This is achieved in collaboration with the International Partnering Institute (IPI) to identify ongoing or recently completed partnered - projects. By distinctively selecting the sample the researcher is able to find individuals who experience the phenomena, are representative of the ideal population, can be assessed through survey or observation on information related to the research question (Miller et al., 2011) . Miller et al., ( 2011) further argue how important it is to understand the mult ilayered and temporal structure within the phenomena as it w ill guide the analyses approach. Project risk factors such as size and complexity coupled with collaborative practice information are characteristics of each project. Some projects (e.g., partnere d - projects) employ collaborative practices during project delivery such as creating partnering charters, partnering workshops, and using partnering scorecards and performance surveys to objectively measure how well teams are performing on pre - determined go als and objectives. These project characteristics, though limited by the number of case study projects, will offer both quantitative and qualitative information to enhance the intended data analysis for this study. The proje ct team level , as a unit of anal ysis , becomes useful as project - specific information is integrated into analyses and used to triangulate data across multiple case studies (Campbell & Fiske, 1959) . Collaborative working occurs during many project d elivery approaches and encourages team integration. Collaborative project delivery practices such as p roject p artnering seek to increase team integration and owner value while reducing waste and inefficiencies during project delivery (Lahdenperä, 2012; Pishdad - Bozorgi & Beliveau, 2016) . The underlying motivation behind collaborative project delivery arrangements is to spread risks and rewards 13 evenly among project stakeholders (i .e., owner, design, and contractor). This inherently influences individual, sub - team, and project performance outcomes due to behavioral norms enabled through formal and informal contract agreements. Informal relational contracting strategies as those men tioned above build solidarity, flexibility, and trust within AEC project teams. Thus, partnered - projects are well - suited to represent collaborative project delivery methodologies because they can respond to both teams of teams and relational governance the ory. Teams of teams are widespread in AEC projects, yet are not systematically investigated in AEC literature . how goal alignment affect s performance in AEC project teams whe n collaborative project delivery practices are followed Two levels of analyses are relevant to this research question , project /team level , and individual - level (see Figure 1 - 1 above) . The specific questions intended to address each level of analysi s are given below: Project Level Questions What collaborative project delivery practices impact project goal alignment in partnered - projects? Do project risk factors impact collaborative project delivery practices and, thus, goal alignment ? 14 What are the relat ionships between project risk factors , collaborative project delivery practices , and goal alignment in partnered - projects? Does goal alignment affect project performance ? Individual - level Questions Does individual goal alignment affect individual perform ance perceptions ? If so, is this relationship moderated by individual TMS ? The research design is central to any research study and requires many thought trials. This provides not only the plan but the structure required to ef fectively answer research questions and control variance (Kerlinger & Lee, 1999) . The goals of this study are obtained by investigating the variance across AEC project teams using multiple case study evidence from six projects . This study utilizes a mixed - methods research approach to collect, analyze, and draw inferences among project and individual - level data. The research documents and study data collection instruments were presented to the Institutional Review Board (IRB) at Michigan State University for appro val. The documents include d the survey instruments, structured interview questions, and other protocols required for human subject research. This study received an expedited approval due to its voluntary identifiers are requested from study participants. 15 To achieve the main objectives of this research , a mixed - method approach is followed . A mixed - method approach facilitates exploratory research using two or more types of data collection and data analysis m ethods (i.e., qualitative , quantitative ). These data are used to integrate findings and draw inferences from a single study or theoretical perspective (Miller et al., 2011) . The two types of data, per Miller et al. ( 2011) , enhances data collected sequentially and provide s a complete understanding of the phenomenon. Thus, multiple research questions and hypotheses are asserted as in sections 1.5 - 1.7 above. There are three stages followed in this study. Project - specific data such as collaborative project delivery practice s (e.g., use of partnering charters, scorecards, workshops) and project team rosters are collected from main project participants (e.g., partnering facilitator, owner representative, or Construction Manager/General Contractor (CM/GC) project manager ). Stru ctured interviews are performed to collect data from key stakeholders representing the owner team (e.g., owner, owner representative, or other key stakeholder involved in project delivery) regarding project performance outcomes (e.g., cost, schedule, quali ty, and conflict resolution). Using project rosters, q uantitative data is gathered from surveys administered to project team members . Data collected from surveys are used o f goal alignment, TMS, and performance outcomes . The survey is used to collect data at or near project completion. It is anticipated perceptions will change during the life cycle of a project, thus vary across individuals. Therefore, surveys are initiated at or near project completion to capture a rece 16 During stage one, a partnered project delivery framework (A. E. Sparkling, Mollaoglu, & Kirca, 2016) is initiated vi a case study evidence to verify its applicability to project level data analysis (Sohani, 2016) . From this initial investigation, the framework is refined and used to develop a comprehensive survey incorporating literature review and industry feedback. At stage two, the researcher co llects data from the identified partnered - projects and subsequent project teams using surveys, structured interviews , and email correspondence. Last, data analysis is completed among both qualitative and quantitative data. The researcher uses pattern - matching and content analysis as a means to provide a cross - case synthesis. Concurrently, q uantitative data are analyzed using confirmatory fa ctor analysis and multiple regression /correlation analysis (MRC) . Chapter 3 gives a detailed description of the stages and mixed - method approach followed in this study while the expected deliverables are covered 17 next. Figure 1 - 2 : Three primary stages foll owed in this study as part of the research approach. The expected theoretical and practical contributions to the body of knowledge as a result of this research are : (1) A framework demonstrating the relationship between project risk facto rs, collaborative project delivery practices, goal alignment, TMS, and performance outcomes; (2) Test an evaluation metric for goal alignment and utility of TMS metric to investigate AEC collaboration during project delivery; Qualitative Data 1. Content analysis 2. Case study tactics (i.e., pattern - matching and cross - case synthesis) Quantitative Data 1. CFA procedures for measurement and model validation 2.Multiple regression/correlation analysis 3. Descriptive statistics Data Analysis 1. Coordinate study goals/needs with IPI 2. Contact project leads to assess willingness to participate 3. Determine if selected sample of projects is representative for study goals 4.Collect data using web - based surveys (i.e., quantitative data), structured interviews (i.e., qualitative data), and email correspondence to capture project level data (i.e., qualitative data). Data Collection 1. Analyze prior case study findings (Sohani, 2016) 2. Refine and extend prior partnered project delivery framework (Sparkling et al. 2016) 3. Develop survey metrics Developing Metrics based on Previous Research 18 (3) Best practice guidance on collabo rative contracting practices and associated behavioral attributes which underlie effective imple mentation; and, (4) Future guidance on key project team metrics to explore processes integral to better performance outcomes in collaborative project delivery metho dologies. This dissertation is structured around seven chapters. Chapter 1 provid ed an introduction to relevant literature underpinning this research and included the problem statement, research goals and objectives, research scop e , research questions, stated hypotheses and propositions , research design and approach , and expected deliverables. Chapter 2 provides an in - depth review of the literature for this research focusing on relational governance, team theory , goal alignment, an d underlying characteristics of project partnering. The methodology is presented in Chapter 3 which includes constructs, metrics, and survey development. Detailed results are reported in Chapter 4 stemming from cross - case synthesis procedures . Chapter 5 ex plicate s the results of data and model validation while Chapter 6 present s findings and discussion from both multivariate regression analyses and cross - case synthesis . Chapter 7 conclude s this research by summarizing the findings, conclusions, limitations, theoretical contributions and directions for future research. 19 CHAPTER 2 LITERATURE REVIEW The chapter begins by introducing the current state of the literature related to collaborative working arrangements in construction projects. More importantl y, why the issue of team integration and its implications to project performance continues to remain unclear. This is followed by an introduction to project partnering, contractual , and procurement practices which challenge project teams. These theories ar e beneficial in unlocking misguided assumptions often attributed to newer construction practices and their benefits to increased performance. Next, the literature shifts to provide an in - depth exploration into the well - defined team and relational governanc e theories generally found in organizational research directing attention to AEC collaborative working implications. The AEC industry is increasingly challenged with improving the efficacy of project team performance through collaborative work ing arrangements (Suprapto et al., 2015) . Collaborative working arrangements synonymous with relational contracting methods such as alliancing, joint ventures, and project partnering are all comprised of interorganizational project teams (Rahman & Kumaraswamy, 2005; Suprapto et al., 2015) . Recent research shows that cohesive teams are perpetuated by strategies to facilitate team integration (Franz, Leicht, Molenaar, & Messner, 2010; Franz & Leicht, 2016) . Efficient knowledge sharing and processing systems also called transa ctive memory systems (TMS), are integral to cohesive project teams and their tasks coordination (Comu et al., 2013) . 20 A strong TMS allows two or more people to cooperatively and efficiently encod e, store, retrieve, and communicate information from different subject experts (Hollingshead, 1998b; Kyle Lewis, 2003) . A TMS exists within interorganizational project teams and can guide the success of th e project. In the AEC industry, project teams are generally responsible for the success of a project although from different perspectives. These teams must work to align competing goals between their respective organizations versus those related to the pr oject. An opportunity exists within collaborative projects which challenge this interesting dichotomy by way of feedback systems. According to Manley and Chen (2015) , project governance mechanisms whether formally (e.g., open book cost accounting or shared risks structures) or informally (e.g., integra ted team selections or relationship workshops) instituted in contracts provides feedback enhancing project outcomes. From this, learning routines are developed among project teams. It is within these learning routines where individuals and organizations ga in feedback which increases their collaborative project understanding (Manley & Chen, 2015) . T his feedback information is brought forward into future projects and is related to positive performance outcomes. The idea of dynamic learning feedback system in AEC literature is presented in the collaborative model asserted by Manley and Chen (2105). Fi gure 2 - 1 shows their conceptual model for a dynamic learning system among multiple stakeholders and across organizational levels during project delivery. In the model, collaborative relationships established during a project life cycle are at the center of the feedback loop. According to Manley and Chen (2015), a three - stage sequential learning process exist s and is comprised of exploratory, 21 transformative, and exploitative learning. Exploratory learning occurs during routines of such as workshops where tea m members freely communicate and share knowledge to achieve mutual goals and objectives. Figure 2 - 1 : Collaborative model: dynamic learning capability amongst multiple stakeholders across organizational levels over the projec t life cycle , with performance feedback (Manley & Chen, 2015). Transformative learning is the process whereby individuals exchange, disseminate, and codify knowledge for future use during interpersonal interactions. Knowledge gained from 22 collaborative proj ect experience through exploratory and transformative learning is made available or matched to future conditions. This is referred to as exploitative learning. In the exp erience is increased. As a result, the operating and governance mechanisms lead to empowered decision - making behaviors among project teams (Tuuli & Rowlinson, 2009) . G roup learning and feedback encourages team members to commit resources towards multiple goals and objectives (DeShon, Kozlowski, Schmidt, Milner, & Wiechmann, 200 4) . This is important within construction projects due to the interdependencies of organizations in project delivery. Broadly, t here is an ongoing effort to efficiently manage resources always seeking to reach an optimal point. However, the effects of fee dback are generally not investigated. This feedback system helps the team develop a systematic way to realign efforts and resources to the project when deviations from goals are detected. In doing so, a system emerges that allows the team to manage knowled ge and information related to the project and can be essential to performance outcomes. The performance of misaligned teams has been connected to the type of feedback instigated (e.g., diagnostic information and process related) and can vary based on the makeup of the team (Johnson, Hollenbeck, Scott DeRue, Barnes, & Jundt, 2013) . For example, interorganizational teams operating under a project partnering arrangement (e.g., construction project teams) will respond to feedback different ly than an intraorganizational self - managed team (e.g., product development team). It is, therefore, anticipated that transactive memory systems vary across organizations and impact performance outcomes especially when feedback is present. This shared know ledge system enables interorganizational project teams to 23 efficiently coordinate tasks and information processing during project delivery. What motivates and limits the effectiveness of many organizations working in teams are transaction costs in project d elivery (Li et al., 2013) and building a truly integrated team (Franz et al., 2016) . Ultimately, project risks and uncertainties are minimized with increased team cohesion, integration, and a well - developed TMS . Partnering in the context of AEC industry is more commonly defined as - term commitment between two or more organizations for the purpose of achieving specific business (Construction Industry Institute [CII], 1989) . Construction partnering is a netwo rk of self - organizing project teams which includes feedback systems that help to align project objectives with shared business goals and expectations (Bennett & Peace, 2007; Construct ion Industry Institute [CII], 1989) . These interorganizational project teams are typically represented as owners, designers, and contractors. Briefly, project partnering was developed in the 1980s by the U.S. Army Corps of Engineers as a way to mitigate construction disputes using joint workshops between owners and contractors (CII, 1989). It started as a voluntary arrangement between these two parties and has since evolved into a practice that is followed formally in contracts (Lahdenperä, 2012) . The approach commonly uses a partnering charter to establish agreed upon goals and objectives for the project and its participants. A partnering champion or facilitator is used to help lead the process during workshops and partner ing meeting. The primary motivation is to ensure project objectives are the focal point for the team. It is also important to communicate a 24 succinct decision - making process to resolve issues or disputes. Project partnering is also underutilized in the cons truction industry, while longstanding and classified as a best practice (Construction Industry Institute [CII], 1996, Lahdenpera, 2012). Partnering literature broadly covers partnering from the perspective of success attributes (Black et al., 2000) , performance outcomes or benefits (Gransberg et al., 1999) , and emergent collaborative working environments (Jacobsson & Roth, 2014) . Various approaches are offered in the literature to implement partnering based on both informal aspects (e. g., philosophically focusing on trust, good - will, commitment) and formal tools (e.g., procedures and processes such as workshops, scorecards, etc.) (Crespin - Mazet, Havenvid, & Linne, 2015) . Partnering implementation is best achieved when properly aligned with project risk fac tors (Eriksson, 2010) . Crespin - Mazet, Havenvid, and Linne (2015) , in their case study, found relational congruence between the project team as integral in decisions to pursue partnering. This supplants traditional notions which assert complexity, uncertainty, and risks as the most prominent factors informing this decision (Eriksson, 2010) . What this suggest s is the need to facilitate longer - term relationships, shifting from project partnering to strategic partnering. R elationships become more solidified by trust, commitment, and commonality of goals due to increased interactions instigated during project partnering (Crespin - Mazet et al., 2015) . An underlying question is whether these gains are achievable by managing the behavioral attributes on a single partnered project. 25 The success of project partnering is typified within industry reporting (International Partnering Institute [IPI], 2017 ), yet has not been fully researched in a manner that permits robust evidence. These risk factors are also critical when building collaborative project teams during the procurement process (Rahman & Kumaraswamy, 2005) . For example, project partnering is well - suited for routine projects lacking size and complexity and, therefore, are identif ied as low in risk (Gransberg & Scheepbouwer, 2015) . Risk factors will also interact with collaborative practices which positively impacts the relational behaviors of the project team (Suprapto et al., 2015) . Thus , it is important to adequately access this risk and ensure the level of partnering practices are fitting for the project (Eriksson, 2010) . Based on this d iscussion, the following proposition is given: Proposition 1: Project risk factors and the level of collaborative project delivery practices in partnered - projects are positively related ; According to the IPI ( 2017b) , partnering should be implemented based on certain perceived risk factors such as project value, complexity, political significance, and the experience of the team. Risk related factors can be disaggregated into contractual and pr oject risk which occur across the various phases of construction project delivery. According to McKim ( 2005) , project risks are often underestimated b y construction professionals while contract related risks are ignored altogether. This study takes in to account the external types of risk using t he matrix in Figure 2 - 2 which are then codified into partnering agreements and charters. The figure illustrate s how various risks levels are determined, identifies risk factors, establishes the desired level of engagement, reports expected benefits and costs to implement, and recommended partnering elements for success. 26 When followed, the criteria guide the level of partnering implemented for a project. These levels range from one to five with one being low risk and five the highest amount of risk. The project risk factors are scored based on the following: 1. Project value or cost a. Micro/Short duration ($0 - $5M) b. Sma ll ($5M - $10M) c. Medium ($10M - $25M) d. Large ($25M - $250M) e. Very Large/Mega ($250M - $500M) 2. The d egree of complexity (e.g., short timelines, schedule constraints, uncommon materials or designs) a. Standard complexity b. Moderated complexity c. Increased complexity d. Hi gh complexity e. Highly technical/complex design and construction 27 Figure 2 - 2 : IPI Vertical Construction Project Partnering Scalability matrix for collaborative partnering. 28 3. Political significance (e.g., place of importance, client size, organizational images at stake, strategic project) a. Unlikely b. Likely c. Probable d. High visibility/oversight 4. The d egree of relationships (i.e., previous working experiences with other stakeholders, contractors, subcontractors, agencies, constructio n manager, etc.) a. Established relationships b. Newly formed relationships c. New project relationships or High potential for conflict based on past experiences Risk evaluations using the criteria above becomes critical to assist project teams, not only in their d ecision to partner but to what extent. In fact, Gransberg et al. (1999) specifically point out that projects larger than $5M are ideal to incorporate formalized partnering practices. This with those of the project as opposed to individual motives. The advantages of partnering are generally proffered by examining these partnering characteristics. Project partnering, as defined in this study, is the pro cess by which partnering processes and tools are incorporated either formally or informally into project delivery. Many partnering characteristics are shown as beneficial to drive th e success of projects by way of aligning project teams. Process feedback c an be general or specific, which is typically found in project partnering strategies. For instance, a survey of 264 construction professionals suggests top 29 management commitment and belief in an integrated process is a highly important characteristic of pa rtnering (Ospina - Alvarado et al., 2016) . Meanwhile, they found training to be of medium importance and rated facilitators as neutral. Despite this, man y of the partnering characteristics found in the literature are shown to drive the success of team and project performance. Another example is that successful partnering can be attributed to the use of facilitators (E. W. L. Cheng & Li, 2002) . When formalized in partnering contracts, the role of a professional neutral facilitator is to guide partnering workshops and meetings, lead efforts to collect partneri ng scorecards and communicate feedback from various metrics. The facilitator is also responsible for promoting cooperation between parties while remaining neutral on project specific content. They can also guide the team in formulating their partnering cha rter. A partnering charter codifies the mission statement, goals and objectives, and guiding principles for the project (Larson, 1997) . Ultimately, a facilitator keeps the project team focused on achieving their defined par tnering goals and objectives. Project goals are generally broad and encompass ideas such as maintaining a safe project, meeting customer satisfaction, minimizing rework with a commitment to quality, and delivering the project on - time and under budgeted c osts. Project goals are solidified through mission and value statements within project charters. This continuously reminds project teams of the holistic commitments to which they have agreed upon. In contrast, project objectives should be measurable, attai nable, results - oriented, and time - bound. Project objectives may include metrics on cost and schedule growth, claims and disputes, or number of change orders related to errors and omissions associated with the design or construction (Gransberg et al., 30 1999) . According to Earley (1990), goals influence performance outcomes and stim ulate self - confidence, effort, and task strategies when feedback is present. In project partnering, towards common project objectives. The discussion above result s in the second proposition: Proposition 2: perceptions in partnered - projects are positively related Garnering the support and sponsorship of top management has also been identified as critical in partnering implementation (Cheng & Li, 2002) . This assuages project team member concerns that manp ower, resources, finances, and adequate time ha ve been allocated towards partnering processes from the home office. Senior managers engaged in the partnering process have an opportunity to help identify and assess project risks, create proactive plans to m anage and control risks, and establish decision - making processes to deal with them as they arise. Moreover, top management should empower all field level team members to make decisions in the interest of the project when problem - solving is required (Ng et al., 2002) . These feedback processes send motivational and directional cues to individuals encouraging them to adjust their efforts or come up with new performance strategie s (Earley, 1990) . Other partnering tools and processes used during project delivery also encourage collaboration and point to success. Some of these are attracting high levels of involvement from all stakeholders of the project (i.e., owner, designer, contractor, subcontractors, and end - users) and t he development of a dispute resolution ladder which includes a facilitator when appropriate. The objective of a dispute or issue resolution ladder is to resolve issues quickly and 31 at the lowest level possible rather than letting problems escalate. When the y do escalate , the likelihood of cost and schedule impacts are increased. Achieving full engagement from stakeholders helps to continuous realign the vision and goals of the project with those of the project team. Additional elements of partnering such as project scorecards for benchmarking project goals, holding partnering meetings, workshops to partnering concepts offer feedback clues to the project team (Eriksson, 2010; Ng et al., 2002) . Performance feedback made available through scorecards can help individua ls identify the need to adjust actions, however , does not always provide strategies to make adjustments. Thus, the role of professional facilitators becomes even more important. According to Cheng and Li (2002) , partnering processes are reactivated when stakeholders are fully vested in its benefits and trust others intentions with its use. A summarized list of partne ring practices is given below: Kick - off partnering workshop used to develop the partnering charter; Partnering charter that outlines: o Mutual goals and objectives; o Partnering maintenance and close - out process, partnering sessions and attendees, the frequen cy of meetings; and, o A clear dispute resolution plan mutually agreed upon by partnering participants; Partnering specifications; 32 Engagement of a professional neutral third - party partnering facilitator; Partnering training; Executive sponsorship demonstr ating top management commitment and support for the partnering process; Early involvement of key stakeholders in the decision - making process; Multi - tiered partnering (i.e., executive, project team, stakeholders); Subcontractor on - boarding/off - boarding wher e relevant parties participate in partnering sessions; Focused Action Strategic Teams (FAST) empowered for field - level decision - making as a means of timely issue resolution Monthly scorecards for continuous feedback on project team performance; and, Disp ute resolution ladders. All of these feedback processes and practices are closely related to other characteristics found during project delivery as reported by Sparkl ing, Mollaoglu, and Kirca, (2016) in their research synthesis of partnering literature. Research points out the broad characteristics of AEC partnering literature and potential links among these characteristics (Mollaoglu et al., 2015; Sparkling et al., 2016) . In their syntheses, 72 partnering studies are classified into several prominent categories using a meta - analytic review process. The work posits a clear taxonomy regarding partnering characteristics, 33 specifically categorizing them as drivers during partnered project delivery, project team characteristics , and performance outcomes among others. Using the framework in Figure 2 - 3 , a path to improved performance outcomes is proposed. Figure 2 - 3 : Partnered project delivery framework adopted from Sparkling et al. 2016 The path follows the three categories defined as 1) Drivers during partnered project delivery best practices followed during contractual, procurement, and par tnering practice related activities; 2) Project team characteristics Performance outcomes Improved project and organizational performance attribute d to partnering implementation. Project performance benefits include improved cost, schedule, and quality/safety performance, along with strategies to manage conflict. Organizational performance related outcomes are those attributes such as creating and ma intaining lasting relationships with partnering teams or working to improve organizational reputation in the industry. Other performance outcomes are those of individuals' and project teams'. Individual performance outcomes are those internal 34 attitudes an d beliefs held by individuals while team performance is concerned with developing team cohesion and trust. Sohani (2016) implemented this framework via case study and found that partnered projects followed many of the drivers asserted as practices. In th is analysis, partnering practices were connected to increased performance outcomes teams experienced during project delivery. This framework is further supported by a research synthesis of AEC partnering literature ( Sparkling et al., 2016). Based on their analysis, the top elements of the framework categories as seen in the literature are reported. Using the results of cumulative research or meta - analytic techniques in AEC research is indirectly validated as a consequence of this study ( Sparkling et al., 2016) . Sparkling et al. (2016) used meta - analytic procedures to their advantage by reviewing AEC partnering literature over several decades to spot trends, gaps, and direction for future research. For instance, in their synthesis, t hey not only provide a taxonomy of the literature but offer clues to connect the research streams. One particular stream has dominated the literature focusing on drivers during partnered project delivery (i.e., practices, contractual, and procurement) and the impact on partnering success. The literature points out many critical feedback attributes which are implemented in practice such as partnering workshops, project scorecards, clearly defined goals and objectives, etc. (Chan et al., 2004; Deborah Hughes, Williams, & Ren, 2012) . The top practice elements as a result of their synt hesis are shown in Table 2 - 1. Despite the importance of formal tools, contractual and procurement related elements as purported by Sparkling et al. (2016) are also useful to the effectuate project team performance. 35 Table 2 - 1 : Analysis of Partnered - Project Delivery Framework Category: Drivers during delivery ( Sparkling et al., 2016) Top Drivers During Delivery Number of times elements investigated in literature Total number of studies investigating each subcategory (% of Total # of studies or 73) Practice Elements (D1) 51 (70%) Implementing partnering workshops for project teams 22 Properly communicating mutual goals and objectives 14 Benchmarking and monitoring partnering process (e.g., Project surveys) 12 Es tablishing clear and compatible goals for project teams 11 Using team building sessions across organizations 11 Enabling free flow of information across organizational boundaries 10 Engaging a neutral/third - party partnering facilitator to guide the p rocess 9 Using integrated technology systems that encourage collaboration 7 Establishing well - defined roles and lines of responsibility 5 Implementing an effective problem - solving process 5 Contractual Elements (D2) 40 (55%) The use of incent ives/fees /risk - reward/ or gainshare - painshare agreements 14 Explicit contracting language and/or form of contract 12 Conflict identification and resolution strategy established in contracts 9 Shared equity arrangements indicated in contracts 7 U sing a partnering agreement 6 Equal power/empowerment afforded to all project teams 6 Procurement Elements (D3) 25 (34%) Early involvement of key participants (e.g., designer / contractor / subcontractors) 13 Selection of parties with partner ing experience 8 Considering previous work experience with other members 6 Selection of parties with technical expertise 6 This study asserts an emerging framework underlies the frequently reported and researched practice elements. It is ant icipated that practice elements can be further isolated. 36 For example, forming joint project charters and including mutual goals and objectives help project teams in goal alignment. Meanwhile, the structure is brought in goal alignment by way of clear dispu te resolution/ problem - solving processes are followed, field level decision - making is encouraged, and professional facilitators are used to guide project teams. Feedback mechanisms such as partnering training/team - building sessions, partnering workshops, a nd scorecards are used as follow up processes to reinforce goal alignment. This study demonstrate s that project risk factors and collaborative project delivery practices (i.e., partnering drivers or characteristics) influence goal alignment in AEC project teams. Case study evidence supports the original partnered - project delivery framework and further exclaims the importance of both contractual and procurement drivers, in addition to practices (Sohani, 2016) . The case study examined partnered - project data consisting of meeting minutes and project scorecards for an airport project with high - risk factors. Multiple feedback processes were incorporated into the project, some were using a neutral th ird - party facilitator, holding monthly partnering sessions, colocation of project teams, and formation of an issue resolution ladder. According to the partnered - project framework, these are all considered drivers during project delivery and offer feedback to the project team. Based on the analysis, the level of feedback was positively relate d to the performance of the project team and, subsequently, project performance. In particular, early involvement of contractor and subcontractors (i.e., procurement ele ment) in the design process was shown to be positively related to team attributes such as establishing mutual trust and encouraging team commitment. Thus, our limited understanding of how softer metrics can be used to monitor 37 project performance is emergin g yet requires additional validation and testing from an alternative perspective and leads to the third proposition. Proposition 3: - projects are positively related Sparkling et a l. (2016) purport several other useful elements that are beneficial to understand how team characteristics and performance outcomes are investigated in AEC literature. Table 2 - 2 illustrates the top project team level performance elements. Performance outco mes are shown i n Table 2 - 3. Table 2 - 2 : Analysis of Partnered - Project Delivery Framework Category: Team Characteristics ( Sparkling et al., 2016) Top Project Team Characteristics Number of times elements investigated in literat ure Total number of studies investigating each subcategory (% of Total # of studies or 73) Project Team Level Elements (T1) 40 (57%) Establishing mutual trust within project teams 24 Staying committed to the project teams goals and objective s 15 Using integrated project teams 11 Maintaining commitment to entire partnering process 7 Committed to win /win attitudes 5 Developing mutual interests among the project team members 5 Individual - level Elements (T2) 11 (15%) Individual s able to maintain positive attitudes 6 Working with integrity during the process 2 Maintaining enthusiasm in the partnering process 2 38 Table 2 - 3 : Analysis of Partnered - Project Delivery Framework Category: Project Performance Outcomes ( Sparkling et al., 2016) Top Project Performance Outcomes Number of times elements investigated in literature Total number of studies investigating each subcategory (% of Total # of studies or 73) Cost Performance Elements (P1) 31 (42%) Meeting costs targets for the project 10 Improved cost savings during project delivery 7 Reduced additional expenses due to changes and other concerns during project delivery 6 Claims cost are reduced as a percent of the original cost 5 Quality / Safety Performance Elements (P2) 27 (37%) Improved the quality of the project 14 Increased client and/or end - user satisfaction 8 Improved safety performance for the project 8 Reduced environmental issues and/or complaints 5 Reduce wasted work or re - work 5 Improved overall project design 5 Schedule Performance Elements (P3) 29 (40%) Projects are able to meet scheduling targets 12 Reduced overall time in delivering the project 6 Faster project delivery 6 Conflict Resolution Elements (P4) 21 (29%) Reduced disputes among project teams 10 Improved resolution of claims and helps to avoid issue escalations 7 Reduced litigation resulting from unresolved conflicts 7 The above illustrates useful benefits offered from meta - analytic approaches and techniques, yet AEC literature is short on its implementation. A clear framework and taxonomy 39 are developed using a research synthesis that provides a useful aggregation of theory. Meanwhile, a priori evi dence alludes to its advantages and clearly implicates team attributes as crucial to understanding project partnering (A. E. Sparkling et al., 2016) . Utilizing this as a guide, an emerging framework is developed for this study to un derstand collaborative project delivery approaches in the context of partnering. Team researchers continuously seek to explain and understand team effectiveness and performance. This is due to the fact teams ar e common components of organizational two or more people who interact, dynamically, interdependently, and adaptively toward a common and valued goal/objective/missi on, who have each been assigned specific roles or functions to perform, and who have a limited life - (Salas, Dickinson, Converse, & Tannenbaum, 1992) . Teams and groups are often used interchangeably, thus follows in this study. An introduction to relational governance theory is described in the next section followed by teams of teams and fee dback theories. The notion of relational governance emerged from those seeking to understand the nuances of formal contracting and relational contracting (Carson, Madhok, & Wu, 2006; Poppo & Zenger, 2002) . The se minal work of Williamson ( 1979) on transaction cost economics ( TCE ) gav e rise to the phenomenon of opportunism and governance structures illustrating how opportunistic influences persist in formal contracts. However, relational contracting literature is critical of TCE 40 as missing out on the social and relational embeddedness in the exchange (Granovetter, 1985) . An emergent relational governance structure develops as organizations establish shared values and agreed - upon processes as part of interorganiza tional relationships (Zaheer & Venkatraman, 1995) . Within this governance stru cture, trust is sought via social processes which encourage flexibility, solidarity, and information exchange (Poppo & Zenger, 2002) . Unforeseeable events are overc ome in the social process by way of flexibility. According to Poppo and Zenger ( 2002) , flexibility facilitates adaptation when untimely events happen during the exchange. Meanwhile, they assert solidarity is a means to create interorganizational problem - solving and commitments as organizations take actions and adjust behaviors. B ehaviors such as mutuality and cooperation are further espoused through the use of information sharing regarding goals and objectives. These social processes embedded in relational exchanges safeguard against transaction hazards such as ambiguity, volatility, uncertainty, and opportunism. Moreover, th e set - up costs are lower than formal contracts which are mor e complex and inherently signal distrust (Carson et al., 2006) . The principle component of relational governance is trust which is the extent to which fair negotiations are expected, commitments sustained, and belief that others parties will take a ction to fulfill future obligations (Claro, Hagelaar, & Omta, 2003) . Trust plays a vital role in the economic exchange as a means to curtail asset specificity concerns. As interorganizational relationships develop and interactions reoccur over time, social elements and relational norms become more salient (Z aheer & Venkatraman, 1995) . While interpersonal trust is important, interorganizational trust attenuates negotiated costs because agreements are reached sooner and parties are willing to reach a quick consensus (Zaheer, McEvily, & Perrone, 1998) . In part, 41 because intero rganizational trust in conditions of relational governance structure and processes becomes institutionalized and does not solely rely on the boundary spanners [i.e., team members with external ties to access resources for the group (Katz, Lazer, Arrow, & Contractor, 2004) ]. Collaborative project delivery methodologies, such as partnering, include both formal contracting and informal/relational contract ing governance forms. This is further revealed by L u et al. ( 2015) who assert contractual and relational gove rnance are complementary attributes with positive effects on project performance. For example, traditional contracting terms are specified with equitable risk allocations and followed by parties working on partnered - projects while relational governance is effectuated when benchmarking metrics, partnering workshops, and conflict resolution strategies are not memorialized in the contract. This occurs s pecifically when s formed seeking goal congruence toward explicit project goals and objectives established in partnering charters. This mentality generally reduces differences between individual and organizational goals (Ouchi, 1980) . Collaborative projects may rely on hybrid approaches drawing from both informal and formal mechanisms in negotiating the transactions involved in project delivery (Chen & Manley, 2014) . The primary purpose behind these governance mechanisms is to improve performance outcomes while harnessing goal alignment objectives (Chen & Manley, 2014) . These processes develop stronger team relationships around common goals and enhance performance outcomes (P. Davis & Love, 2011) . Th is discussion leads to the hypothesis below: 42 Hypothesis 1 : Individual performance in partnered - projects is positively goal alignment perception . T he notion of teams of teams is pervasive in practice if one takes a moment to assess the organizational landscape. Construction projects are ripe with teams of teams due to the way projects are delive red. As an example, a construction company may have seve ral projects running concurrently each having a Project Director, Project Manager, Superintendent, Project Engineer, Safety Manager, and administrators. This comprises a team which is then joined with other similar team structures to deliver a construction project. Teams, like these, must work efficiently and effectively to accomplish complex tasks or goals. For example, some partnered projects use focused action strategic teams (FAST) to facilitate quick and timely exchanges of ideas and information to re solve issues. These smaller subgroups, or teams of teams, are typically comprised of responsible work parties and are often divided by specific work scopes such as the quality control and assurance team. Meanwhile, outputs whether tangible or intangible ar e fruits from processes and are used to give feedback on how well performance objectives are met (Ilgen & Moore, 1987; Nadler & Tushman, 1980; Nadler, Mirvis, & Cammann, 1976) . Performance conditions are concerned with quality, quantity, or other measurable outputs that are obtainable from processes. This information is used to provide performance feedback to individuals or teams. Several factors are at play when teams work towards collective goals. The performance of the team can be constrained by team composition, work structure, task characteristics, a nd 43 shared cognition among others (Salas, Cooke, & Rosen, 2008) . It is, therefore, important to und erstand these various factors and how they relate to construction project teams. Team composition largely consists of underlying attributes such as personality, cognitive ability, motivation, and cultural factors (Salas et al., 2008) . One can only imagine the hosts of issues personalities may impose on team performance. For example, team performance is positively related to te am conscientiousness, agreeableness, extraversion, emotional stability, and openness to experience (Mathieu, Maynard, Rapp, & Gilson, 2008) . Therefore, team performance can be restricted when team members lack some of these key attributes. Cognitive ability, however, has v aried implications regarding team performance. It is positively related to learning performance during task execution when the workload is evenly distributed and when task - related knowledge is desired over longer time frames. Conversely, cognitive ability is less credible when the team structure is not appropriate for the tasks (Mathieu et al., 2008) . Team norms, communication structure, and even work assignments can also affect team performance. These factors underlie the work structure under which teams operate (Salas et al., 2008) . An effective communication structure is imperative to a successful project team. One type of communication structure is that of a transactive memory system (TMS). A TMS, developed by Wegner (1987) , is a cooperative system to store, retrieve, and communicate information for person to person (Lewis, 2003) . This communication structure allo ws a team to efficiently coordinate, communicate, and access specialized knowledge of individuals. In the AEC industry, the implications of team cognitive processes and emergent states are easily evinced. Project teams come together for the common purpo se to complete a project 44 based on owner requirements communicated and executed by many different organizational units. During the traditional delivery processes (e.g., design - bid - build (DBB),) effective teams struggle to establish common goals and objectiv es, freely share information, remain open to input from all team members, and work seamlessly across organizational boundaries (Baiden, Price, & Da inty, 2006) . It is a necessity for multidisciplinary teams to integrate knowledge from different domains, yet, is hindered due to misaligned cognitive processes which are seemingly complex (Kotlarsky, van den Hooff, & Houtman, 2015) . Although this working system has persisted for many years, the construction industry lacks effective teams (Baiden & Price, 2011) . The emergent cognitive process within FAST teams is team mental models (TMM) and transactive memory systems (TMS). A TMM can be thought of as convergent knowledge representation regarding information held in common by a team (Mohammed, Ferzandi, & Hamilton, 2010) . It is likely executive teams experience a TMM whereby everyone has a shared understanding of project goals and obje ctives. Furthermore, the TMM can help them efficiently access information necessary to execute tasks such as deciding on a particular heating, ventilation, and air conditioning (HVAC) system that satisfies both design intent and stakeholder requirements. A s a project moves into construction, a TMS develops within the core project team. It is this distributed knowledge structure that allows the core team to rely on the specialized expertise of all team members and to coordinate information flow. An example TMS is exhibited during a meeting to review an electrical equipment submittal which includes the party is anticipated to provide information in the review process base d on their specialized area 45 of knowledge. Design changes or inadequate designs are marginalized during these team submittal review processes where expectations are coordinated, clearly communicated, well - defined, and received by everyone. Moreover, in thes e instances, all project stakeholders are able to provide feedback and insights based on their knowledge, skills, abilities, experience, and other tacit knowledge. Feedback serves multiple purposes ranging from error - correcti on to identifying problems and providing clarity for goals (Nadler, 1979) . For team members to allocate cognitive and behavioral resources to tasks, DeShon et al., ( 2004) suggest feedback should be specific to individuals or their subsequent teams, but not both. In this self - regulatory process, individuals compare feedback information against individual goals and team goals. Then, a behavioral choice is made to prioritize the level of effort and strategy expende d towards these goals ultimately affecting performance outcomes. To achieve benefits from feedback individuals working in teams must receive, process, and react to the feedback mechanism. This is often problematic when the feedback frequency is irregular or i nfrequent; however, learning from feedback is contingent upon project risk levels, evolves over the project lifecycle , and across projects (Manley & Chen, 2015) . Thus, in AEC partnered projects feedback is provided systematically based on key project factors such as cost, size, complexity, and duration (Gransberg et al., 1999) . These feedback monitors are set by the te am early when project goals are identified and whe n decisions are made to use a neutral third - party facilitator. 46 Feedback intervention theories are not unified on the effect feedback has on performance. According to Kluger and DeNisi (1996) , the rel ationship between feedback and behavior is complex and often contradictory in its interpretation. The feedback body of literature, also known as feedback intervention, commonly asserts that feedback intervention improves performance despite varied results over its nearly 100 years of research. In particul ar, some researchers have found feedback intervention has no effect, negative effects, or debilitates performance altogether. To address this dilemma, Kluger and DeNisi (1996) conducted a meta - analysis in order to document the state of the literature, prop erly unify feedback intervention theory, and to integrate existing theories into a new feedback intervention (FI) theory. F eedback intervention theory merges components and ideas from the following theories into a singular theory: control theory (Annett, 1969), goal setting theory (Locke & Latham, 1990), action theory (Frese & Zapf, 1994), action identification theory (Vallacher & Wegner, 1987), multiple - social cognition theory (Bandur a, 1991), and portions of learned helplessness theory (Mikulincer, 1994). In doing so, they offer five basic arguments for feedback intervention theory: 1) Behavior is regulated by comparisons between feedback to goals and standards, 2) goals or standards are hierarchically organized, 3) feedback - standard gaps that receive attention are active in behavior regulation processes because attention is limited, 4) generally attention is directed towards moderate levels in the hierarchy, and 5) feedback interventi ons shift the focus of attention and impacts behavior. Based on these conjoined theories and assumptions, they assert FI induces task motivation, task learning, and meta - task processes. 47 Whether investigating feedback from the perspective of goal - setting or control theory, the behavior is directed towards goals or standards. Feedback (i.e., intervention or not) allows people to evaluate their performance with respect to goals or standards. This comparative inspection results in a feedback sign relative to the goal (e.g., positive or negative sign in relation to the discrepancy between performance and goal). Several options are available to individuals in response to the feedback - standard discrepancy. One can eliminate the discrepancy, according to control t heory, by changing behavior so future feedback is adjusted, change goal or standard to align with present feedback, refuse feedback, or avoid the situation (i.e., physically or mentally). Similarly, goal setting theory maintains people may respond to feedb ack - standard discrepancies by striving to achieve the goal, change the goal, refuse feedback, or abandon commitments one has made to goals. Available literature and recent research has examined effective teams in the context of AEC, where well - defined con cerns with communication, trust, and collaboration are illuminated (Chan, Chan, & Ho, 2003; Cheu ng, Ng, Wong, & Suen, 2003; Dewulf & Kadefors, 2012; Larson, 1997; Naoum, 2003; Xue, Shen, & Ren, 2010) . Research into AEC partnering purports workshops consisting of feedback or partnering health indexing (Puddicombe, 1997) offer direct information as to the current effectiveness of the arrangement (Cheung, Suen, & Cheung, 2003; Mollaoglu et al., 2015) . With this in mind, it is important to understand how underlying team knowledge pr ocessing systems (TMS) interact with the l evel of commitment and self - efficacy of individuals when seeking to achieve team goals. This leads to the second hypothesis for this study: 48 Hypothesis 2 : MS moderates the relationship between individuals alignment perceptions and indiv idual/team performance in partnered - projects. Lewis (2003) systems was spawned by (Wegner, 1987) , who first observed how groups in close relationships it is easil y retrieved from others when needed henc e; a transactive memory system is greater than the sum of its parts or individual memories (Wegner, Erber, & Raymond, 1991) . A TMS entails a shared division of c ognitive labor regarding encoding, storing, retrieving, and communication information from different subject experts (Ho llingshead, 1998b; K. Lewis & Herndon, 2011) . This system efficiently ensures new information entering a group is properly allocated to the correct member who is responsible for it. This information is added to the pertinent knowledge already held by the member and is available to be quickly retrie ved, communicated, and integrated with the tasks related knowledge when needed by the group. Kyle Lewis (2003) put forward three categories to discern a TMS as specialization (i.e., the differentiated structure of the reliability of other members knowledge), and coordination (i.e., effective and well - orchestrated knowledge processing system). These categories are used to aggregate ns to create a team score of TMS. This measurement offers a unique approach to diagnose integration within teams. 49 Theory and research suggest TMS facilitates quick and coordinated access to specialized expertise, thereby improving group performance. The performance attributed to TMS is explained by the unique knowledge structure that develops. More importantly, this depository of knowledge is coded and stored in a systematic process that allows for easy retrieval or elicitation from group members. An effe ctive TMS is further enhanced by the dynamic interplay as teams communicate, interact, and execute tasks in groups (Lewis & Herndon, 2011) . As a result of this guidanc e, this study contends team theory is useful to explore the association between partnered project feedback during collaborative project delivery and performance outcomes. Forming teams that work well together is commonly a scribed to the success of projects. From an organizational perspective, Hoegl and Gemuenden (2001) establish performance measures of team collaboration using commu nication, coordination, balanced member contributions, mutual support, effort, and cohesion as measurement indicators. In their opinion, perceptions of team performance are directly related to teamwork qualities and those of team leaders and managers. Simi larly, Suprapto et al. (2015) purport how key teamwork qualities, as these, serve as mediators in regards to relational attitudes, collaborative practices, and joint capability with project performance in construction projects. However, the effectiveness i s constrained when collaborative practices become formalized and results are taken for granted (Bresnen & Marshall, 2002) . In spite of this, it is becoming more apparent that collaborative working arrangements, such as par tnering, hold clear implications to help increase the efficacy of project teams. In particular, those working across organizational boundaries. 50 Construction projects are inherently filled with risks and commonly rely on the contracts to h elp spread the risks across its participants. It is challenging to account for all unforeseeable risks ahead of commencing work, thus contract conditions have been traditionally used. Traditional contracts work against joint risks management necessary post contract stage when challenges are encountered extending beyond the contract language (Rahman & Kumaraswamy, 2004) . Thus, collaborat ive working arrangements surfaced to encourage team integration and to deal with decades filled with adversarial mentalities among fragmented construction processes (Lahdenperä, 2012) . The concept of integration has taken on several meanings in the AEC industry. The term construction (Betts, Fischer, & Koskela, 1995) . The premise of computer integrated construction was the implementation of technologies which could facilitate frequent data and knowledge sharing among proje ct participants (Teicholz & Fischer, 1994) . Through this, they explicated several goals for computer integrated construction to achieve its aim such as the rapid deployment of high - quality designs, quick and cost - effective construction with the ability to deliver data and models to the end users of the facility. Integration, according to Teicholz and Fischer (1994), begins with integrated computer applications that allow for concurrent design and construction. The use of integration continued to progress from computer integrated construction to information between differen t proj (Baiden, Price, 51 & Dainty, 2003) . Integration has also different individual or organizational goals, cultures, and needs into a single cohesive unit (Jaafari & Manivong, 1999) . Although many different definitions persist, its purpose remains the same being to align project teams into cohesive and collaborative units with common objectives. A fully i ntegrated team, according to Baiden et al. (2003), includes the client/owner team and is considered a team of teams with clearly defined skills and professional roles necessary to satisfy project objectives. This integrated team is best achieved by incorpo rating contractually binding agreements such as dispute resolution ladders to help resolve conflicts into collaborative contracting arrangements (Gransberg & Scheepbouwer, 2015) . Recent literature on team integration and col laboration demonstrate an emerging area of research within construction project teams (Comu et al., 2013; B. Franz et al., 2016) . Team integration theory is well - documented in other areas such as organizations, sports, military, and academics (Chiocchio & Essiembre, 2009) . Yet, construction project teams are ripe with informatio n that can inform theory based on common understandings of social dynamics. Thus, researchers are attempting to better understand the role of integration, group cohesion, and transactive memory systems as they relate to project performance occurring during collaborative working approaches (Comu et al., 2013; Franz et al ., 2016) . Franz et al. (2016), explores a sample of 204 completed projects following different delivery methods to compare cost , schedule, and quality performance. They assert the literature is disparate, or according to Davis (1971) not well - organized in which project delivery methods are most appropriate for team integration and the extent integration affects 52 performance outcomes. As with this study, they insist the how and why are generally not explored when thinking about this relationship. Thus, the data in their study is analyzed to explore these relationships against different delivery methods. Two distinct theories are a priori in their work, group cohesion, and team integration. In their opinion, team integration consists of interfirm interactions and shared culture that is developed during project delivery. Meanwhile, group cohesion is the point at which a new identity is established and the team members join togethe r in a common culture committed to project goals while also trusting, respecting, and clearly communicating as a team. Using these two theories, the researchers investigate mediators between project delivery and performance. Franz et al. (2016) determined key components of project delivery methods influence team integration and group cohesion. Specifically, the timing of involvement from team members factors into whether team integration is achieved. Group cohesion is increased through cost transparency an d procurement processes that place less emphasis on price but, rather focus es on qualifications or relational attributes. According to Franz and Leicht (2016) , the literature is shifting in our understanding of project delivery a pproaches and how collaborative tools redefine traditional classifications. They separated established project delivery methods into latent classe s such as Class I (i.e., Design - Bid - Build [DBB]), Class II (i.e., Design - Bid - Build with early procurement), Cl ass III (i.e., Construction Management At - Risk [CMR]), Class IV (i.e., Design - Build [DB]), and Class V (i.e., Multiple DB and Integrated Project Delivery [IPD]). Thus, processes are equally important as the project delivery methods utilized for a project a nd influences team behaviors. 53 Comu et al. (2013) outlined another new perspective on collaborative working arrangements and a relationship to transactive memory systems (TMS). In their opinion, firms operating globally across geographical and other bounda ries require unique strategies to overcome ambiguity. Virtual teams are also seeking to increase collaboration, despite their geographical distances as they are believed to inherently be more complex than traditional project networks. Thus, the researchers insist both technology and relational intervention (i.e., facilitators) affect project performance. Comu et al. (2013) assert in their evaluation of global virtual project networks (GVPN) that facilitators affect performance outcomes so long as they d eliver process feedback as opposed to project content. Stated differently, process facilitators engaged in content related discussions impact GVPNs performance working in virtual collaborative workspaces . According to Comu et al. (2013) , the development of a cohesive TMS is related to collaborati ve effectiveness in the project network. Using data from four simulated global engineering project networks, C omu et al. (2013) examine the development of cohesive subgroups. A strong cohesive network is posited to include higher frequency and duration of interactions during group collaboration, thus affecting performance. They compare ratios of nodes with ties wi thin the subgroup against that of nodes outside the subgroup. These networks are also separated by facilitator versus non - facilitated groups. Interestingly, the findings are insightful in that facilitators did not help maintain nor support cohesive collabo ration in TMS. Additional support for this findings is, therefore, available for future investigation. The second finding as a result of their work is that TMS development and cohesiveness are positively associated with non - facilitated networks. 54 Contrastin gly, this study anticipates feedback via facilitation is positively associated with TMS development and/or cohesiveness, however, it is based on process facilitation rather than content facilitation. The two aforementioned studies demonstrate a need to understand the key elements underlying collaborative working from a team theory perspective. The key components of collaborative working are 1) commitment; 2) cooperation and communication; 3) trust; 4) common goals and objectives; and a 5) win - win philos ophy (Yeung et al., 2012) . Commitment refers to a shared sense of ownership to project goals, objectives, and successful outcomes for all project team members. As can be expected, these essential elements are interrelated such that open and honest communication inspires trusting relationships to form within project te ams. Meanwhile, establishing common goals and objectives helps to direct attention to controlling transaction costs associated with a project (Walker & Chau, 1999). A te am focus ed on common goals and objectives reinforces other attributes beneficial to inc reased collaboration and project success. Other supporting elements found with in AEC literature are 1) agree d - upon problem resolution methods; 2) continuous improvement strategies; 3) facilitated workshops; 4) equitable risk - reward structure; 5) declared statement of common objectives; 6) agreed - upon gain - sharing/pain - sharing or bonus incentive program; and, 7) formal contract to which binds all parties to the agreements (Yeung et al., 2012) . These core elements and supporting attributes are typically found in contracting approaches bent on developing integrated proj ect teams such as IPD, strategic or project alliancing, and project partnering. Yet, it is the shared knowledge and information processing system that helps the teams pick up efficiencies and 55 increase their performance. Interestingly, outputs from collabor ative processes will impact team effectiveness. For instance, consistent engagement from top management, integrity monitoring using team scorecard surveys, team empowerment, and feedback in the form of acknowledgment and/or celebrations are processes that are likely to influence individual behaviors and build trust (Pishdad - bozorgi & Beliveau, 2016) . Team regulatory and goal alignment processes determine the amount of behavioral resources one will allocate to team performance goals (Kozlowski & Ilgen, 2006) . The use of scorecards serve s as early warning indicators to the project team. This information can be used to provide genuine acknowledgments to thank team members for their contributions . Moreov er, can affect team goal alignment and motivation towards team performance outcomes. This understanding asserts a bottom - up approach to increasing team effectiveness, rather than top - down through contractual project delivery methods. Collaboration betw een different organizations is critical to accomplish common goals and factors heavily into performance outcomes (Dietrich et al., 2010) . According to Dietrich et al. (2010), there are five high - quality characteristics demonstrated in collaborative projects being: communication, c oordination, mutual support, aligned efforts, and cohesion. These characteristics are generally present among project teams when agreed - upon goals are established, clear and open conflict resolution strategies are used, and effective communication systems are employed. They assert, other characteristics are joint problem solving, trust, and goal congruence are all present within collaborative project teams. Thus, these teams may adjust their behavior in the relational exchanges with other organizations base d on the 56 those of the team (Stephen & Coote, 2007) . This occurs when formal and explicit language is not included in contracts. In fact, Stephen and Coote (2007) arg ue that relational behaviors are best aligned with goals when supportive leadership is involved. Generally, team effectiveness and integration hav e been attributed to specific project delivery methods (Mollaoglu - Korkmaz, Swarup, & Riley, 2013) . In contrast, Franz and Leicht (2016) offer an alternative taxonomy of project delivery methods which e xplains how they share common characteristics and can be situated within one of five emerging categories. These characteristics a re used to reorganize delivery types based on the timing of contractor involvement in design stages, procurement strategy, contractual arrangements with the owner, contractor selection and award criteria, and payment terms establish the various delivery ap proaches. With this in mind, construction project team integration and cohesion are clearly influenced by these characteristics (Franz, Leicht, Molenaar, & Messner, 2010) . For example, early involvement of contracto rs in the schematic design phase helps integrate their knowledge performance on a project. This suggests that increased team integration and cohesion is not confined to specific project delivery methods, rather results from manipulated team behaviors effectuated by owner decisions. Another alternative explanation is wh ether collaborative practices adopted for projects are the right fit, especially considering how feedback levels and, subsequently, the collaborative experience of project teams may vary when these practices are followed. 57 The construction industry generally relies on relational/collaborative project delivery arrangements to facilitate better goal alignment between project teams and project objectives (Zuo, Chan, Zhao, Zillante, & Xia , 2013) . Project delivery approaches such as IPD and project partnering emerged in the US in 1998 and 2005, respectively (Lahdenperä, 2012) . These multi - party contracting practices seek to increase team integration and own er value while reducing waste and inefficiencies during project delivery (Lahdenperä, 2012; Pishdad - Bozorgi & Beliveau, 2016) . The underlying motivation behind collaborative pr oject delivery arrangements is to spread risks and rewards evenly among project stakeholders (i.e., owner, design, and contractor). Project delivery approaches such as IPD and project partnering clearly incorporate collaborative practices and processes th at affect organization and management strategies, contracts, project team communications and their behaviors. Pishdad - bozorgi and Beliveau (2016) separated the following processes, referred to as traits, into four categories: 1. Organizational and Management Strategies a. Early involvement of key participants b. Jointly developed project target criteria c. Collaborative decision making d. Intensified early planning and design e. Champion/facilitator f. Building Information Modeling (BIM) g. Lean construction 2. Con tract a. Multi - party contract 58 b. Relational contract c. Shared financial risk/reward based on project outcome d. Risk identified and accepted early e. Liability waivers between key participants 3. Communication a. Open communication b. Colocation c. Information sharing 4. Behaviora l a. Transparent financials (Open book accounting) b. Mutual respect, trust, and collaboration c. Pre - existing relationships between parties d. Lean culture Many of these processes and/or traits behold feedback signals that change individual and team behaviors durin g collaborative delivery. The success of both the team and project can suffer, as a result, when not completely aligned. Collaborative approaches such as IPD, Lean construction, and project partnering share many similarities which encourage collaboratio n and influence the extent to which goal alignment occurs within project teams. Although, several differences persist with project partnering such as the use of benchmarking, partnering workshops, and partnering sessions. These collaborative practices impl emented during partnered project delivery provide more performance related feedback as opposed to process feedback which is also cr itical to help regulate behaviors. 59 Goal alignment is engaged by involving key participants early on and jointly developing p roject goals which build collaboration and trust. In this manner, project team s begin to align their knowledge and intentions with those of the project, rather than working from individual silos (Pishdad - bozorgi & Beliveau, 2016) . From a contractual standpoint, holding multiparty agreements that incl ude shared risks and reward structures militate against traditional mindset towards innovation, especially when their success is contingent upon that of the entire team (Lahdenperä, 2012) . Lean construction is another example of collaborative working arrangements that help project teams overcome project complexities, uncertain conditions, and competing goals often found among dive rse teams (Matu rana, Alarcón, Gazmuri, & Vrsalovic, 2007) . According to The Lean Construction Institute (2016), there is a concerted effort within the construction industry to continuous ly improve, generate value, remove waste in processes, work on flow efficiency, and optimize the whole construction project delivery process . The underlying goals are to always maintain respect for people involved in the project. The motivation behind Lean construction is to minimize waste and increase performance using strategies such a s building information modeling (BIM), The Last Planner System (LPS), value stream mapping, target value design, and set based design (Smith, Mossman, & Emmitt, 2011) . BIM is an information - rich 3D modeling software tool that allows project teams to virtually build a construction project ahead of physical construction. Tools like BIM help to integrate design and construc tion teams delivering more value for owners. 60 in project - based environments (Smith et al., 2011) teams by holding people accountable to promises made to other members of the project team. It is this ability to keep commitments that create a feedback loop, increasing cohesio n and, ultimately, affects performance outcomes. Value stream mapping is a process to determine the current value - added stream of the project to identify and eliminate waste. The process optimizes the design as part of project delivery and helps to avoid reworking that typically occurs later on during construction. Target value design is used to get early stakeholder involvement in the design process focused on set - based de sign is a process whereby stakeholders and subject matter experts (SME) work to develop a various solution to product and production design problems and decide on a solution at the last responsible moment as a team . When used, set - based design permits conc urrent design options to develop from which the project stakeholders can choose from. Each of these Lean construction processes, according to Smith et al. (2011), inherently require intense communication and collaboration. In project partnering, another collaborative working arrangement, FAST teams are assembled to resolve change issues and problems occurring during construction processes. There may be executive teams (e.g., top management from owner, contr actor, and design companies), core teams (e.g., f ield project managers from respective organizations), or stakeholder teams (i.e., end - user groups both internal and external to the project) all of w hom participate in joint workshops (International Partnering Institute (IPI), 2016). Team cohesion begins t o take place during these ongoing workshops, although this may be limited during team 61 formation (Salas, Grossman, Hughes, & Coultas, 2015) . Therefore, the frequency of partneri ng workshops permits greater team cohesion as they move through the phases of team development. This dy namic process, according to Kozlowski and Ilgen (2006) , emerges as teams task driven problems. This active intervention and synergistic environment promotes collective planning, organizing, and controlling of the project goals and objectives. In the event changes do occur, teams are able to de - couple tasks to minimize project impacts. However, a larger group of people working to review and approve a submittal may experience social loafing (i.e., reduced effort and motivation while working collectively as opposed to individually) within the team (Lam, 2015) . When this happens the overall performance of the team can be affected, especially if team me mbers in the field fail to offer tacit knowledge based on their experience since they are closest to the actual work. Collaborative approaches like partnering aim to bridge certain dilemmas using tools such as formalized partnering charters which empower a ll team members to participate, remain open - minded, to ask questions and also commit to hav ing fun during the entire partnering process. The propositions presented below are used to investigate the partnered - pr oject level research questions qualitatively. Figure 1 - 3 shows the propositions for the relationship between project risk factors, collaborative project delivery practices, and project performance. Proposition 1: Project risk factors and the level of colla borative project delivery practices in partnered - projects are positively related ; 62 Proposition 2: perceptions in partnered - projects are positively related ; and, Proposition 3: Indivi - projects are positively related . Separately, Figure 2 - 4 shows the proposition for the asserted cross - level interaction in this study. The hypotheses intended to quantitatively test the theoretical model and relationships among the variables at the individual /team level of analyses are given below . These are specifically used to estimate the direct and moderating effects between goal alignmen t , TMS , and performance measured during partnered - project delivery . The hypotheses are shown in Figure 2 - 5 and described below. Goal Alignment Project Level Data / Qualitative Investigation P2 P3 P1 Collaborative Project Delivery Project Performance Risk Factors Figure 2 - 4 : Relationships and propositions between project risk factors, collaborative pro ject delivery practices, and project performance 63 Hypothesis 1 : Individual performance in partnered - projects is positively goal alignment perception . Hyp othesis 2 : MS moderates the relationship between individuals alignment perceptions and individual/team performance in partnered - projects. F igure 2 - 5 : The r elationship and h ypotheses between g oal alignm ent, transactive memory s ystems, and individual/team performance Collaboration is becoming a central idea and expectation within the construction industry even though discontinuities exist between projects. The characteristics and risk associated with the project offer clues to understand whether a well - coordinated knowledge and information systems (TMS) moderates the effects of goal alignment on performance outcomes. Based on the literature, a need exists to parse out the structural component of t eams that cause the variations in this link. The literature often alludes to a strong connection between collaborative project delivery approaches and improved project performance, yet frequently fails to critically inspect the behavioral aspects of teams. Thus, this study fills this gap by identifying how project Individual/Team Performance Goal Alignment H2 H1 TMS Individual Level Data / Quantitative Investigation 64 risk factors and collaborative project delivery practices influence these relationships . Moreover, this study firmly establis hes the relationships among goal alignment , TMS , and performance outcom es . The methodology followed to achieve the aims of this study are described in the next chapter. 65 CHAPTER 3 METHODOLOGY This chapter presents and describes the research methods used in this study. This study uses a mixed - methods approach to how do collaborative project delivery practices affect goal alignment and performance in AEC project teams attributes of co llaborative project delivery approaches which result in better performance. Focusing on partnered - projects as a type and subset of collaborative AEC project delivery approaches, this research examines multilevel data via multiple - case study evidence. A bri ef review of the research questions is presented, followed by the methodological steps required to address the research questions. Few of these studies, until recently, have examined how team related factors are related to project performa nces in the context of AEC project teams (Comu et al., 2013; B. Franz et al., 2016) . A gap exists in the AEC literature to advan ce this topic, in particular, the role goal alignment and TMS have in the project delivery process. Based on the literature, g oal alignment perception s among AEC project teams and cohesive transactive memory systems are important for individual , team, and project performance. Thus, this research intends to address the how goal alignment affect s performance in AEC project teams when collaborative project delivery practices are followed Project Level Questions : What collabora tive project delivery practices impact project goal alignment in partnered - projects? Do project risk factors impact collaborative project delivery 66 practices and, thus, goal alignment? What are the relationships between project risk factors, collaborative p roject delivery practices, and goal alignment in partnered - projects? Does goal alignment affect project performance? Individual - level Questions : Does goal alignment affect individual performance? If so, is this relationship moderated by TMS? In the AEC l iterature, a need exist to identify the underlying attributes of collaborative project delivery approaches which result in better performance outcomes . To respond, several goals and objectives are established and reiterated next. The primary aim of this research is to explore the relationships between project risk factors, collaborative project delivery practices, goal alignment, transactive memory systems (TMS), and performance outcomes in AEC project teams. Focusing on partn ered projects as a type and subset of collaborative AEC project delivery approaches, the specific objectives of this study are to develop : 1. Develop a framework demonstrating the relationship between project risk factors, collaborative project delivery pract ices, goal alignment, TMS , and performance outcomes by; 1. Qualitatively examining the following at partnered - project level: i. The links among project risk factors , collaborative project delivery practices , and project performance ; and, 67 2. Quantitatively examining the following at individual - level in interorganizational AEC project teams: i. The relationship between individual /team performance , goal alignment , and TMS . 2. Test an evaluation metric for goal alignment and utility of TMS metric to investigate AEC collabora tio n during project delivery; 3. Help facilitate collaborative contracting in construction projects by identifying key characteristics individuals in common. 4. Provide theoretical contributions to AEC literature understanding collaborative p roject delivery methodologies as a form of relational governance. Sound empirical research is grounded in a strong understanding of pertinent literature, identifying the gaps for research, and positing rch question to fill the gap (M. S. Davis, 1971; Eisenhardt & Graebner, 2007) . According to Yin, ( 2003) , various strategies are available to researchers wh ich can help answer the research question. The different strategies are unique to the research question explored. These research strategies are experiments, surveys, archival analyses, history, and case studies. Experiments are intended to test impacts of intervention on an outcome while controlling for other external factors using a control and experiment group (Creswell, 2009). Given that partnered construction projects are unique endeavors, it is challenging to randomly assign individuals to distinct con trol and experiment groups to assess effects across AEC project teams and projects. Therefore, multiple 68 case studies are investigated to explore project team dynamics within the context of partnered - projects. In this format, both qualitative (e.g., partner ing charter, partnering scorecards, project meeting minutes, partnering session documents) and quantitative (e.g., surveys ) data are available. The mixed - methods researc h approach followed in this study builds upon an emerging perspective to understand in terorganizational project teams in AEC literature and analyze subsequent data (Korkmaz, 2007) . In this study, individuals are embedded within teams (e.g., owner teams, design teams, and contractor teams) and teams are nested within case study projects. Therefore, three levels o f analyses become pertinent in this study, individual - level , team level, and project level. According to Yin (2003), multiple units of analysis offers greater flexibility to inspect data for consistent patterns across units and cases. Moreover, mixed - metho ds enables researchers to use two or more types of data collection and data analysis methods (i.e., quantitative, qualitative). These data are used to integrate findings and draw inferences from a single study or theoretical perspective (Miller et al., 2011) . Another advantage of mixed - methods is the ability to triangulate the findings of multiple forms of data, quantitative and qualitative (Campbell & Fiske, 1959) . Project - specific data such as collaborative project delivery practices (e.g., us e of partnering charters, scorecards, workshops) and project team rosters are collected from main project participants (e.g., partnering facilitator, owner representative, or Construction Manager/General Contractor (CM/GC) project manager ). Structured inte rviews are performed to collect data from key stakeholders representing the owner team (e.g., owner, owner representative, or other key stakeholder involved in project delivery) regarding project 69 performance outcomes (e.g., cost, schedule, quality, and con flict resolution). Project - level data are collected using a qualitative approach. These project - level data are analyzed in parallel with survey data using pattern - matching, content analysis, and cross - case synthesis to help integrate findings. Meanwhile, p roject risk factors and collaborative project delivery practices become fixed effects in quantitative analyses. Q uantitative dat a, via project rosters, are gathered from surveys administered a t or near project completion . Data collected from surveys are us ed o f goal alignment, TMS, and performance outcomes . This research design permits group comparisons (e.g., owner teams, design teams, and contractor teams) and statistical testing using factor analysis and multivariate r egression /correlation analysis (MRC) . These data are also aggregated forming group mean scores to inspect team level effects within case study projects. A partnered project delivery framework and relevant literature guide the methodology f or this research. The steps forming the research process followed in this study is illustrated in Figure 3 - 1, along with the chapter in which it occurs. The steps for this research include reviewing the literature, selecting the research strategy, developi ng construct measures, data collection procedures, data analysis, model validation, results from both cross - case synthesis and multivariate regression analyses, then, findings and conclusions. 70 Figure 3 - 1 : Research process and steps followed in this study . The population considered for this study consists of project participants and stakeholders involved in construction projects working under partnering arrangements in the U.S. The o bjective is to collect data from partnered - projects and their subsequent project teams. The project teams are represented by owners, design engineers, contractors, and subcontractors. The final study participant sample (n) size was 125 potential survey res pondents from six case study projects. Survey data were collected beginning in January 2018 , over a period of three months . 71 The mixed - methods research followed in this study uses metrics based on extensive investigation of team theory and AE C partnering literature. This study investigate s relationships for the conceptual multi level model which suggests individual /team AEC project team goal alignment (i.e., independent variable ) affects individual / team performance (i.e., dependent variables), yet is moderated by TMS (i.e., moderator variable ) . The project level model purports a clear relationship existing between project risk factors , collaborative project delivery methodologies, and project performance outcomes in partne red - projects. One key team theory, transactive memory system is a sub - domain of SMM thus is accessible for measurement using similar strategies (Mohammed, S., & Dumville, 2001) . For example, the shared mental model construct contains three characteristics that permit measurement. These characteris tic are elicitation , structure representation , and representation of emergence (DeChurch & Mesmer - Magnus, 2010) . The measurement approaches help interpret the extent of convergence or similarity Elicitation methods capture the content of the model using similarity ratings, card sorting tasks, concept mapping, or rating scales (DeChurch & Mesmer - Magnus, 2010). The elicitation method parses key elements of the task to understand the content based on participants responses. Structure representation is used to illustrate agreement regarding the similarity of team members as it is represented in a model (DeChurch & Mesmer - Magnus, 2010; Mohammed et al., 2000). While Kozlowski and Klein (2000) suggest representation of emergence responds to 72 the overall climate measuring both the content and strength of consensus at the focal le vel of analysis. This study use d the elicitation strategy to understand whether a strong TMS is present within partnered - project teams. The independent variable in this study is goal alignment while transactive memory system is a moderator variable of in dividual, team, and project performance. Performance constructs are the dependent variables that will be investigated in this study. These data will also be aggregated to form a group mean for team level analyses. The latent constructs in this study entail both formative and reflective indicators. Reflective indicators are observed variables perceived as reflective (i.e., effect) indicators of an underlying construct or latent variable (e.g., TMS, personality, attitude, etc.). In contrast, formative indica tors are observed variables perceived as formative (i.e., cause, causal) indicators that are assumed to cause a latent variable (e.g., Socio - economic status (SES), etc.) (Diamantopoulos & Winklhofer, 2001) . This st udy assesses quantitative data regarding goal ali gnment, TMS, individual/ team performance using reflective indicators associated with reliable metrics in team literature (Hoegl & Gemuenden, 2001; Jap, 1999; Kyle Lewis, 2003) . Most of the items are assessed using a five - point Likert s cale (e.g., 1 - strongly agree to 5 - strongly dis agree) or by answering multiple choice questions (e.g., role in the project ) within the survey. Meanwhile, qualitative data on project risk factors , collaborative project delivery practices , and project perform ance are measured using formative indicators resulting from the AEC literature. 73 The following sections are used to describe the constructs and metrics for qualitative data collected in this study. The constructs are proj ect risk factors , collaborative project delivery practices , and project performance . Project risk factors are assessed using best practice guidance and AEC literature which assert key attributes involved in the effectiveness of partne ring (Gransberg et al., 1999 ; IPI, 2016 ) . Utilizing these risk factors, structured - interview questions are developed to ascertain the desired level of partnering anticipa ted for the project. This allowed the researcher to determine how risk factors are related to collabo rative project delivery practices, individual/ team , and project performance. The variables and measures used to investigate project risk fa ctors are displayed in Table 3 - 1 . These variables are assessed using a scoring system to differentiate between certai n factor s such as project risks with potential impacts on cost/time, complexity, and political significance. To do so, each category is scored from 1 - Not important to 5 - Very important. The questions also include a contextual portion related to the specific case study project. As an example, schedule risks with potential impacts on cost/time includes options to select from such as none, limited, and many. 74 Table 3 - 1 : Measures used to assess common project risk factors (Gransbe rg et al., 1999; IPI, 2016). Common Project Risk Factors: Evaluation Method Not important=1 to Very important=5 Number of project risks with potential impacts on cost/time (e.g., complex design and construction, public - private partnership, compressed s chedule, uncommon materials, etc.) Few Moderate Many 1 2 3 4 5 Schedule risks with potential impacts on cost/time (e.g., liquidated damage and/or incentives) None Limited Many 1 2 3 4 5 Project team relationships Team ha s worked together before and has solid partnering foundation Team has no prior experience working together but has partnering foundation Team worked together before but no partnering foundation Team has not worked together and has no partnerin g foundation 1 2 3 4 5 Team partnering experience Experienced Some experience Most team members new to partnering 1 2 3 4 5 Political significance and community interest High visibility (significant strategic project) Probable (organiza tion image at stake) Likely, depending on the size of the client and place of importance Unlikely, unless in a place of importance 1 2 3 4 5 Complexity High (i.e., highly technical and complex design and construction; short timeline/ schedule con straints, uncommon materials, new supply chain, etc.) Increased Moderate Standard 1 2 3 4 5 Project Delivery Method DBB DB CM/GC CMR Other 1 2 3 4 5 75 Each measure of project risk shown in Table 3 - 5 is given a score us ing the following equation: In the equation above, the response options for each variable are : (1) Few, (2) Moderate, or (3) Many (1) None, (2) Limited, or (3) Many (1) Team and partnering experience high, (2) Team experience low and partnering experience high, (3) Team experience high and partnering experience low, or (4) Team and partnering experience low (1) Experienced, (2) Some experience, (3) Most team members new to partnering (1) Unlikely, (2) Likely, (3) Probable, or (4) High visibility (1) Standard, (2) Moderate, (3) Increased, or (4) High Collaborative project delivery practice s are measured using metrics developed in a partnered - project delivery framework and AEC literature review (Mollaoglu & Sparkling, 2015) . Based on 76 the li terature, some of these practice elements are the use of partnering workshops for project teams, establishing mutual goals and objectives, and project surveys to monitor partnering processes (C han et al., 2004; D Hughes, Williams, & Ren, 2012; Deborah Hughes et al., 2012) . These formal and informal governance strategies help project team members align their goals and objectives ba sed on previous experiences. The measure to understand the level of collaborative project delivery practice followed and used in this study is purp orted in Table 3 - 2 . The structured interview questions also use yes, or no responses to certain items. Additionally, the survey intends to capture the importance of each prac tice using a Likert score ranging from 1 - Not important to 5 - Very important. The scores for collaborative practices are calculated based on the equation below: Whereby, (1) Yes, or (0) No (1) Yes, or (0) No (1) Yes, or (0) No 77 Table 3 - 2 : Full description of collaborative project delivery practices construct, metrics and measures used in this study (Chan et al., 2004; IPI, 2016; Mollaoglu & Sparkling, 2015). Level of Collaborative Project Delivery Practices: Evaluation Method Yes No Not important=1 to Very important=5 Contractual Related Practices Professional facilitator was used i n this project. 1 2 3 4 5 A shared equity arrangement was indicated in contracts. 1 2 3 4 5 A partnering charter was used in this project. 1 2 3 4 5 A proactive conflict management tool that added structure to collaborative problem - solving processes was used in this project. 1 2 3 4 5 Equal power/empowerment was afforded to all project teams and team members in decision - making processes. 1 2 3 4 5 An incentive/fee/risk - reward/ or gainshare - painshare agreement was established in co ntracts. 1 2 3 4 5 Procurement Related Practices Parties were selected based on partnering experience. 1 2 3 4 5 We selected team members based on previous work experience with other team members. 1 2 3 4 5 Parties were sele cted based on technical expertise. 1 2 3 4 5 There was early involvement of key participants (e.g., designer/contractor/specialty subcontractors) during schematic design (SD). 1 2 3 4 5 1 2 3 4 5 Project Related Practices Partnering workshops were held for this project. 1 2 3 4 5 Partnering scorecards were used in this project. 1 2 3 4 5 There were two or more project teams located together in a common office (i.e., colocation). 1 2 3 4 5 Partnering training/team - build ing sessions were held for this project. 1 2 3 4 5 Measurable and achievable milestones were established to determine the success of the project. 1 2 3 4 5 Project teams openly exchanged information across organizational boundaries (e.g., Buildin g Integrated Modeling (BIM)) 1 2 3 4 5 Quarterly partnering meetings were used in this project. 1 2 3 4 5 Monthly partnering meetings were used in this project. 1 2 3 4 5 Multi - tiered partnering was used in this project (i.e., executive, co re team, stakeholders) 1 2 3 4 5 Specific task force used for conflict and issue resolutions 1 2 3 4 5 78 This goal alignment measure is based on goals and objectives elicited in case study partnering charters, therefore, is specifically aligned with each case study project. For example, some projects included safety, schedule, budget, and submittals as goals in their partnering charters with well - defined performance metrics ( sample case study partnerin g charter, goal aligning objectives, and performance metrics shown in Figure 3 - 2 ). Goals and objectives items identified in partnering charters were used to measure this construct using a five - point Likert scale (i.e., 1 - strongly disagree to 5 - strongly agr ee) to rate perceptions of individuals on their project - specific goals. This study collected partnering scores from case studies to indirectly investigate goal alignment. Figure 3 - 2 : Sample Case Study Project Scorecard, Goal Aligning Actions, and Performance Metric 79 The scores for goal alignment are calculated based on the equation below: Whereby, (1) Strongly disagree to (5) Strongly agree (1) Strongly disagree to (5) Strongly agree The project performance construct used in this study is developed from extant AEC literature. Project performance entails three first - order variables or elements from which measures are determined. The three ele me nts are cost, schedule, and quality and safety performance . These also include owner satisfaction perceptions. Cost refers to outcomes regarding cost growth and additional expenses as a result of changes or other conditions during project delivery (Grajek , Gibson Jr., & Tucker, 2000; Gransberg et al., 1999; Yeung, Chan, Chan, & Li, 2007) . Schedule refers to time performance such as being ahead or behind as compared to original contract completion dates (Grajek et al., 2000; Gransberg et al., 1999; Yeung et al., 2007) . Quality and safety performance is concerned with the quality rat ings, reducing the amount of wasted work or rework, and end - user satisfaction of the project. Meanwhile, safety performance is centered on accident rates (Yeung et al., 2007) . This study investigates this construct using a five - point Likert scale (i.e., 1 - not satisfied to 5 - very satisfied ) , among other things, to rate perceptions of owners on their responses to questions within the survey. Table 3 - 3 describes the measures of the project performance used in this study. 80 Table 3 - 3 : Full description of the project performance constructs, metrics and measures. Cost Performance: Evaluation Method Cost Growth Contrac t Project Cost: $____________ Actual Project Cost: $____________ (Indicate if values are estimated) Total Cost of Partnering Facilitator, meetings expenses, surveys, etc.: $____________ Estimated Savings Due to use of Partnering Cost Savings as a % of Project Budget: %____________ with the Cost Performance of this project Not satisfied=1 to Very satisfied=5 1 2 3 4 5 Schedule Performance: Evaluation Method mm/dd/yy Planned Start Date (Starting from P lanning): Planned Completion Date: Planned # of Work Days (Construction): ___________ Actual Start Date (Notice to Proceed): Actual Completion Date (Substantial Completion): Actual # of Work Days (Construction): _____________ Satisfaction with the Schedule Performance of this project Not satisfied=1 to Ve ry satisfied=5 1 2 3 4 5 Quality and Safety Performance: Evaluation Method Quality Performance of this project Not satisfied=1 to Very satisfied=5 1 2 3 4 5 Performance of th is project Not satisfied=1 to Very satisfied=5 1 2 3 4 5 81 P roject performance scores are calculated using the following equation: Whereby, (1) Strongly disagree to (5) Strongly agree (1) Strongly disagree to (5) Strongly agree (1) Strongly di sagree to (5) Strongly agree (1) Strongly disagree to (5) Strongly agree Cost and schedule growth are objective metrics upon which construction projects are often compared (Touran, Gransberg, Molenaar, & Ghavamifar, 2011) . These common metrics were used to report performanc e ratings based on project partnering documents from the six case studies. Cost and schedule growth performance ratings are computed using the following equations: Meanwhile, par tnering facilitation cost as a percentage of the original project budgeted cost is calculated using the equation: 82 These computations are used to make objec tive comparisons across cases. A performance ranking is given to each case study project using the cost and schedule growth ratings. These rankings were similarly based on the aforementioned ratings. The following secti ons are used to describe the quantitative constructs and metrics in tended for this study. The constructs are goal alignment , transactive memory systems (TMS), individual/ team performance . Goal alignment as a const ruct is measured by assessing the congruence among individuals working in AEC project teams. The process of collaboration and goal alignment across organizational boundaries involves learning curves in working as a team, bringing together varied skills, an d investments in time and resources. Based on the literature, there are many collaborative practice elements used to align project teams such as the use of partnering workshops, establishing mutual goals and objectives, and involving key project stakeholde rs early in the design and construction project processes (Chan et al., 2004; Hughes et al., 2012 a ; Hughes et al., 2012 b ) . This study use s a measure of goal alignment to investigate causality among coordinated efforts across organizational boundaries (Jap, 1999) . The measure is based on reflective indicators shown in Table 3 - 4 . This study investig ate s this construct using a five - point Likert scale (i.e., 1 - strongly disagree to 5 - strongly agree) to rate perceptions of individuals on their responses to questions within the survey. 83 Table 3 - 4 : Full description of goal alig nment construct and measures used in this study. Goal Alignment: Evaluation Method (Multiple Choice: Likert Scale) Goal Alignm ent Mechanisms (Chan et al., 2004; Jap, 1999) Mutual goals and objectives in the partnering charter were communicated effectively. Clear and compatible partnering goals were established by the project team. I generally agreed with project - related goals established by the project team. My attitude towards project - related goals established by the project team were similar. My goals for the pr oject were in close alignment with the project team. A three - dimension scale is used to measure the second construct transactive memory system . The first order variables are specialization, credibility, and coordinati on with five items for each dimension (Lewis, 2003). TMS allows researchers to assess how well team members understand who possess specialized knowledge (specialization), how well they trust and rely on that knowledge (credibility), and the way this knowle dge is efficiently organized (coordination). The widely adopted scale is used in team research to measure transactive memory systems (Lewis, 2004; Rau, 2005; Zhang, Hempel, Han, & Tjosvold, 2007; Zheng, 2012). The items in the scale capture responses using a five - point Likert scale ranging from 1 (strong agree ) to 5 (strongly dis agree). To measure TMS in this study the follow ing variables shown in Table 3 - 5 are assessed (Lewis, 2003; Zhang, Cheng, & Fan, 2015) . 84 Table 3 - 5 : Full description of transactive memory system construct, metrics and measures used in this study (Lewis, 2003). Transactive memory system: Evaluation Method (Multiple Choice: Likert Scale) Coordination The project team worked toge ther in a well - coordinated fashion to complete the project. The project team had very few misunderstandings about what to do during construction. I believe we accomplished our task for the project smoothly and efficiently. Credibility I was comforta ble accepting procedural suggestions from other team members I was confident relying on the information that other team members brought to the discussion When other members gave info rmation, I wanted to double - check it for myself. (reversed) Specialization I understand what skills my team members have and domains they are knowledgeable in. The specialized knowledge of several different team members was needed to complete the proj ect. In AEC project teams, specialization is an inherent property of project teams. For instance, owners, designers, contractors, and subcontractors come together knowingly bringing together their unique expertise to deliver projects. Specialization is one dimension that may not provide meaningful information in the context of AEC project teams, however, is still measured in this study. There a re three dimensions that will be assessed in this study regarding individual/ team performance perceptions. Those are project - related , communication - related , and team - related . Project - related outcomes are investigated using common cost, schedule, quality a nd safety perceptions found in AEC literature (Grajek et al., 2000; Gransberg, Reynolds, & Boyd, 1997; 85 Yeung et al., 2007) . These perceptions are use d to assess how well project teams feel their performance was aligned with owner project goals. The second dimension, communication , is a measure of efficient information and knowledge sharing beneficial for team integration. The measure teases out efficiency, clarity, an d frequency at which communication occurs within project teams (J. S. - C. Hsu, Shih, Chiang, & Liu, 2012) . Meanwhile, team - related refers to relational skills and attributes that are beneficial to the development of team communica tion, cohesion , and trust (Le - Hoai, Lee, & Son, 2010) . Hoegl and Gemuenden (2001) , purport several metrics to understand not only teamwork quali ties but, individual and team performance. Based on their study, a reliable measure of individual / team performance is asserted consisting of several indicators . For example, the measure examines whether equitable decision - making, information sharing, and m utually beneficial outcomes were afforded to all project team members. This study investigate s these dimensions using a five - point Likert sc ale (i.e., 1 - strongly agree to 5 - strongly dis agree) to rate perceptions of individuals on their responses to questi o ns within the survey. Table 3 - 6 describes the measures of the individual / team performance construct us ed in this study. Individual perceptions are aggregated to investigate sub - team performance outcomes. For instance, the owner , design, and construction t eam responses are aggregated forming sub - teams permitting inspections of variables at the team level of analysis. 86 Table 3 - 6 : Individual/ team performance construct, metrics and measures used in this study . Individual /Team Pe rformance: Evaluation Method (Multiple Choice: Likert Scale) Project Related (Grajek et al., 2000; Gransberg et al. 1999; Yeung et al., 2007) We adhered to cost goals for the project. We adhered to schedule targets for the project. W e satisfied the scope of work for the project. We accomplished our task for the project smoothly and efficiently. We achieved project goals established by the project team. We delivered a high - quality project for the owner. There was little rework required in this project. The project was delivered safely without major safety incidents. Communication Relate d (Hsu et al., 2012) There was efficient and effective information sharing among the project team (e.g., RFI responses). There was frequent communication within the team. Team members communicated often in spontaneous meetings, phone conversations, etc. The team members largely communicated directly and personally with each other. Team Related (Hoegl & Gemuenden, 2001) All project team members were treated equal ly in the decision - making process. All project team members worked with the same focus on project objectives. We worked together to share informati on across organizational boundaries. We worked towards mutually beneficial outcomes for all participants. All project information was readily available to everyone involved in the project. We always sought collective identification and resolution of problems. Project team accountability was emphasized for all project outcomes. Several of the measures in this study are underutilized in AEC literature, thus are assessed for adequacy as valid metrics using confirmatory factor analysis (CFA). This is covered in more detail in the analysis section. 87 Potential partnered - projects and teams are identified through collaboration with the International Partnering Institute. Researchers contact ed project representatives v ia email c ommunications and sent relevant information describing the research. Once key project representatives agree d to participate, follow communications were made to determine project characteristics (e.g., project size, duration, project delivery method, projec t type, etc.) and consistency with study goals. In advance of data collection, a pproved human subject protocols are implemented in this study established by the Institutional Review Board of Michigan State University. The research documents and study data collection instruments are presented to the Institutional Review Board (IRB) at Michigan State University for approval. The documents include the survey instruments, structured interview questions, and other protocols required for human subject research. This study receive d an expedited approval due to its voluntary nature clear identifiers are requested from study participants. A pilot su rvey of the intended population was used to test the survey instrument. A surve y provides quantitative information depicting attitudes and opinions of the population under consideration. Survey data is collected via partnered construction projects that are completed or nearing completion. Obtaining field level data places the investi gated phenomena into the real context where implications can be understood and accessed (Yin, 2003) . The rationale is to gain information specifically emanatin g from participants involved in the 88 partnering process, although no field manipulations or experiments are undertaken in this pilot study. The final survey instruments and metrics are revised based on feedback from industry professionals and academics. To understand the extent to which collaborative project practices affects goal alignment and subsequently performance outcomes , several project teams are sampled from a database of construction projects across the U.S. working under partnered - project arrange ments. In addition to survey data, project /team level data are collected via collaboration with key partnering team members and partnering facilitators. Structured interviews are conducted with key representatives of the owner survey questionnaire (Appendix B). The project documents include partnering charters, partnering scorecards, partnering session documents, and project meeting minutes. To access the information, a web - based file sharing platform is established to allow the researcher to collect project data. The information sharing is managed by the key project team members (i.e., Requests to participate in the survey are sent to key project team members (e.g., owner representatives, construction project managers) using email and monitored by researchers. Researchers are included on all correspondences to monito r potential respondents and for follow - up . An online survey system is used to administer and collect survey data. Follow - up is provided to project participants identified by key project team members and includes a link to 89 the online survey. The survey inst rument consists of three major components to assess team project delivery (Appendix A ) . S elf - reporting components in the survey are used to infer the existence of trans active memory systems and capture performance outcomes. According to Lewis and Herndon (2011), indirect measures using indicators are most appropriate to predict the existence of TMS or its effects. This is especially important in setting s where observing or direct measuring is not practical or feasible. As previously mentioned, a mixed - methods research approach is followed in this study. There are two levels of analyses pertinent to this study, individual - level and project level. According to Yin (2003), multiple units of analysis offers greater flexibility to inspect data for consistent patterns across units and cases. Moreover, mixed - methods enables researchers to use two or more types of data collection methods and data analysis methods (i.e., quantitative, qualitative). These data are used to integrate findings, draw inferences within a single study or theoretical perspective, and triangulate findings from multiple forms of data (Campbell & Fiske, 1959; Miller et al., 2011) . The section disc usses qualitative data analysis methods while section 3.9 covers quantitative data analysis methods. Th is stage of the mixed - methods approach is to understand qualitative evidence resulting from the case studies. The project level data are collected using a qualitative approach. Project - specific data is comprised of evidence such as collaborative project delivery practices (e.g., use of partnering charters, scorecards, workshops) and project risk factors (e.g., cost/schedule impacts, partnering team experi ence, complexity, etc.). Mean while , project team rosters are 90 collected from main project participants (e.g., partnering facilitator, owner representative, or Construction Manager/General Contractor (CM/GC) project manager ). Structured interviews are perfo rmed to collect data from key stakeholders representing the owner team (e.g., owner, owner representative, or other key stakeholder involved in project delivery) regarding project performance outcomes (e.g., cost, schedule, quality, and conflict resolution ). These project - level data are analyzed in parallel with survey data using pattern - matching, content analysis, and cross - case synthesis to help integrate findings. Meanwhile, p roject risk factors and collaborative project delivery practices are objectivel y scored during structured interviews. These data are also qualitatively analyzed for links with other variables. Before delving into the case study data analysis and model validation, reliability and validity are discussed. There are effective case study tactics available to ensure validity and reliability criteria are satisfied. In case studies, the researcher is concerned with four design tests being construct validity, internal validity, external validity, and reliability. The four tests, case study tactics, and phase of research in which the tactic occurs , in parentheses () , are described below (Yin, 2003) : Construct validity: Correctly operationalizing measures for the concepts under investigation; Important to specify changes that are to be studied and demonstrate how measures of these changes adequately reflect specific types of changes selected. Use multiple sources of evi dence (data collection) Establish a clear chain of evidence (data collection) 91 Solicit key informants to review draft of the case study report (composition) Internal validity: Establishes a causal relationship in explanatory case studies where certain con ditions are shown to lead to other conditions and are not spuriously related. Do pattern - matching (data analysis) Do explanation - building (data analysis) Address rival explanations (data analysis) Use logic models (data analysis) Use cross - case synthesis ( data analysis) External validity: Establishes the domain where findings can be generalized. Use theory in single - case studies (research design) Use replication logic in multiple - case studies (research design) Reliability: Demonstrates that proper operati ons of a study can be replicated with the same results. Use case study protocol (data collection) Develop a case study database (data collection) This study adopt s these tactics to satisfy the research quality criteria in order to benefit from the richne ss found in case studies. Well - done theory building from multiple - case studies, similar to experiments, can be very objective and allow formal analytical modeling (Eisenhardt & Graebner, 2007) . 92 In this study, the researcher used six case studies to establish a clear chain of evidence. The data co llected from case studies were compiled into a database for further inspection and analyses. Structured interviews with key stakeholders (i.e., owner/owner representative and contractor) are recorded with participates consent to aid in transcriptions. The trans crip tions for validation. These observations are used to give a recapitulation for each case study, analyze data using pattern - matching , and to provide finding s from cross - case synthesis. Using single and multiple - case studies in theory building to establish propositions, theoretical constructs, and provide empirical evidence continues as a valid and rich research strategy (Eisenhardt & Graebner, 2007) . Despite the advantages, Eisenhardt and Graebner (2007) contend challenges abound as to why inductive theory building rather than theory testing is required to understand the specific phenomenon. In response, emerging and competing theories are generally better understood by plac ing the phenomenon in its context, especially this, a multiple case study approach is followed in this study. According to Yin, ( 2003) , there are two general analytic strategies used in case study analyses. The first is to rely on theoretical propositions as a guide which form s the rese arch design of the case study. The propositions also help direct the analysis to specific data so other superfluous information can be ignored. For example, many other things take place during partnered - project delivery which can be tracked such as achievi ng certain project schedule milestones. This information, though important, is irrelevant to the goals of this study. The 93 second strategy is to develop a case description or framework for organizing the case study. This should include the general character istics and relationships that existing within the phenomenon. Both of these strategies can be achieved by using pattern - matching, explanation - building, or time - series analysis (Yin, 2003) . Pattern - Matching Pattern - matching is a case study technique used to compare empirically based patterns with predicted ones. When doing pattern - matching the aim is to inspect the data for expected outcomes (i.e., predicted re sults found and alternative patterns absent), rival explanations (i.e., the presence of certain explanations should exclude the presence of others), or simpler patterns (i.e., pattern - matching is valid with only a few variables or clear differences among d erived patterns). Explanation Building Meanwhile, explanation - building is an iterative process of analyzing case study data to identify causal links. The process begins with initial theoretical statements, then, findings of the cases are compared. The sta tements are revised, if needed, and compared against details in the case. This is followed by additional revisions when require d to assure specific propositions are analyzed, then, compared with the other cases. Yin (2003) cautions against this technique a s the analyses may shift from the original topic under investigation. Time - series analysis is events that change over time. With this type of case study analysis, the res earcher attempts to identify theoretically proposed sequences of events believed to lead to specific outcomes. The 94 events must be explicitly identified by the researcher prior to the investigation. From this, comparisons can be made between the trends in t he data with those of empirically derived data points. Cross - Case Synthesi s The purpose of cross - case synthesis is to identify common patterns among data which provides internal validity. An objective scoring or comparison criteria is established to facil itate comparisons across multiple case studies. Content analyses are conducted among objective data for similarities and/or differences in the features of each case study. Quantitative d ata resulting from this stu dy are analyzed using two distinct approaches which are reviewed in this section. A description of factor analysis is given first. Factor structure analysis is used to ensure the data is representative of constructs asserted in this study. Next, confirmato ry factor analysis, a model validation technique, is discussed. This technique is used to examine if the model is consistent with the data. Multivariate regression correlation analysis (MRC) is described following model validation. MRC is used to analyze m ultiple variables present in statistical models. MRC permits testing relationships among independent and dependent variables while remaining more flexible than analysis of variance ( ANOVA ) and analysis of covariance ( ANCOVA ) (Cohen, Cohen, West, & Aiken, 2013) . Before discussing data analysis procedures, reliability and validity are addressed. 95 There are several tactics available to the researcher to address reliability and validity during data collection. The str ategies from which one can choose varies and depends on how data is captured in the study (i.e., survey instrument, observations, ethnography, etc.). This study uses a mixed - methods approach to collect both empirical evidence and case study data, thus are discussed separately. Section 3.9.1 deals with case study data reliability and validity concerns. Strategies for Survey Measurements Reliability is primarily concerned with the consistency of items within a measure and stability of the measure over tim e. Several methods are available to assess the stability of a measure such as test - retest, alternate forms, and split - half test (Nunnally, 1978). These methods are able to determine the reliability coefficient (i.e., alpha) during conditions of transient, content, random response, and rater error. The internal consistency of a measure is generally associated with interrater reliability. Homogeneity of the sample and test length can affect the variation within rater responses and, therefore, true score and o bserved score variance. Spearman - Brown prophecy is one such method to evaluate internal consistency and the appropriate test length. A valid measure is the notion that we are me asuring that which is intended (Nunnally, 1978) . To inspect the validity of a measure, a multi - trait multi - method (MTMM) procedure or factor analysis (FA) can be followed (Shaffer, DeGeest, & Li, 2016) . The MTMM procedure will help partition the responses and variance according to shared method varia nce, shared trait variance, and other combinations based on a matrix. Perhaps, the measure is encapsulating multiple dimensions that were thought to be unidimensional. This can also be seen in a factor analysis. In factor analysis, responses are inspected to determine the amount of variance based 96 on how it loads onto a single factor and its items. From this, a factor structure emerges which can serve as an indication of dimensionality. The data from indi vidual team members are aggregated to represent information investigated at the team level. Inter - team - agreement can be assessed by calculating intraclass correlation coefficients (James, Demaree, & Wolf, 1984) . T his method is applied as justification for aggregating individual scores into team - level scores, therefore providing inter - rater reliability. The constructs investigated in this study are measured using measurement models which have underlying indicators or factors. For example, TMS has three dimensions with five indicators measuring each dimension. Measurement validity can be assessed in this instance as there are five parameters to estimate and 10 correlations that can be employed to generate estimates. These estimates are used to assess the fit of the model. The internal consistency theorem is used to examine the resulting correlation matrices. This analysis allows the researcher to inspect predicted and obtained inter - item correlations and subsequent ly the validity of the measurement (Hunter, Gerbing, & Boster, 1982) . Confirmatory factory analysis ( CFA ) is explored to assess whether a model fits the data. This method allows one to analyze the data against various goodness - of - fit indices [e. g., comparative fit index (CFI), root mean square error of approximation (RMSEA)]. The application of CFA is also used in convergent and discriminant validity analysis. Results from these analyses validate whether the hypothesized model is consistent with the data to which assertions can be made and discussed. 97 Structural equation modeling (SEM) is used to test the hypothesi zed model in this study (Figure 3 - 3 ). In the model X represents goal alignment , M is transactive memory sys tem, and Y is individual/ tea m performance. The model asserts the three parameters that are available for estimation to test the hypothesized model. Figure 3 - 3 : The Relationship and Hypotheses between Goal alignment, Transactive Memory Systems, and Ind ividual/Team performance The above parameters are estimated by their correlation coefficients according to ordinary least squared estimation (OLS) . These predicted correlations are analyzed against obtained correlations and used to determine if the model i s consistent with the data. From this analysis, MRC analysis using multilevel modeling strategies can be explored. Multilevel data is oftentimes used in the field of psychology research to investigate personal ity, social and organizational behavior. Other uses for multilevel modeling are education and educational policy, communication research, and sociology. Essentially, its practicality applies to any situation in which the data are hierarchical in nature wit h both macro and micro level phenomena (Nezlek, 2008; Snijders & Bosker, 2012) . As an example, in studies involving Individual/Team Performance (Y) Goal Alignment (X) H2 H1 TMS (M) In dividual Level Data / Quantitative Investigation 98 individual pupils in classrooms , the data is comprised of several units of analysis . In this example, level one would be representative of pupils, level two being classrooms or schools. Typically the levels are reflective of the hierarchy in which the sample is selected from a larger population. Multilevel modeling is also applicable to longitudinal data or study comparisons in meta - analysis research (Hox, 2010) . There are several distinct advantages of multilevel modeling over ordinary least s quared (OLS) regression. In multilevel modeling, disaggregated data are available to accurately investigate from any level, whereas in OLS the assumptions of independ ent observations are violated resulting in unfounded conclusions (Snijders & Bosker, 2012) . Moreo ver, multilevel modeling permits a model in which dependent variables are at the lowes t level and explanatory variables are defined at any level including aggregation of level one variables. Figure 3 - 4 is shown to demonstrate the modeling levels in order t o study organizations (i.e., Level 1) and construction projects (i.e., Level 2). Based on the figure, organizational units are clusters of interorganizational project teams nested within construction project units. The complex interplay between characteris tics of the two units can be analyzed via a multilevel model. Thus, the structural dependence explaining within - and between - group variance are efficiently elicited. 99 Figure 3 - 4 : Multilevel sample selected at random from population According to Snijders and Bosker ( 2012) , care must be taken to avoid three areas of potential errors when working with multilevel models and data aggregation. An error can occur inferred. For instance, a variable that is aggregated to level 2 (e. g., team cohesive ratings) used as a metric for team performance are not reflective of reflectiv e of how variables are constructed. When lower level variables are aggregated as the mean of a group, it is interpreted as a structural variable (Hox, 2010) . The mean structural variables are, then, used as the explanatory variable at higher levels of analysis. Alternatively, contextual variables are disaggregated in that they receive the group mean to which they belong at the higher level of analysis (Hox, 2010) . A second well - known source of error is related to ecological fallacies. An ecological fallacy occurs when one falsely asserts correlations between macro - level (i .e., level 2) variables which are then used as substitutions for micr o - level 100 (i.e., level one) variables (Robinson, 1950) . As an example, Robinson (1950) belied that researchers were tacitly abusing macro - level data connecting race to illiteracy as explanatory for same variables at the micro - level of analysis. The third potential error source is that of disregarding the original data structure (Snijders & Bosker, 2012) . A concern is with data aggregation which is used to derive our within - group and between - group regression lines. When the data structure is not firmly c onsidered or understood, misleading interpretations can occur. Knowin g this, multilevel modeling is efficient in handling disaggregated data, and as a result , allows researchers to investigate cross - level interaction effects. The use of random coefficien t models is best when 1) the groups are believed to be unique categories from which conclusions will be made specific to the categories; 2) the groups are asserted as a sample representative of a larger population (i.e., real or hypothetical) and conclusio ns are made by the researcher regarding this population; 3) testing the effects of group - - group variability; and, 4) dealing with small group sizes (i.e., 2 - 100) which assumes independent and identical distributed group effects (Snijders & Bosker, 2012) . In satisfying the condition s above, this study implores the benefits of a random effects model. The data structure and a basic multilevel model are presented next. In this study, the researcher examines two - le vel data represented by AEC project teams (i.e., macro - level) and individuals (i.e., micro - level). One basic multilevel model underlying this data structure is described by the foll owing random effects equation: 101 Or, In the model above, there is continuous variable . These are modeled as a function of the intercept for ea ch level 2 unit ( ), error ( ), and the variance of ( ) is the level 1 random variance. The regression coefficient for is (i.e., average effect of goal alignment in group ) and similarly , the r egressi on coeff icient for is (i.e., average effect of TMS in group ) . The interaction effect has a regression coeffi cient f or given by (i.e., average effect of interaction between goal alignment and TMS in group ) . The random and f ixed sl ope intercepts for the above are given by: The full model , when rewritten, is given as: The full model above is used for analyses at the individual - level while controlling for the sub - team level. The individua l model above asserts individuals are nested within sub - teams. Moreover, the effects of individual goal alignment and individual TMS are expected to vary 102 randomly with individual performance . The model also includes the mean effects of sub - team goal alignm ent and team TMS or within team effects on individuals. In the model above, the effects of sub - team goal alignment and sub - team TMS on sub - team performance is fixed across projects. This chapter postulates a clear methodology to which this study follows in addressing the research needs. The objective of this study is to better understand the relationships between project risk factors , collaborative project delivery practices , goal alignment , t ransactive memory systems (TMS) , and performance outcom es in interorganizational AEC project teams. These objectives are achieved by following a research process as describe d in this chapter. The research process entails the selection of an appropriate research strategy, developing study metrics, and establish ing data collection procedures. This chapter finishes with the procedures for both quantitative and qualitative data analyses. Based on t he structure of the data, multi level modeling techniques are described in detail as well - suited for this study along wi th case study analysis. Meanwhile, the chapter discusses factor analysis techniques used to understand the factor structure of constructs, along with, how these approaches are used in model validation . 103 CHAPTER 4 QUALITATIVE ANALYSIS This chapter gives the results from case study project documents and structured interview data used in qualitative analyse s . In all, six case study projects and their teams were investigated to understand the relationships between project risk factors , collaborative project delivery practices , goal alignment , and performance outcomes . It begins with a brief description and summary of each case study in section 4 .1 . Then, pattern - matching and cross - case analyses are presented in section s 4 .2 and 4 .3 to summarize study finding s. The following sections offer a brief introduction to each case study investigated in this study. Six case studies were investigated varying in size, duration, and project types. The researcher conducted structured in terviews with two individuals from each case study project, representing both the owner and contractors. P roject details were gathered from project documents, correspondence with project participants during structured interviews, and the faci litators from each case study. This project involved installation and green infrastructure improvement s associated with a storm sewer system. The project, located within a local community, was part of a multibillion - dollar citywide Sewer System I mprovement Program (SSIP). Several unique green features were incorporated in the project. For example, permeable pavement systems and rain gardens were installed to effectuate natural soil filtration processes. The goals included improvements to residenti al streets, more pedestrian and bike - friendly corridors, and green spaces. In addition, 104 the City wanted to reduce stormwater runoff and its impact on the sewer system. The budget cost for the project was $5.59 million while the scheduled work duration was 375 workdays. The stormwater management system is designed to accommodate 950,000 gallons of stormwater each year. In addition, the project comprised two acres of impervious surface that were improved. The green features accounted for 18,444 square feet of permeable pavement/concrete and 2,250 square feet of rain garden. The project infrastruct ure encompassed eight city blocks in a major metropolitan city. The scope of the project was increased during construction to include stormwater infrastructure and surface replacements for two additional city blocks. The project followed a traditional DBB project delivery approach and included a formal partnering process. Partnering was institutionalized in project documents and requirements including the use of a neutral third - party partnering facilitator. The partnering structure for the project is described next. Partnering facilitation services for the project included the development of an informal partnering charter, a partnering kick - off session, three par tnering workshops, performance monitoring via scorecards, partnering meeting minutes, an issue resolution ladder , and action plans to reconcile issues. The project team identified nine goal - aligning objectives for the project as shown in Table 4 - 1. The pr oject charter also included specific action items and responsible leads for each of the primary goals. 105 Table 4 - 1 : Case S tudy #1 Project Charter Goals and Performance Metrics Goal - aligning Objectives Performance Metrics Safet y Zero accidents Excellent housekeeping Schedule Early completion Substantial completion by 3/8/17 Final completion by 5/18/17 Budget Potential Change Orders and Change Orders minimized Contingency not exceeded Cost savings documented Submitta ls Submittal process timely and well - managed Environmental Compliance No non - compliance notifications Focused on cleanliness Green Infrastructure Quality No rework Green Infrastructure (GI) supplier plans and specs clearly defined Communication/Coo rdination Organization streamlined Responsive decision - making process Community Appreciation No community complaints Timely notifications of scheduled work Team and Project Recognition Respe ct and trust for all team members No issues escalated above field project team Project recognized for GI advanced work achievements A decision - making and issue resolution ladder w ere also utilized in this case study project. There were four levels an d associated FAST teams designated at each level. Those were Field level, Project level, Program level, and Executive level ordered from lowest to highest. Additionally, it included time escalation triggers (e.g., one day, up to one week, etc.) and a proce ss to drive the decisions (e.g., discuss onsite ASAP, document in meeting minutes, email, and conference call). The project team can also instigate objective mediation from the partnering facilitator to help resolve issues . 106 Comments from Case Study #1 S urvey Respondents process. It was a pleasure working with everyone involved, and the results speak for themselves. The project was completed on time, within budget and with hi adjustment s to reach successful project completion. Communication was excellent. The contractor and all team members This case st udy project encompassed an 18.4 - mile long in - service water tunnel repair. The water tunnel serves a large population and many stakeholders (e.g., Public Utilities) which made it critical to complete the work efficiently. The water system required a complet e shutdown for extended periods of time making it even more important to get buy - in from multiple - stakeholders and project participants. The project consisted of thorough pre - inspections of the water tunnel system , lining repairs and complete replacements , rock coring to improve the tunnel system, and numerous safety precautions. Due to the extreme safety conditions imposed on workers, safety was the number one priority on the project. Therefore, extensive precautions were in place such as having specific safety personnel and rescue teams dedicated to the project, frequent safety audits, along with ventilation systems and monitoring throughout the entire tunnel span. Safety concerns ultimately ended up becoming part of the partnering charter. 107 The b udgeted cost for the project was $4.96 million and was scheduled for completion in 255 work days. The case study project was delivered using a DBB project delivery approach and included partnering as a tool to help the project achieve its goals. A neutral third - par ty facilitator was used in the project. An informal partnering charter was developed and followed by the project team. In addition, facilitation services for the team included three partnering workshops, scorecards, partnering meeting minutes, an issue res olution ladder , and decision focused action plans. The goals of the project primarily revolved around safety, schedule, contract documentation, and clear communication amongst the teams. Eight goals shown in Table 4 - 2 were identified by the project team a nd served as goal - aligning objectives. Table 4 - 2 : Case S tudy #2 Project Charter Goals and Performance Metrics Goal - aligning Objectives Performance Metrics Shutdown Management Achieve completion milestone dates as required Safety and Health No injuries or accidents No lost or forgotten personnel left during shifts No environmental health issues during work (e.g., air, water, etc.) Schedule Shutdown completed by 2/27/17 @ 5:00 PM PST Contract Documentation No outstandin g change orders Communication/Coordination No communication/coordination misunderstandings Field Issues No issues raised above Project Leadership level identified on issue resolution ladder except unforeseen conditions Quality No wasted work or rework Integrated Team Responses Exceptional teamwork achieved The project team established five different levels of decision - making in their is sue resolution ladder. Those five level s from lowest to highest were Field leadership, Si te leadership, Project leade rship, Program leader ship, and Executive leadership. The Program level includes other project team members such a Senior Program Manager from the Public Utilities 108 Commission. A Dispute Resolution Advisor was also identified as part of the issue resolution procedures. This individual can be engaged informally by the project team at any level to help facilitate the decision - making process. The issue resolution ladder included key triggers for escalation and decision - making processes to resolve each issue at t he lowest level possible. Comments from Case Study #2 Survey Respondents Many team members were brought on after the start of the project, thus expectations and integration int o the team w ere problematic. Staffing needs to be finalized before the start , allowing for integration, roles and expectations to be fully addressed Check Locally Based Enterprise ( LBE ) qualifications prior to approval o f them as subcontractors under th e prime contractor. This lack of qualifications led to many disagreements on the project. I feel that it is important to get the project team thinking as a "we" in lieu of a "you and I" situation. I feel that by doing this everyone is working to complete the project and there are fewer areas for arguments which slows down the project. This was an extremely critical project for the ratepayers and the Public Utility . Everyone involved with the project understood the importance and worked together as a tea m. This case study project involved an estimated $165 million investment into removal and repairs to a major interstate highway and 13 associated bridges along a busy metropolitan community. 109 This ongoing project has a scheduled durat ion of 595 workdays for completion. The construction work included removing and replacing concrete surfaces and bridges that also support drainage systems, along with paving and patching. A unique feature of this case study was the replacement of a 49 - year - old , 1.63 - mile long bridge span which carries 37 million vehicles per year. As a result, complex demolition procedures were required and had to be reviewed well in advance of construction. I ntelligent transportation equipment systems were also installed a s part of the project. Some elements of the intelligent transportation system were electronic signage, cameras, and traffic monitoring systems which allow the Department of Transportation (DOT) to communicate travel times, accidents, and other mobility iss ues with drivers. Due to its proximity to oil refineries, railroads, residential communities, and utility infrastructure systems, the project team identified and frequently discussed project risks. Attention was directed towards Maintenance of Traffic (MO T) and safety of all individuals whether working onsite or traveling the surrounding areas. Some unforeseen project risks uncovered during construction were inferior steel decks which needed to be replaced along with some structural steel elements. These a dded change order cost and scope to the project. Meanwhile, the project team had to factor winter conditions into the construction duration and processes. The project delivery approach in this case study was DBB and incorporated partnering into its proj ec t requirements. The team used a neutral third - partnering facilitation service to help develop a formal partnering charter, facilitate partnering workshops, provide partnering meeting minutes, goals and 110 objectives. The partnering facilitator also used team building exercis es to help establish trust and shared objectives within the project team. Quarterly partnering workshops were held on the project with the team having completed nine, thus far . The project team also leveraged an issue resolution ladder with four levels of decision - making teams and defined escalation times (i.e., Level 1 Field : Immediate, Level 2 Project : ½ day, Level 3 Program : 1 day, and Level 4 Executive : 2 days). The formal partnering charter and mission statement captured mutual goals and objectives for the project stakeholder s. This document was solidified with signatures from all the respective parties. Table 4 - 3 illustrates the goals and performance metrics which the pr oject team followed in this case study. In all, 12 key goals were identified by the project team to which they revisited and evaluated during partnering progress meetings. Not only did the team include project goals but encouraged organizational opportunit ies for contractors and subcontractors. Table 4 - 3 : Case Study #3 Project Charter Goals and Performance Metrics Goal - aligning Objectives Performance Metrics Safety and Security Delivered safe and secure working environment fo r all stakeholders Schedule Meet project schedule and milestone dates Budget Complete project on budget Timely Decision - Making Make timely decisions to avoid delaying work Opportunity to Profit Contractor/Subcontractors afforded fair profit opportun ity Environment No environmental impacts Quality No wasted work or rework; Installed for longevity Risk Management Anticipate risks and collaboratively address all emerging issues Seizing Opportunities Always seek project improvements at every stag e Team Respect and Community Respect for team and concerns of all stakeholders including community Pride and Fun Make project a fun process and maintain pride for the finished project Open Relationship with FHWA Open communications/working relationsh ip to drive timely issue resolution Focus on Project First Empowered team focused on project goals 111 Comments from Case Study #3 Survey Respondents number of attendees was lower. With the large group settings and many specialist s in the room , - both Contractor and Owner - there remained a few personality traits that I believe contr ibuted to not allowing for a full y open and honest communication experience which in turn tends to hurt the partnering process. Overall to see that safety was This partnering case study project entailed complete demolition and replacement of a major metropolitan bridge with two new bridges to handle both eastbound and westbound traffic. This historic bridge dates ba ck to the 1950s and has been in operation for close to 60 years in the U.S. The project was surrounded by three railroad systems, a river, residential communities, existing businesses, supported approximately 20 public utilit y systems, and served as a majo r traffic corridor to downtown events. The project was part of a larger state infrastructure investment focused on improving safety, reducing traffic congestion and delays, and modernization objectives for the interstate system. As with many roadway and bridge projects, MOT and safety are of great concern to the project team. As such, the team followed careful planning and preparation strategies for bridge 112 demolition especially considering the required use of explosives. Other project risks involved in th e project were construction work 115 feet above active railroad systems and 136 feet above a major shipping corridor along the river below. Poor soil conditions were also identified early - on by the project team which required additional shoring systems. T his $272.99 million joint venture (JV) project was delivered using a DB project delivery approach. The planned schedule duration for this project was 1,303 workdays. The project is in its final stage of completion with minor finishing activities remaining (e.g., painting). Partnering is institutionalized in construction contracts within this state DOT. In addition, this state DOT has embraced best - value DB project delivery across this and many other projects despite being publicly funded projects. A for mal partnering process and charter were developed in this project. A neutral third - party facilitator worked with the project team to develop its partnering charter, partnering mission statement, lead partnering workshops held quarterly and tracked partneri ng health using scorecards. Team building exercises were incorporated into partnering workshops to effectuate trust and goal - alignment amongst the project team. The facilitator also provided partnering meeting minutes and helped the team establish an issue resolution ladder. The issue resolution ladder , in this case study , was comprised of four decision - making levels being Field level, P roject level, District level , and E xecutive level. The project team also had access to a Dispute Re solution Board (DRB) t o help resolve issues objectively. Each level included an agreed upon escalation time ranging from one day to 14 days maximum. The partnered - project team put forward sixteen varied goals and objectives for the project. The focus ranged from project - specifi c goal aligning items such as safety for the public and project 113 participants and budget cost and schedule goals to broader community objectives. In particular, the project team wanted to ensure the local contracting community participation and Disadvantage d Business Enterprises (DBE) programs would be strengthened as part of this investment. Table 4 - 4 below describes the goal - aligning objectives for the project and strategies to monitor progress. Goals were codified to the project with signatures from repre sentative stakeholders from each organization involved in the project. Table 4 - 4 : Case Study #4 Project Charter Goals and Performance Metrics Goal - aligning Objectives Performance Metrics Safety 100% commitment to safety for w orkforce and public Quality Construct project to last and serve public interest Cost and Schedule Complete project on budget and schedule Environment Respect for environment Issue Resolution Respectfully, timely, without DRB involvement and no litig ations Decision - Making Resolve issues and decisions at lowest level possible Partnership Partner and maintain relationships with all stakeholder (e.g., project, public, local government, railroads, permitting agencies) Relationships Trust and respec t all project stakeholders Focus on Project First Attitudes of success and fairness with focus on project first Risk Management Anticipate issues and mitigate risks early with clear plans in place Quality Management Leverage independent quality managem ent to ensure project exceeds design life Communication Open and effective communications Team Empowerment Empowered team focused on project goals Recognition Strive for award - worthy quality, aesthetics, and partnering Community Involvement Grow D isadvantaged Business Enterprise (DBE) program and local contractor participation Traffic Maintenance Effectively manage maintenance of traffic issues Comments from Case Study #4 Survey Respondents nting. The responses to this survey are based on the project as a whole. The negative responses are reflective of the issues 114 evident. There have been challenges in painting the bridge, but hopefully , those issues Design - build Team (DBT)) need to totally buy - in to partnering. One person not committed to it will s ink the entire effort. Both parties need - not an everyday roll in most projects. My responsibilities meant working w ith the owner, contractor, designer and Independent Quality Firm - all of which I found put tructors, Quality Control ( QC ) , Quality Assurance ( QA ), and Quality Oversight (QO). Too much overlap, especially between the QC, QA , and QO The team functioned well together on almost all issues. The painting issue was challenging and seemed to be handled differently regarding collaboration and open communication. Because of the pace/nature of work, most communications/information is shared electronically. It is very important to have information technology ( IT ) involved througho ut the project duration to ensure the output from various software used is transmitted seamlessly to all. 115 This $3.1 million traditional DBB project was part of a holistic plan for a major Port Authority to upgrade and improve the con ditions of three - miles of seawall constructed over 100 years ago. A 257 workday duration was established for the project team to complete the scope. The scope for this project was to complete roof repairs and modifications on a prominent Port Pier which se rves tourist and local businesses. The Pier project required hazardous material abatement, demolition, subsurface investigations, and repairs to accommodate the new roof system and surface coatings. Some challenges associated with the project were to wor k within close proximity to tenants occupying local businesses, managing pedestrian traffic around the worksite, and mitigating environmental impacts. There was also a concern to avoid any impacts on a protected migratory bird population nesting during the construction period. Partnering requirements are institutionalized in construction contracts by this local government, thus were employed on this case study project. An informal partnering charter, workshops, partnering meeting minutes, benchmarking sur veys to monitor performance, and an issue resolution ladder w as incorporated into the project via a neutral third - party facilitator. Three partnering workshops were held by the facilitator to parse out key project risks, performance goals, and success fact ors for the team. The results from the initial workshop were utilized to inform goal - aligning objectives (see Table 5 - 5 ) found within the ir informal partnering charter. 116 Table 4 - 5 : Case Study #5 Project Charter Goals and Perform ance Metrics Goal - aligning Objectives Performance Metrics Safety No injuries or accidents Maintain a clean and safe project site Update local business near project of progress regularly Schedule Achieve scheduled milestones Substantial completion 11/14/17 Final completion 12/14/17 Budgeted Cost Less than 5% change orders RFIs/Submittals Exceed contract specified time (one week or less) All processed within two weeks with no exceptions Environmental Impacts No environmental non - compliance re ports Quality Control/Assurance Weekly inspections and cleanliness No quality non - compliance reports Communication/Coordination Clearly defined communication chart followed Public Relations Excellent relationships with local business near project si te No complaints or disruptions Teamwork Plan for long - term relationships Nine goal - aligning objectives were identified by this project team along with clear performance metrics for each. Additionally, responsible leads were tasked with monitoring an d helping to assure these goals wer e achieved during the project. Notwithstanding project goals, the project team committed to an issue resolution ladder. The issue resolution ladder, as with case study #2, included five decision - making levels and time e scalation processes. This also includes a similar strategy to engage a Dispute Resolution Advisor during the issue resolution process. Comments from Case Study #5 Survey Respondents and added them 117 he y were pro - active and solution - o The final case study project investigated in this study involved renovations to a m useum b uilding and its e xhibits. The budgeted cost for the project was $5 million with an anticipated work duration of 270 workdays. The project was delivery as using a DBB project delivery method. Given that this was an existing museum facility, there were many challenges faced by the project team. The exhibits were temporarily relocated to allow continued c ommunity enjoyment and use during construction. The scope of the project was to provide a complete interior renovation and upgrades to the museum. Some of the improvements were increased programmable space within the facility for events or classes , interac tive habitat - based exhibits, and transforming the museum into a learning space for visitors and the broader community. Several sustainable design elements were implemented in the project. They used recycled/reclaimed materials and wood, energy - efficient lighting, and low - flow plumbing fixtures in bathrooms. Other environmental friendly strategies such as locally sourced materials, recycling construction waste, and reusing materi als in the project reclaimed during construction demolition activities. The project was jointly funded by S tate grants, City Parks and Recreation budgeted dollars, and community fundraising support (i.e., collected by a non - profit organization that provides funds for museum activities and improvements). Considering this, m yriad st akeholders 118 were actively engaged in the design , daily oversight, construction project spending and scheduling, project partnering , and other decision - making processes. The non - profit organization maintained an ongoing wish list with targeted fundraising ac tivities to support their added scope. Partnering was implemented on the project to bring the stakeholders together around common goals and objectives. The project team instituted an informal partnering charter developed through partnering workshops. The workshops were led by a neutral third - party facilitator that maintained partnering meeting minutes and scorecards designed to monitor performance goals. The facilitator also worked with the team to formulate an issue resolution ladder with five levels. The se issue and decision - making levels, like the other case study projects, comprised a Field level, Project level, Program level, Senior Program level, and an Executive Leadership level. The partnering facilitator was offered as a mediator when requested by the project team. As with most issue resolution ladders, the project team was bent on containing issues and resolutions at the lowest level possible. The project team, with guidance from the facilitator, developed clear goals for the project. These beca me goal - aligning objectives within their partnering charter and were used as performance metrics during partnering workshop evaluations. Their partnering charter and goal - aligning objectives are shown i n Table 4 - 6. 119 Table 4 - 6 : Case Study # 6 Project Charter Goals and Performance Metrics Goal - aligning Objectives Performance Metrics Safety /Environmental Protection No i ncidents or accidents No non - compliance notices Schedule Achieve substantial completion by 9/1/16 Achieve final completion by 10/1/16 Budgeted Cost Within budgeted cost of all funding sources Minimize change orders Funding stakeholders satisfaction with project Communication/Documentation Smooth and timely process Quality Meets or exceeds design expec tations Aesthetically pleasing to all stakeholders Public Satisfaction Total public satisfaction, excitement, and enjoyment Communication/Coordination Clearly defined communication chart followed Teamwork Plan for long - term relationships The pro ject absorbed many challenges and risks as a result of several factors. One was an u nanticipated electrical design and utility service improvement. The new electrical service was required by the local governing authorities and caused significant cost and s chedule impacts. Another identified risk was managing the multitude of stakeholders with competing interests. This risk posed schedule concerns because payments were originating from different sources at varying times in the project. Thus, the project team had to be cognizant to monitor and control spending across the project duration . The case study projects we re detailed in the proceeding sections. Key project characteristics and partnering attributes were also summarized to illustrate how partnering wa s followed. Th e next section provides a pattern - matching of the case study projects. Comments from Case Study #6 Survey Respondents we were contracted with that had an in fluence on the outcome. This, along with the 120 issues with the electrical and fire line services were the primary complications on the - a quality built project , and an enjoyable experience of the team - w ere in large part due to the personal commitment of the The primary aim of this research is to explore the relationships between project risk factors , collaborative project delivery practices , goal alignment , transactive memory systems (TMS), and performance outcomes in AEC project teams. Using qualitative data, this section provides a pattern - matching of case study projects. It begins by examining the broader characteristics of the case studies investigated in this study. Then, it illustrates some commonalities and deviations within the data to help show relationships among the case study projects. D ata collected from six partnered case study projects w ere examined wi thin this study. Projects varied in size, project type, complexity, and duration . Based on some of these factors, the number of people participating in partnering processes differed acros s projects. They were also located in two distinct regions in the U.S., West coast and the Midwest. Table 4 - 7 shows the background characteristics for each case study project including project size, scheduled duration, project type, location, number of pa rtnering participants, and number of partnering workshops . Based on this analysis, it shows that while projects are limited in regions, a good mixture of projects w as represented in the analyses. For example, 121 projects ranged in size s with two micro ($0 - $5M ), two small ($5 - 10M), one large ($ 25 - 250M), and one very large /mega (>$250M) according to IPI vertical partnering matrix. Although, in horizontal partnering matrix projects ranging in sizes from $10 - 250M are combined into the large category or level 3 (IPI, 2017a) . Table 4 - 7 : Characteristics of Case Study Projects Project Size ( * $M) Schedule (* * Workdays) ** * Project Type Location No. of Partnering Participants No. of Partnering Workshops Case Study #1 5.59 345 Horizontal West 6 3 Case Study #2 4.96 255 Large Infrastructure West 17 3 Case Study #3 1 49.96 595 Horizontal Midwest 30 9 Case Study #4 272.99 1303 Large Infrastructure Midwest 52 10 Case Study #5 3.10 257 Vertical West 9 3 Case Stu dy #6 5.00 270 Vertical West 11 3 The projects examined in this study were also grouped by project types being: two horizontal, two large infrastructure projects, and two vertical pro je cts . When looking at the number of partnering participants it is clea r that larger project included a greater number of participants. This is attributable to the number of contractors and stakeholders involved in the construction project. Similarly , the partnering workshops for projects less than $10 million in size were li mited to three. These workshops, according to partnering documents, included a partnering workshop at the initial project kick - off, one midway thru the project duration, and a * $M U.S. dollars in millions; ** Workdays excludes holidays and weekends ; ** * Project types included among others vertical (e.g., office buildings), horizontal (e.g., roadways), and large infrastructure (e .g., tunnels, bridges , or major highway infrastructure) 122 final workshop at or near project completion. Meanwhile, those above $10 million held partnering workshops on a quarterly basis. Partnering cost as a percentage of the original contract is given next. The cost of partnering ranged from 0 .01% of the original contract value to 0 . 48%. From Table 4 - 8 it is obvious that partnering costs a re minor as compared to the overall construction budgeted cost. Table 4 - 8 : Case Study Partnering cost as a percentage of original contract value. Partnering Cost ($) Orig. Cost ($M) Partnering Cost as a Percentage of Orig. Co st (%) Case Study #1 4 , 000 5.59 0.07 Case Study #2 8 , 500 4.96 0.17 Case Study #3 20 , 000 1 49.96 0.01 Case Study #4 19 , 000 272.99 0.01 Case Study #5 15 , 000 3.10 0.48 Case Study #6 10 , 000 5.00 0.20 A brief inspection of project delivery approach follo wed in each case study project showed that DBB was most prominent. This approach accounted for five of six cases in the data while case study #4 was delivered using DB. Next, cost and schedule performance outcomes are presented. The case studies were also examined to understand cost and schedule growth. To do so, the original and actual cost/schedule informat i on was collected from project participants. Two of 123 the projects are ongoing, thus, actual cost information is te ntative (i.e., case study #3 and #4). Cost growth for micro/small projects is shown i n Figure 4 - 1 . Figure 4 - 1 : Cost Growth for Micro - Small Projects or less than $10M Evaluating the cost growth change from original to ac tual cost by case study it demonstrates that case study #1 and case study #5 experience minor cost growth, 5% percent and 6% percent respectively. Case study project #2 has moderate cost growth at 13% percent while case study project #6 found significant c ost growth being 26% percent. According to project team members, this is directly correlated with a major change in scope for the project. An unplanned electrical design change and new electrical utility service upgr ade increased the project cost. The la rge/mega projects were also evaluated to understand cost growth based on original versus actual cost. Figure 4 - 2 illustrates the results of two case study projects which Orig. Cost ($M) , 5.59 Actual Cost ($M) , 5.88 Orig. Cost ($M) , 4.96 Actual Cost ($M) , 5.61 Orig. Cost ($M) , 3.10 Actual Cost ($M) , 3.30 Orig. Cost ($M) , 5.00 Actual Cost ($M) , 6.30 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 Orig. Cost ($M) Actual Cost ($M) Case Study #1 Case Study #2 Case Study #5 Case Study #6 124 were greater than $25 million. Both projects are seeing minimal cost growth or 3% perc ent (Case study #3) and 1% percent (Case study #4). Although important, both projects are ongoing which may affect these findings. Despite this, it is anticipated case study #4 will hold pretty consistent since they are in the final stage of completion onl y doing minor finish work (e.g., surface painting). Figure 4 - 2 : Cost Growth for Large - Mega Projects or greater than $25M Schedule growth was also examined across all case study projects in this study. The analysis looke d at scheduled durations in workdays (i.e., calendar days excluding holidays and weekends). In addition, this analysis considered the original planned duration for each project and compared it against actual project durations. The results of this analysis are shown i n Figure 4 - 3. Orig. Cost ($M) , 149.96 Actual Cost ($M) , 155.00 Orig. Cost ($M) , 272.99 Actual Cost ($M) , 275.60 0.00 50.00 100.00 150.00 200.00 250.00 300.00 Orig. Cost ($M) Actual Cost ($M) Case Study #3 Case Study #4 125 Figure 4 - 3 : Schedule Growth fo r Case Study Projects based on Workday D urations Based on the analysis, three case studies exper ience d considerable schedule savings. Case study #5 was completed in ha lf the anticipated schedule duration even considering getting off to a late start. According to the project team, a hazardous material abatement contractor brought onboard for the project was replaced due to lack of experience. This change delayed the proj ect by three months, yet the project still finished ahead of schedule. This case study project also reported a scope add during construction of an additional Pier renovation. They reported this project also finished ahead of schedule. Case study #4 also fi nished considerably ahead of schedule. The project team attributed their success to high levels of collaboration and coor dina tion . This schedule improvement was due, in part, to 24 - hour workdays to make up for delays associated with downtown events and red esigns to shoring systems. The schedule growth for the case studies as a percentage are as follows: 10 % percent (C ase study #1), - 28% percent (C ase study #2), 0% percent or on target (Case study #3), - 19% 345 255 595 1303 257 270 378 183 595 1050 128 759 0 200 400 600 800 1000 1200 1400 Case Study #1 Case Study #2 Case Study #3 Case Study #4 Case Study #5 Case Study #6 Actual Schedule (Workdays) Orig. Schedule (Workdays) 126 percent (Case study #4), - 50% percent (Case study #5), and 181% percent (Case study #6). These results are shown in Table 4 .9 which communicates both cost and schedule growth, ranked on best overall project performance. Table 4 - 9 : Overall Project Performance Ranking Cost G rowth Schedule Growth * Rank Case Study #4 1% - 19% 1 Case Study #3 3% 0% 2 Case Study #5 6% - 50% 3 Case Study #2 13% - 28% 4 Case Study #1 5% 10% 5 Case Study #6 26% 181% 6 As previously mentioned, case study #6 saw drastic increases in their schedule due to unanticipated work. However, case study #1 also experienced a small increase in their completion schedule. According to the owner team, this was explained by a conti n gency add which in creased the storm water sewer replacement from six to eight city blocks. The primary aim of this research is to explore the relationships between project risk factors , collaborative project delivery practices , goal alignment , transactive memory systems (TMS), and performanc e outcomes in AEC project teams. This section provides a cross - case synthesis of case study projects based on qualitative data. It asserts the study propositions and delves into partnering attributes and perceptions according to key metrics put forward in Chapter 3. From this analysis , it demonstrates how each proposition is supported or rejected. * Overall Project Performance Ranking based on Lowest Cost and Schedule Growth Comparison. 127 The qualitative analysis in this section addresses the following questions by examining partnered - project level data in the form of partnering charters, partne ring meeting minutes, scorecards, and partnering workshop evidence. These data were inspected alongside structured interviews with project stakeholders. This allowed the researcher to investigate t he links among project risk factors , collaborative project delivery practices, and project performance. T hree propositions were utilized to qualitatively test these assumptions: Proposition 1: Project risk factors change the level of collaborative project delivery practices implemented in partnered - projects. Pro position 2 : Collaborative project delivery practices are directly related to alignment perceptions in partnered - projects. Proposition 3 : directly related to project performance in partnered - proje cts. The propositions in this study are addressed by inspecting case study evidence and drawing conclusions from emerging patterns. Trends within these data are also utilized to make assertions in supporting or rejecting study propositions. The case studies are initial ly evaluated in a side - by - side comparison for each variable i.e., project risk factors and collaborative project delivery practices . These comparisons also include project characteristics (e.g. , project cost, project duration, project type, and project delivery method) , overall project performance rankings , and critical analysis to add richness and context to the results from the se analyses. 128 Table 4 - 10 combines all six case study projects to p ermit an evaluation of project risk factors . An interesting pattern emerging from this figure. Case study projects with the highest rating in political significance and community interest were rated among the best performing projects. This held true across case study projects from varied sizes, durations, and project types (i.e., horizontal and infrastructure). Intuitively this finding is banal, in that, one would anticipate projects having potential long - term impacts on communities will strive for timely a nd successful completion. For instance, case study #3 commented during structured interviews on the disruptive nature of the construction work to traffic and neighbors living nearby as very important in daily decision - making. However, they trivialized the impact partnering had on their project success which was made more prominent during a partnering workshop . In this workshop , a noticeable friction was present. willingness to resort back to force accounts (i .e., cost of work based on time and material) due to lack of trust in pricing for added changes in scope. This is often attribute d to the project delivery method (e.g., DBB creating an environment for contractors to monopolize changes ), yet positive result s were found in other case studies following this methodology. Interestingly, case study #4 , the only Design - Build project, showed the highest overall project performance ranking. It was also one of two projects with the strongest team and partnering expe rience within the project team. This may explain why interviewees purportedly identified partnering workshops as a great tool to continuously realign the team. Given their existing relationships, it appeared easier to push aside differences in this case st udy unlike case study #3 . Similarly , case study number #2 found it challenging for team integration to take hold when project team members were not engaged in a timely manner. The project was also one 129 with fewer partnering workshops (i.e., three workshops being one at initial project kickoff, one midway through the project , and a final at the end) due to its size. Thus, it may be plausible the frequency of partnering workshops should be aligned with the project teams collective experience working together a nd/or in the partnered - project arrangement. Despite this, the interviewees from case study #2 gave high marks for the success of the project and its team members. Table 4 - 10 : Project Risk Factors and Overall Project Performanc e Rank for Case Studies. Project Risk Factors Case Study #1 Case Study #2 Case Study #3 Case Study #4 Case Study #5 Case Study #6 Project Size ($M) 5.59 4.96 149.96 272.99 3.10 5.00 Project Duration (Workdays) 345 255 595 1303 257 270 Project Type Hori zontal Infrastruc. Horizontal Infrastruc. Vertical Vertical Project Risks* 1.5 3.0 3.0 3.0 1.0 3.0 Schedule Risks* 1.5 3.0 3.0 3.0 2.0 2.0 Project Team Relationships** 1.5 2.0 2.0 1.0 1.0 3.0 Partnering Team Experience*** 1.5 1.5 2.0 2.0 1.5 2.5 Polit ical Significance and Community Interest**** 3.0 4.0 4.0 4.0 4.0 3.5 Complexity**** 3.5 4.0 4.0 4.0 2.0 2.0 Project Delivery Method DBB DBB DBB DB DBB DBB Overall Project Performance Rank 5 4 2 1 3 6 Next, the case studies are reordered based on the ir overall project performance ranking to help clearly illustrates other emerging trends within these data. *1 - Few to 3 - Many; **1 - Strong team and partnering experience to 4 - No team or partnering experience ; ***1 - Exp erienced to 3 - No Experience; ****1 - Standard to 4 - High 130 Table 4 - 11 shown below similarly combines project risk factors , yet aligns projects based on their overall project performance rank. Another patter n obs erved from this analysis is the notion that many project and schedule risks are also largely seen by project s with higher project success. Some of those risks permeated their way into case study partnering charters. Case study # 2 specified a clear tur nover date and time as a metric for the team knowing how vital it was to bring the water infrastructure back online. They also included goals formulated around safety and health considerations. Given the inherent risk of working inside confined spaces such as tunnels, it was no surprise the team made worker safety a prominent goal in their partnering charter. Project and schedule risk as these tend to drive a more concerted team effort in projects. Although, this benefit gets attenuated when liquidated dama ge clauses for delays or fines for safety incident s become part of contracts. Meanwhile, case study projects with the highest level of complexity are also among those associated with increased project performance. Projects high in complexity (e.g., many s takeholders and downstream impacts if not projects are not delivered on time) tended to find ways to quickly align their goals for the benefit of the project. It should be noted that p roject team experience and partnering experience for the top four projec ts were relatively strong (i.e., ranges from 1.0 to 2.0 with one being a strong team and partnering experience ) across these cases. As previously mentioned, this offers insight as to why these projects were able to find greater successes. 131 Table 4 - 11 : Comparing Project Risk Factors among Overall Project Performance Rank Project Risk Factors Case Study #4 Case Study #3 Case Study #5 Case Study #2 Case Study #1 Case Study #6 Project Size ($M) 272.99 149.96 3.10 4.96 5.59 5.00 Project Duration (Workdays) 1303 595 257 255 345 270 Project Type Infrastruc. Horizontal Vertical Infrastruc. Horizontal Vertical Project Risks* 3.0 3.0 1.0 3.0 1.5 3.0 Schedule Risks* 3.0 3.0 2.0 3.0 1.5 2.0 Project Team Relationships** 1.0 2.0 1.0 2 .0 1.5 3.0 Partnering Team Experience*** 2.0 2.0 1.5 1.5 1.5 2.5 Political Significance and Community Interest**** 4.0 4.0 4.0 4.0 3.0 3.5 Complexity**** 4.0 4.0 2.0 4.0 3.5 2.0 Project Delivery Method DB DBB DBB DBB DBB DBB Overall Project Performanc e Rank 1 2 3 4 5 6 Enhancing t he results, Table 4 - 12 brings in a comparison between the case study projects with the highest and lowest rankings. Results from this analysis demonstrate differences in both complexity and project team relationships, yet cases closely mirror one another in project risks. However, similar patterns emerge among political significance and community interests, project risks, a nd partnering team experience. One critical observation is how case study #6 reported many project ri sks and higher visibility for political significance and community interests while having drastically different outcomes. During structured interviews, the interviewees commented on how the project took on new challenges with added scope. *1 - Few to 3 - Many; **1 - Strong team and partnering experience to 4 - No team or partnering experience ; ***1 - Experienced to 3 - No Experience; ****1 - Standard to 4 - High 132 The additional wo rk (i.e., new electrical service requirement from utility) caused significant cost and schedul e impacts on the project which affected overall project performance perceptions. This is explained further as part of proposition testing section . Table 4 - 12 : Comparison of Highest and Lowest Overall Project Performing Case Study. Project Risk Factors Case Study #4 Case Study #6 Project Size ($M) 272.99 5.00 Project Duration (Workdays) 1303 270 Project Type Infrastructure Vertical Project Risks* 3.0 3.0 Schedule Risks* 3.0 2.0 Project Team Relationships** 1.0 3.0 Partnering Team Experience*** 2.0 2.5 Political Significance and Community Interest**** 4.0 3.5 Complexity**** 4.0 2.0 Project Delivery Method DB DBB Overall Projec t Performance Rank 1 6 The project risk factors from these case studies are also investigated for patterns by looking for trends or commo nalities and/or differences between project types. Table 4 - 13 is us ed to illustrate these attributes. A rich patte rn emerges w hen controlling for project type ( e.g. , infrastructure, horizontal, and vertical) against overall top - rated project performance across four case studies. Results demonstrate that both high performing case studies share similar risks perceptions while emanating from two different project types. For instance, both case study #3 and #4 equally reported increased project and schedule risks which can be attributed to the similarly of construction ( e.g ., highway and bridge repair *1 - Few to 3 - Many; **1 - Strong team and partnering experience to 4 - No team or partnering experience ; ***1 - Experienced to 3 - No Experience; ****1 - Standard to 4 - High; Indicates similar attributes 133 projects with many a ctive traffic routes nearby). Meanwhile, the lowest performing group differ across each risk category. What this illustrates is that both case st udy number one and six may have misalignment in their level of collaborative partnering. Especially considering that case study number one had a known level of community interest and increased complexity. With regards to case study #6 , there were many risks in the project such as working with tives. They also had a project team with limited experience working together as shown by their ratings for project team relationship (i.e., 3.0 Team has no prior experience working together but has partnering foundation ) and partnering team experience (i .e., 2.5 Some experience) . Table 4 - 13 : Comparison of Project Type and High/Low Overall Project Performing Case Study. Project Risk Factors Case Study #4 Case Study #3 Case Study #1 Case Study #6 Project Size ($M) 272.99 149.96 5.59 5.00 Project Duration (Workdays) 1303 595 345 270 Project Type Infrastructure Horizontal Horizontal Vertical Project Risks* 3.0 3.0 1.5 3.0 Schedule Risks* 3.0 3.0 1.5 2.0 Project Team Relationships** 1.0 2.0 1.5 3.0 Partnering Team Exper ience*** 2.0 2.0 1.5 2.5 Political Significance and Community Interest**** 4.0 4.0 3.0 3.5 Complexity**** 4.0 4.0 3.5 2.0 Project Delivery Method DB DBB DBB DBB Overall Project Performance Rank 1 2 5 6 *1 - Few to 3 - Many; **1 - Strong team and partnering experience to 4 - No team or partnering exp erience ; ***1 - Experienced to 3 - No Experience; ****1 - Standard to 4 - High; Indicates similar attributes 134 Like case study project #2, this amplifies an instance where early team member involvement (e.g., public utility officials) and more frequent partnering workshops may have staved off the late timing for the added scope. The next section provides an analysis used to test study propositions. To test the propositions in this study the theoretical framework is used. Three propositions are tested in this study: Proposition 1: Project risk factors and the level of collaborative project delivery practices in partnered - projects are positively related ; Proposition 2: perceptions in partnered - projects are positively related ; and, Proposition 3: oject performance in partnered - projects are positively related . Figure 4 - 4 recaps the metrics investigated during those interviews. The response risk influences t heir case study project. Then, a Likert scale is used to further discern its level of importance in each case study. A similar methodology was followed with respect to collaborative project delivery practices . Key case study participants selected approp riate responses which were used to determine how collaborative practices were executed within their case study project. The evaluation method for this category is also shown in Figure 4 - 5. There were three distinct 135 categories from which structured intervie w respondents could select. Those category options were contractual related practices, procurement related practices, and project related practices. F ig ure 4 - 4 : Common Project Risk Factors Evaluation Method 136 Based on the theoretical framework in F igure 4 - 6 , a direct relationship is present among variable s using evidence from each case study project. Data supporting this framework are based on perceptions among individual s from each case study project along with case study project documents such as partnering charters, partnering scorecards, and insights f rom structured interviews. Figure 4 - 5 : Collaborativ e Project Delivery Evaluation Method 137 Figure 4 - 6 : Theoretical Framework for Relationships among Project Risk Factors, Collaborative Project Delivery Practices, Goal Alignment, and Project Performance. These data were analyzed to gene rate a score for e ach variable shown in Figure 4 - 6 . The Few=1, Moderate=2, and Many=3) which is multiplied by the rating of the item using a Likert scale ranging f rom 1 - Strongly disagree to 5 - Strongly agree . These data were weighted based on the number of choices available i.e., five response options were equally weighted as 0.2 and multi plied by the actual response. T his value is, then, converted to a 100 point sc ore by multiply by 100 and divided by the total number of response s for each category . The scores shown in Table 4 - 14 are sorted by case study number and used to test the three propositions presented in this study. This table shows the results from struct ured interviews used to capture project risk , collaborative practices , project performance , and goal alignment perceptions from each case study project. Additionally, it gives information gathered from case study partnering documents such as partnering cha rters, partnering scorecards, and insights from structured interviews. Goal Alignment Project Level Data / Qualitative Investigation P2 P3 P1 Collaborative Project Delivery Practices Project Performance Risk Factors 138 Table 4 - 14 : Case Study Scores used in Proposition Testing sorted by Case Study Number Structured Interviews Case Study Partnering Documents Case Study P rojects Project Risk Factor Score Co llaborative Practice Score Project Performance Score Goal Alignment Score Project Charter Goals Goal Alignment Actions Case Study #1 44 43 88 90 9 23 Case Study #2 80 71 100 92 8 13 Case Study #3 75 41 85 71 13 49 C ase Study #4 64 64 85 78 10 43 Case Study #5 44 43 100 93 9 23 Case Study #6 57 36 55 93 7 14 Proposition 1: Project risk factors and the level of collaborative project delivery practices in partnered - projects are positively related . The first propositi on (i.e., P1 in Figure 4 - 4) in this study investigates the relationship between project risk factors and collaborative project delivery practices. Project risk factors were measured during structured interviews with key owner representatives using Likert s cale ratings and analyzed according to the aforementioned analyses above. The results from structured interviews and scorecards used to capture project risk factors , collaborative project delivery practices , project performance , and goal alignment percepti ons from each case study project are shown in Table 4 - 15. The table highlights the relationship between these the two variables in proposition 1 using score s computed from structured interview and scorecard data . The table is sorted on project risk score s from highest to lowest against collaborative project delivery practice scores. 139 Table 4 - 15 : Results from Structured Interviews and Project Scorecards sorted by Project Risk Scores Structured Interviews Case Study Projects P roject Risk Factor Score Collaborative Delivery Practices Score Project Performance Score Case Study #2 80 71 100 Case Study #3 75 41 85 Case Study #4 64 64 85 Case Study #6 57 36 55 Case Study #5 44 43 100 Case Study #1 44 43 88 A noticeable p attern exists regarding project risk factor scores for each case study represented in the table. Case study #2 purportedly has the highest project risk factor score (i.e., 80) while case studies #1 and #5 both received the lowest project risk factor score in these data (i.e., 44). Juxtaposing project risk factor scores with collaborative project delivery practice scores, it appears four case study projects have clearly aligned their collaborative project delivery practices with project risk factors . In othe r words, when project risk s such as complex design and construction, compressed schedules, and uncommon materials are perceived as low , the importance of collaborative practices is minimized. In fact, case study #1 reported the ability to take on the addit ional scope in the form of another two blocks of water main replacement and surface repairs associated with the low risk to the project schedule and budget. Two of the six cases do not support the proposed relationship in proposition 1: Case s tudies #3 a nd #6. Interestingly Case study #6 has the lowest project performance score while case study #3 also has one of the lower performance scores. These results led to the development of a new proposition: 140 Proposition 1a - If collaborative project delivery prac tice s are not positively aligned with the level of project risk in partnered - projects, then project performance will be negatively affected . Table 4 - 16 is used to explicitly probe project risk factors identified and scored by case study project participant s. The case studies are shown in order based on their overall risk scores. In addition, the specific project risk factors and characteristic s of each case study are shown. Table 4 - 16 : Project Risk Factors and Overall Risk Scor es Case Study #2 Case Study #3 Case Study #4 Case Study #6 Case Study #5 Case Study #1 Project Risk Item Score Project Risk Factors Partnering Team Experience 100 90 90 70 90 90 88 Political Significance and Community Interest 100 90 80 80 30 25 68 Schedule Risks 90 90 90 30 40 44 64 Project Team Relationships 100 100 60 67 27 20 62 Complexity 45 40 47 42 50 45 45 Project Risks 45 40 20 53 25 38 37 Overall Risk Score 80 75 64 57 44 44 61 Project Characteristics Project Size ($M) 4.9 6 149.96 272.99 5.00 3.10 5.59 Project Duration (Workdays) 255 595 1303 270 257 345 Project Type Infrastruc. Horizontal Infrastruc. Vertical Vertical Horizontal Project Delivery Method DBB DBB DB DBB DBB DBB Despite the variation within project risk factor scores, five of six cases scored highly on partnering team experience. In Table 4 - 16 overall project risk scores are highest (e.g., ranged 141 from 90 - 100) when project teams purportedly have limited partnering experience. This may serve as an earl y warning to project teams that collaborative project delivery practices are not aligned or s hould be equally aligned . Meanwhile, projects with high visibility due to the impact on surrounding communities also maintain higher risk scores. This trend is mo st prominent in case study #2 where the project had increased exposure because it involved shutting down a regional water source for the community. Thus, the stakes for the project team were heightened which resulted in increased awareness and adhered to p roject goals outlined in the partnering documents. One particular item was the inclusion of advance inspections to the tunnel lining prior to repairs and construction. Unlike case study #6, there was a clear plan to deal with unforeseen conditions to help mitigate schedule impacts. The project team reported completing the project well ahead of schedule due largely to this single item. The second highest project risk factor among these cases was political significance and community interest. Looking at this closer both case study #3 and #6 appeared disjointed from their collaborative project delivery practices , yet they ranked this risk factor as relatively important. Again, maybe project teams should place more stock in collaborative project delivery practi ces when risk factors are known. Especially those regarding limited partnering experience and newer relationships, political significance and community interest, and potential schedule risks. Further evidence supporting proposition 1 is given in Table 4 - 17. Based on this table, another interesting revelation arises relating to case study #3 and #6. Both case studies noted 142 the inclusion of incentive s, fees, risk - reward, or gainshare - painshare agreement s in contracts. During structured interviews , it was di scovered that these were specific to liquidated damages intending to account for late project delivery . Neither case study reported the necessity to enforce liquidated damages, however, it should be noted that this tends to exacerbate project risk percepti ons when not seen as equitable in multiparty agreements. Using the above analyses, this study asserts limited support for the relationship between project risk factors and collaborative project delivery practices suggested in proposition 1. Several prac tical applications and theoretical implications are listed next. Practical Applications When project risks are relaxed collaborative project delivery practices are easily aligned. Limited partnering team experience and projects with high visibility are e arly warning signs for increased collaborative practices. Avoid disincentives such as liquidated damages unless subsequent incentives or risk sharing arrangements are included in contracts. 143 Table 4 - 17 : Collaborative Project De livery Practices sorted by Overall Score Collaborative Project Delivery Practices Case Study #2 Case Study #4 Case Study #5 Case Study #1 Case Study #3 Case Study #6 Contractual Related Practices Professional facilitator was used in this project. A shared equity arrangement was indicated in contracts. A partnering charter was used in this project. A proactive conflict management tool that added structure to collaborative problem - solving processes was used in this project. Equal power/empowerment was afforded to all project teams and team members in decision - making processes. An incentive/fee/risk - reward/ or gainshare - painshare agreement was established in contracts. Procuremen t Related Practices Parties were selected based on partnering experience. We selected team members based on previous work experience with other team members. Parties were selected based on technical expertise. There wa s early involvement of key participants (e.g., designer/contractor/specialty subcontractors) during schematic design (SD). Project Related Practices Partnering workshops were held for this project. Partnering scorecards were used in this project. There were two or more project teams located together in a common office (i.e., colocation). Measurable and achievable milestones were established to determine the success of the project. Partnering train ing/team - building sessions were held for this project. Project teams openly exchanged information across organizational boundaries (e.g., Building Integrated Modeling (BIM)) *Quarterly partnering meetings were used in this project . Multi - tiered partnering was used in this project (i.e., executive, core team, stakeholders) Specific task force used for conflict and issue resolutions Overall Collaborative Practice Score 71 64 43 43 41 36 144 Theoretical Im plications Departing from literature, early involvement of key participants ( e.g., designer/contractor/specialty subcontracto rs) during schematic design) did not appear significant in this study ( Baid en et al., 2003; Cheng & Li, 2002; Pishdad - Bozorgi & Beliveau, 2016) although this was alluded to by project participants as important. Colocation, a common collaborative project delivery practice, was identified in only one case study despite tremendous benefits asserted in the literature (Baiden et al., 2003; Pishdad - Bozorgi & Beliveau, 2016) . Relational governance structures are best when incentives are included in contracts, yet disincentive s may disparately affect performance outcomes (Lu et al., 2015) . Proposition 2: perceptions in partnered - projects are positively related . This propos ition examines the direct relationship between collaborative p roject delive ry practices and goal alignment as presented in Figure 4 - 4 . To examine this relationship, this study uses partnering documents (e.g., partnering charters, workshop meeting minutes, and partnering scorecard surveys) and data from structured int erviews. The results from these are highlighted in Table 4 - 18. From initial inspection, no discernable pattern emerges among these data. When collaborative project delivery practices are at the highest rating, goal alignment scores are also rated the m hig h est . The same is true for the lowest scores for these two categories. Despite this inconsistent pattern, another al ternative explanation persist. 145 Table 4 - 18 : Results from Structured Interviews and Project Scorecards sorted by Collaborative Project Delivery Practice Scores Structured Interviews Case Study Partnering Documents Case Study Projects Project Risk Factors Score Collaborative Project Delivery Practices Score Project Performance Score Goal Alignment Score* Project Charter Goals Goal Alignment Actions Case Study #2 80 71 100 92 8 13 Case Study #4 64 64 85 78 9 23 Case Study #5 44 43 100 93 7 14 Case Study #1 44 43 88 90 9 23 Case Study #3 75 41 85 71 13 49 Case Study #6 57 36 55 93 10 43 The goal alignment scores are measured using d efined goal alignment objectives and actions put forward by case study project teams. These are clearly elicited in partnering charters developed at outset of a project either during planning or early during phases of constructi on. A closer look at the number of goals in each partnering charter and number of goal alignment actions show an increasing trend as collaborative project delivery practices decrease. In other words, an inverse relationship appears between these two variab les. Case studies with higher collaborative project delivery practices appear to require fewer goal alignment actions or metrics to hold the project team accountable. This leads to a new proposition: Proposition 2a Higher collaborative project delivery practices require fewer goal alignment actions or metrics to hold the project team accountable in partnered - projects. Additional insight s are found by discriminating explicitly for high and low project risk factors , collaborative project delivery practice s , goal alignment , and project performance among these six cases. To do so, each score computed in Table 4 - 18 was converted to a high or 146 low value by center ing the score about the mean for each of the four categories. This allowed the researcher to search for variation among all categories within these data. The results are shown in Table 4 - 19. In this analysis, a pattern materializes when inspecting the relationship between collaborative project delivery practices and goal alignment from a different purv iew. When both project risk factors and collaborative project delivery practices goal alignment project performance. Next, when project risk factors and colla borative project delivery practices , they goal alignment does not immediately afford increased project performance. A goal alignment project performance perceptions. This indicates misalignment among project teams, likely attributable to poorly matching project risk factors with the collaborative project delivery practices . Especially considering the high number of goal alignment actions taken up in project charters intend ing to help project teams manage their goals and objectives. The researcher made similar observations when visiting case study #3. united despite the various collaborative pra ctices designed to do so. This left many project team members in a place of mistrust and may have led to lower than expected project performance ratings. Therefore, project teams should be mindful that collaborative project delivery practices are only effe ctive when they bring increased alignment among project teams. Practical and theoretical applications based on proposition 2 are offered next. 147 Table 4 - 19 : Results from High/Low Score Analysis sorted by Case Study Structured Interviews Case Study Partnering Documents Case Study Projects Project Risk Factors Collaborative Project Delivery Practices Project Performance Goal Alignment Project Charter Goal Alignment Actions Case Study #1 Low Low High High 9 23 Case Study #2 High High High High 8 13 Case Study #3 High Low Low Low 13 49 Case Study #4 High High Low Low 10 43 Case Study #5 Low Low High High 9 23 Case Study #6 Low Low Low High 7 14 Practical Applications Project teams should be cautious when many goal aligning action surface during partnering workshops; Leading indication that project team may be strained to align individual goals with those outlined in the project charter Project deemed as having reduced project risk and following a limited number of collaborative practices should not fall into complacency; D espite the ability to keep their teams aligned , these projects are susceptible to undesirable project performance outcomes Theoretical Implications Social loafing arises in this proposition where project teams or members may take a backseat because the number of goal alignment actions becomes overwhelming (Harkins, 1987; Lam, 2015) . 148 Larger project teams may also succumb to the social loafing phenomenon during partnering workshops and not fully commit to the partnering process (Lam, 2015) . Goal alignment influences performance outcomes when feedback is present which is missed when project teams or individuals are not engaged in partnering and its feedback processes [ i.e., worksh ops, scorecards, meetings, etc.] (Earley, 1990; Lahdenperä, 2012; Pishdad - bozorgi & Beliveau, 2016) . Proposition 3: ormance in partnered - projects are positively related . The third propos ition examines the direct relationship between goal alignment and project performance as presented in Figure 4 - 4 . When examining the relationship between goal alignment and project perfo rmance a clear trend is present. Table 4 - 20 below , again, shows the results from structured interviews used to capture project risk factors , collaborative project delivery practices , project performance , and goal alignment p erceptions from each case study project. 149 Table 4 - 20 : Results from Structured I n terviews and Project Scorecards sorted by Goal Alignment Scores Structured Interviews Case Study Partnering Documents Case Study Projects Project Risk Factor Score Coll aborative Delivery Practices Score Project Performance Score Goal Alignment Score* Project Charter Goals Goal Alignment Actions Case Study #5 44 43 100 93 9 23 Case Study #6 57 36 55 93 8 13 Case Study #2 80 71 100 92 13 49 Case Study #1 44 43 88 90 10 43 Case Study #4 64 64 85 78 9 23 Case Study #3 75 41 85 71 7 14 Table 4 - 20 highlights the relationship between goal alignment and project performance . Goal alignment scores are gathered from individual ratings against specific goals outli ned and monitored as part of their project charters . Figure 4 - 7 presents a sample scorecard with resu lts Figure 4 - 7 : Sample Case Study Proje ct Scorecard, Goal Aligning Actions, and Performance Metric 150 from one of the case study project charters investigated in this study. Each case study project followed a similar strategy which includes action items and performance metrics for the team to use in their ratings. trend is illustrated connecting goal alignment to project performance. Generally, when goal alignment is high project perfo rmance also receives a high score (see Table 4 - 20) . Case study #6 appears inconsistent with the trends. In this case study , the team experience d a project set - back resulting from the added scope for an unforeseen code requirement. An electrical service wa s required to be upgraded as part of the renovation work. This required significant communication and coordination with the electric utility, designers, owners, and contractors. As a result, the project was delivered over two years later than original ly pl anned. According to partnering documents, the project was financially constrained due to multiple funding sources and spending stipulations (i.e., 45 percent State grant, 30 percent City Parks and Recreation Department , and 25 percent n ot - for - profit fundra ising by the organization). This can limit the amount of resources available for early site investigations. guidance during the design phase. As a not - for - profit organizatio n (i.e., in this case , the owner), it can be asserted that they may not have been a sophisticated buyer of construction work and may have experienced breakdowns in communication with the designer leading to this u nforeseen major scope addition. Given all t his, an explanation surfaces as to why the project maintained a high goal alignment level, yet reported a lower project performance score. In contrast case studies #2 and #5 demonstrate a clear relationship between goal alignment and project performance. 151 T he top two projects (i.e., case study #2 and case study #5) followed common metrics that differed from case study #6 . For instance, an emphasis was placed on continuously improving document management systems for processing submittals and requests for info rmation ( RFIs ). This allows information to move quickly across organizations when decisions are required. Another disparate finding among the top and lowest cases is the inclusion of a monitor to help encourage collaboration and integrated teams to develop , especially around problem - solving. These results lead to the development of a new proposition: Proposition 3a: - Collaborative delivery practices should accommodate continuous improvement in facilitating information exchange among team members in partne red - projects. Interestingly, the goal alignment actions for most case studies are similar across traditional areas of focus in the projects (e.g., schedule, safety, and cost) . A full listing of case study project goal alignment met rics is displayed i n Tabl e 4 - 21 . This table clearly illustrates the top three goal alignment metrics used as 1) schedule, 2) safety, and, 3) communication and coordination. All of the metrics shown in the table were used as a means to track and monitor how well the project teams w ere aligned with the project goals identified by the team . One of the top performing projects, case study #2, also shares goal alignment attributes with case study #3. Both projects found it important to maintain confidence in their issue resolution ladde rs. This allows the project team to focus on ways to keep decisions at the lowest possible levels to assure timely responses. They also sought means to anticipate project risks early and often. Their aims were to address the issues cohesively as a team yet resolving them at the field level. 152 Table 4 - 21 : Goal Alignment Metrics used in Case Studies sorted by Project Performance Ano ther observation from Table 4 - 21 was how the project with the greatest number of goal alignm ent metrics (i.e., case study #3 ) did not score equally high in project performance rformance over a longer time span. For example, case studies #3, #4, and #6 all lasted for longer duration s (e.g., Goal Alignment Metrics and Ratings* Case Study #2 Case Study #5 Case Study #1 Case Study #3 Case Study # 4 Case Study #6 # of Case Studies Following Metric Safety 4.4 4.6 4.7 4.2 4.1 5.0 6 Schedule 5.0 4.6 4.2 3.6 3.8 4.0 6 Communication and coordination 4.3 4.6 4.6 3.2 3.7 4.9 6 Budget 4.0 4.6 3.1 4.4 4 Quality control/quality assurance 4.7 3.4 3.9 4.8 4 Environmental compliance 4.4 4.3 4.0 3.9 4 Contract Documents/Submittals/RFIs 4.2 5.0 4.1 3 Public/Neighborhood relations 4.8 4.0 4.5 3 Team and project recognition while having fun 4.9 3.6 4.0 3 Timely decision making at the lowest level possible using issue resolution ladder 4.4 3.5 3.8 3 Integrated team response/Teamwork 4.8 5.0 2 Community appreciation 5.0 4.6 2 Anticipate project risks and address emerging issues collectively 4.8 3.5 2 Trust and respect each other and all the partners on the project including all stakeholders (e.g., public, local government, utilities, permitting agencies, etc.) 3.4 3.7 2 Focus on the Project first and empower people to follow this principle 3.4 3.9 2 R easonable profit incentive for contractors 4.2 1 Seize opportunities to improve project outcomes when they arise (Lessons Learned) 3.3 1 Total Number of Goal Alignment Metrics 7 9 9 13 9 7 Project Performance Score 100 100 88 85 85 55 153 595, 1050, and 759 workdays respectively) as compare to the other three case study projects (i.e., 378, 183, and 128 workdays for case studie s #1, #2, and #5 respectively) . Additionally, case stu dy projects #3 and #4 both had nine scorecards and partnering workshops up to the poin t of data collection. Figure 5 - 8 illustrates the trends over time for these two cases. Figure 4 - 8 : Goal Alignment ratings from Scorecards (Large - Mega) Examining Figure 4 - 8 it becomes more apparent how project teams may not always remain contiguously aligned with the project goals. This is evinced by case study #3 where the goal alignment scores dipped over the project duration then rebounded near the later period of the recorded project team ratings. Case study #3, based on observational data during a si te visit, frequently dealt with challenges regarding pricing change orders. The level o f trust had been challenged the team confronted risk management from disparate perceptions as the owner versus the contractor. When pricing extra work, the contractor encountered risk in 3.00 3.20 3.40 3.60 3.80 4.00 4.20 Scorecard 1 Scorecard 2 Scorecard 3 Scorecard 4 Scorecard 5 Scorecard 6 Scorecard 7 Scorecard 8 Scorecard 9 Case Study #3 Case Study #4 154 pricing added scope too low while the owner believed they often were forced to accept a price that was too high. The alternat iv e was to utilize force accounts which are intended to resolve concerns when negotiated pricing between the owner/design and contractor are untenable. Thus, it becomes obvious that an individual rate r would find it hard to disentangle these types of concerns from project performance perceptions. Although, the researcher observed the project team was willing to agree levels of trust will never reach 100 percent. They also maintained a sound level of in formal dialogue and offered kudos to many team members shortly after a spirited problem - solving exercise orchestrated by their neutral third - party partnering facilitator. Table 4 - 22 further communicates how goal alignment and project performance are relat ed. The table identifies additional case study characteristics being 1) a formal partnering charter codified with signatures was used; 2) timing of partnering charter implementation; 3) the parties involved in the development of partnering charter; 4) clar ity of goals communicated in partnering charter; and, 5) n umber of goal alignment metrics. Follow up procedures used by each case study proj ect are also shown in Table 4 - 22 . These include number of partnering workshops, number of attendees at partnering ki ckoff meeting, number of partnering scorecards surveys completed, and number of team building exercises. There is one clear trend among the projects with lower project performance ratings. The number of stakeholders involved in developing the partnering c harter was complete with the owner , design, contractor, and others for the lower performing projects. This indicates a challenge may persist in achieving and maintaining goal alignment when multiple stakeholders are involved. Moreover, there may be a tendi ng to include competing goals and objectives. 155 An example of competing interests arises in case study #3 whereby the partnering charter includes two goal alignment metrics as follows: This idea, while seemingly harmless, may have influenced the travails the project team experience d with trust. project in line w ith cost and schedule goals. Meanwhile, the contractors team is generally concerned with making a profit and must be delighted to find this codified in the partnering charter. This creates a perfect storm for skepticism among project teams casting doubt o n motives for negotiated prices of extra work. Table 4 - 22 : Goal Alignment and Congruence items from Partnering Documents Goal Alignment and Congruence Case Study #2 Case Study #5 Case Study #1 Case Study #3 Case Study #4 Case Study #6 Project Charter Formal Charter including Project Team Signatures N N N Y Y N Timing of Charter (e.g., # of months before or after construction (CNST) ) 4 - M onths after C NST 2 - M onths before CNST 5 - M onths after C NST 1 - M ont h before CNST 2 - M onths before C NST 2 - M onths after C NST Parties involved in Developing Charter (e.g., Owner - O, Designer - D, Contractor - C, Other Stakeholders - S) O, C O, C O, D, C, S O, D, C, S O, D, C, S O, D, C, S Clarity of Goals (i.e., More than th ree clear and detailed action items) N Y N Y Y N Number of Goals 9 9 7 13 9 7 Follow up Procedures Number of Workshops 3 3 3 9 9 3 Number of Attendees at Partnering Kickoff Meeting 17 9 6 30 52 11 Number of Partnering Scorecard Surveys 1 5 2 9 9 2 Team building training lessons N N N Y Y N Project Performance Score 100 100 88 85 85 55 156 Proposition 3 is supported in this study based on the analyses above. Next , a list of both practical application and theoretical implications are given. Further details are provided in Chapter 6 findings and discussions. Practical Applications Focus on a core set of goal alignment metrics that are detailed around a clear project objective . Anticipate goal alignment deviates over projects with longer dura tions and make provisions to continuously reinforce them . Avoid competing goal objectives within partnering charters . Theoretical Implications Relational governance permeates these case studies (Carson et al., 2006; Ouchi, 1980; Williamson, 1979) o Flexibility, solidarity, and information exchanges strategies followed to encourage trust . o Benchmarking met rics, partnering workshops, and conflict resolution strategies are not memorialized in the contracts . o Opportunisms concerns due to memorializing conflicting goals in project charter resorting back to TCE theory . Goal aligning feedback and goal congruence surfacing within case studies 157 o Neutral third - party facilitators used during partnering workshops to continuously bring project teams into alignment around project goals and objectives (Cheng & Li, 2002; Manley, Mcfallan , & Kajewski, 2009) . o Shared risk and reward structures did not necessarily lead to improved performance, although these were primarily risk structures [ e.g., liquidated dama ges for late project completion] (Manley et al., 2009; Pishdad - Bozorgi & Beliveau, 2016) . Chapter 4 presented the qualitative analyses emanating from this study. A summary of each case study project was given. Then, the researcher used pattern - matching to extend descriptive characteristics for the case studies including cost and schedule results. The chapter then shifts focus towards its determination to illustrate additional patter ns and trends. This cross - case synthesis approach allowed for proposition testing which ends Chapter 4 . The next chapter provides results from quantitative analyses of survey data collected as part of each case study. 158 CHAPTER 5 QUANTITATIVE ANALYSIS Chapter 5 provides quantitative analyses and results based on evidence collected from online surveys. These data were analyzed using Mplus Version 8 software package (Muthén & Muthén, 2017) . The chapter begins by summarizing case study data demographics, then gives a summary of the latent variables and factors underlying each variable in the survey . Next, the results from the model put forward in this study are analyzed using confirmatory factor analysis. The final section ends with hypothesis testing using multiple regression/correlation analysis (MRC) and summarizes the findings. The primary goal for this stage of the research was to quantitatively examin e the following at individual - level in interorganizational AEC project teams: 1. The relationship between individual /team performance , goal alignment , and TMS . To do so, the researcher used confirma tory factor analysis (CFA) to test the proposed factor structure and estimate factor scores for goal alignment , TMS , and individual performance constructs. Multiple regression/correlation analysis (MRC) was used for hypothesis testing of relationships amon g the constructs and respond to the research questions. The model developed in this study was used to test the following hypotheses: Hypothesis 1 : Individual performance in partnered - goal alignment perception . 159 Hypothesis 2 : MS moderates the relationship between individuals alignment perceptions and individual/team performance in partnered - projects. The researcher used Mplus Version 8 (Muthén & Muthén, 2017) in all statistical analyses. All of the statistical analyses performed were used to assess the measurement model, model fit, alon g with the underlying factor structure of the data. The analyses for all data collected in surveys was treated as categorical in CFA and MRC. The case study data demographics are given next followed by results from CFA. Six p ublic case study projects were used in this study to address the research questions and for hypothesis testing. The case study sample d emographics are shown in Table 5 - 1 including among other things project size, schedule, project type, project delivery me thod, and location. These case studies were utilized to investigate the relationships among goal alignment and performance outcomes moderated by transactive memory systems . An online survey was used to collect individual - level data from the case study pro ject participants. Overall there were 125 potential partnering participants across six cases. The online survey was accessed by 69 participants while 51 surveys were sufficiently completed to permit further data analyses. Table 5 - 2 displays the number of s urvey responses, partnering participants, and response rates for each case study. Each case study was represented in online survey data. Four of six case studies had a response rate above 30 percent while case study number #3 was 27 percent and case study number #6 was 20 percent. The final response rate for this study was 41 percent. 160 Table 5 - 1 : Case Study Sample Demographics Project Size ( * $M) Schedule (* * Workdays) ** * Project Type and Delivery Method Location No. of Partnerin g Participants No. of Partnering Workshops Case Study #1 5.59 345 Horizontal /DBB West 6 3 Case Study #2 4.96 255 Large Infrastructure /DBB West 17 3 Case Study #3 1 49.96 595 Horizontal /DBB Midwest 30 9 Case Study #4 272.99 1303 Large Infrastructure /DB M idwest 52 10 Case Study #5 3.10 257 Vertical /DBB West 9 3 Case Study #6 5.00 270 Vertical /DBB West 11 3 Table 5 - 2 : Summary of Study Sample and Responses No. of Survey Respondents No. of Partnering Participants Response Rat e Case Study #1 5 6 83% Case Study #2 9 17 53% Case Study #3 7 30 23 % Case Study #4 22 52 43% Case Study #5 5 9 56% Case Study #6 3 11 27% Overall Participant Totals 51 125 41 % The respondents in this survey also varied across roles in each case s tudy. The data shown i n Table 5 - 3 illustrates the spread among roles identified in survey responses. Based on * $M U.S. dollars in millions; ** Workdays excludes holidays and weekends ; ** * Project types included among others vertical (e.g., office buildings), horizontal (e.g., roadways), and large infras tructure (e.g., tunnels, bridges , or major highway infrastructure) 161 (n=24). Meanwhile, the Contractor group are second with (n=11) response. No external stakeholders responded to the survey. Four individuals selected the role of facilitator, however, they noted in text response the role of Construction Manager or Independent Construction Quality Manager. The role listed a s others included Independent Construction Quality Managers (n=2), Safety Managers (n=2), and Construction Inspectors (n=1). Table 5 - 3 : Respondent Demographics based on Project Role Project Role Case Study #1 Case Study #2 Cas e Study #3 Case Study #4 Case Study #5 Case Study #6 Totals Owner or Owners' Rep 2 2 4 13 2 1 24 Facilitator 0 3 0 1 0 0 4 Designer/Engineer 1 0 1 2 1 0 5 Contractor 2 2 1 3 1 2 11 Subcontractor 0 1 0 1 0 0 2 Other * 0 1 1 2 1 0 5 Total # of Responde nts 5 9 7 22 5 3 51 *Others rated their role as Independent Quality Manager, Safety Manager, or Construction Inspector CFA is used to test a priori hypotheses about relations between observed variables and latent variables or factors (Jackson, Gillaspy, & Purc - Stephenson, 2009) . The object ive is to estimate a population covariance matrix that is compared with the observed covariance matrix. The goal is to minimize the difference between the estimated and observed matrices. CFA is used in this study to validate how well the hypothesized mo del fits the data. Goodness - of - fit summary statistics for the measurement models are confirmed using three fit indices, Chi - Square ( ) test of model fit, root mean square error of approximation (RMSEA), 162 and comparative fit index (CFI). Indications of good model fit using Chi - Square ( ) is a low relative to degrees of freedom with a high p - value ( p > .05), RMSEA values between 0.08 to 0.10 (adequate fit) and less than 0.07 (good fit), and CFI 0.95 (Hooper, Coughlan, & Mullen, 2008) . The latent variables in this study were treated as categorical to improve the data structure. The full measurement model was then estimated using summed scale scores across each of the latent variables . The fit statistics for the final measurement model demonstrates the model fits the data ( = 38.46, df = 33.0 p = 0.24, RMSEA = 0.06 , CFI = 0.98). Detailed results of statistical analyses for the final measurement model are shown in Appendix C. Next, each latent variable is illustrated along with its underlying data structure. Goal Alignment Latent Variable There were five indicators underlying the latent variable for goal alignment . The latent construct for goa l alignment and its indicators are shown in Figure 5 - 1 below. Detailed results of statistical analyses for the latent variable goal alignment are shown in Appendix D. Figure 5 - 1 : Goal Alignment Latent Variable and Factor Indicators Goal Alignment GA1 GA3 GA2 GA4 GA5 E1 E3 E2 E4 E5 163 Each measurement i ndicator represented in Table 5 - 1 is described and shown with its factor loading (see Appendix D for detailed statistical analyses) . The resulting standardized factor loadings and standard errors (S.E.) from CFA were [GA1 = 0. 79 (0.08), GA2 = 0.99 (0.06), GA3 = 0.76 (0.07), GA4 = 0.92 (0.10), and GA5 = 0.78 (0.11). All factors were retained in this analysis as they were significant at 0.05 level of significance. The fit statistics for goal alignment validates the model adequate ly fits the data ( = 7.61, df = 5.0, p = 0.18, RMSEA = 0.10 , CFI = 0.99). Table 5 - 4 : Factor Structure and Factor Loadings for Goal Alignment Factor Structure CFA Factor Loadings (standardized estimates, p < .05 ) Go al Alignm ent (GA) Mutual goals and objectives in the partnering charter were communicated effectively (GA1) . 0.78 Clear and compatible partnering goals were established by the project team (GA2) . 0.99 I generally agreed with project - related goals esta blished by the project team (GA3) . 0.76 My attitude towards project - related goals established by the project team were similar (GA4) . 0.92 My goals for the project were in close alignment with the project team (GA5) . 0.78 Transactive Memory System (T MS) Latent Variable TMS was measured using three sub - factors , coordination, communication, and specialization to form the higher order latent variable. The factor structure for TMS is shown in Figure 5 - 2. 164 The sub - factor s tructure illustrated in Figure 5 - 2 shows eight underlying indicators while results are described in Table 5 - 2 (See Appendix D for detailed statistical analyses) . The results of CFA demonstrated several weak factors of TMS being (CO1) and (CR4). These indicators had either weak loadings or negative residual variances, thus were removed from CFA model to improve model fit. The rest of the factors were retained at 0.05 level of significance and used in CFA to test the model fit. The standardized factor loadings and standard errors (S.E.) a re given as [CO2 = 0.94 (0.08), CO3 = 0.80 (0.10), CR1 = 0.66 (0.12), CR2 = 0.99 (0.05), CR3 = 0.92 (0.04), SP1 = 0.42 (0.15), SP2 = 0.40 (0.16), and SP3 = 0.92 (0.16)]. The fit TMS E1 E3 E2 CO 1 * CO 3 CO 2 Coordination CR 1 CR 2 CR 3 CR 4 * E4 E5 E6 E7 C redibility SP 1 SP 2 SP 3 E7 E8 E9 Specialization Figure 5 - 2 : Factor Structure for the TMS Latent Variable; * I ndicators with weak loadings or negative residual variances 165 statistics for TMS validates the reduced model adequately fits the data ( = 23.62, df = 17.0, p = 0.13, RMSEA = 0.09 , CFI = 0.99). Table 5 - 5 : Factor Structure and Factor Loadings for Transactive Memory Systems Factor Structure Higher - order Factor Loadings (standardized estimates, p< .01 ) S ub - Factor Loadings (standardized estimates, p < .01 ) Coordination (CO) 0.86 The project team had very few misunderstandings about what to do during construction (CO2) . 0.94 I believe we accomplished our task for the project smoothly and efficiently (CO3) . 0.80 Credibility (CR) 0.92 I was comfortable accepting procedural suggestions from other team members (CR1). 0.66 was credible (CR2). 0.99 I was confident relying on the informat ion that other team members brought to the discussion (CR3). 0.92 Specialization (SP) 0.83 I understand what skills my team members have and domains they are knowledgeable in (SP1) . 0.42 The specialized knowledge of several different team members was needed to complete the project (SP2) . 0.40 and knowledge (SP3) . 0.92 Individual/Team Performance There are three dimensions that were assessed in this study regarding the latent variable i ndividual/ team performance . The three sub - factors were project related , communication - related , and team related . The number underlying indicators for each of the sub - factors were (8) project performance , (4) communication performance , and (7) team perform ance. Figure 5 - 3 166 illustrates the latent variable for individual/team performance and its resulting factor structure is shown in Table 5 - 3. Individual/Team Performance Project Performance PP 1 * PP 3 PP 2 PP 4 PP 5 PP 6 * PP 7 PP 8 E1 E3 E2 E4 E5 E6 E7 E8 Credibility Performance CP 1 * CP 3 CP 2 CP 4 E9 E11 E10 E12 Team Performance TP 1 TP 3 TP 2 TP 4 TP 5 TP 6 TP 7 * E13 E15 E14 E16 E17 E18 E19 Figure 5 - 3 : Factor Structure for t he Individual/Team Performance Latent Variable ; * I ndicators with weak loadings or negative residual variances 1 67 Table 5 - 6 : Factor Structure and Factor Loadings for Individual/Team Per formance Factor Structure Higher - order CFA Factor Loadings (standardized estimates, p < .01) CFA Sub - Factor Loadings (standardized estimates, p < .01) Project Performance (PP) 0.70 We adhered to schedule targets for the project (PP2) . 0.62 We sati sfied the scope of work for the project (PP3). 0.90 We accomplished our task for the project smoothly and efficiently (PP4) . 0.83 We achieved project goals established by the project team (PP5) . 0.76 We delivered a high - quality project for the owner ( PP7) . 0.56 The project was delivered safely without major safety incidents (PP8) . 0.74 Communication Performance (CP) 0.74 There was frequent communication within the team (CP2) . 0.91 Team members communicated often in spontaneous meetings, ph one conversations, etc . (CP3) . 0.84 The team members largely communicated directly and personally with each other (CP4) . 0.79 Team Performance (TP) 0.94 All project team members were treated equal ly in the decision - making process (TP1) . 0.76 Al l project team members worked with the same focus on project objectives (TP2) . 0.81 We worked together to share information across organizational boundaries (TP3) . 0.73 We worked towards mutually beneficial outcomes for all participants (TP4) . 0.88 A ll project information was readily available to everyone involved in the project (TP5). 0.73 We always sought collective identification and resolution of problems (TP6) . 0.65 The results of CFA demonstrated several weak indicators of individual/tea m performance being (PP1), (PP6), (CP1), and (TP7), as such were removed to improve model fit. The rest of the factors were retained at 0.05 level of significance and used in CFA to test the model fit. The standardized factor loadings and standard errors ( S.E.) are given as [PP2 = 0.62 168 (0.10), PP3 = 0.90 (0.10), PP4 = 0.83 (0.09), PP5 = 0.76 (0.09), PP7 = 0.56 (0.10), PP8 = 0.74 (0.15), CP2 = 0.91 (0.10), CP3 = 0.84 (0.08), and CP4 = 0.79 (0.10), TP1 = 0.76 (0.09), TP2 = 0.81 (0.07), TP3 = 0.73 (0.06), TP4 = 0.88 (0.07), TP5 = 0.73 (0.07), and TP6 = 0.65 (0.10)]. The fit statistics for individual/team performance validates the reduced model fits the data ( = 101.02, df = 87.0, p = 0.10, RMSEA = 0.06 , CFI = 0.97). Detailed statistical analyses for laten t variable shown in Figure 5 - 3 can be found in Appendix D. Next, t he correlations among all latent variables both higher - order latent variables and sub - factors are shown in Table 5 - 7 . Factor scores generated from the measurement model were used to determi ne correlations among latent variables. Based on this table, all sub - factors are highly correlated with their higher - order factors. The composite reliability for each factor is also included in the table (see Appendix E and Appendix F for detailed statisti cal analyses for composite reliabilities) . Table 5 - 7 : Correlations among Higher - order Latent Variables and S ub - factors Composite Reliability F1 F2 F3 F4 F5 F6 F7 F8 F9 1. Project Performance 0.83 2. Communication Pe rformance 0.82 0.61 3 . Team Performance 0.83 0.74 0.79 4 . Individual/Team Performance 0.84 0.80 0.85 0.99 5 . Coordination 0.77 0.63 0.74 0.79 0.81 6 . Credibility 0. 83 0.55 0.77 0.81 0.82 0.88 7 . Specialization 0. 50 0.53 0 .78 0.77 0.79 0.82 0.87 8 . Transactive Memory System 0.90 0.60 0.80 0.83 0.84 0.94 0.98 0.93 9 . Goal Alignment 0.86 0.55 0.70 0.66 0.69 0.66 0.80 0.71 0.77 Composite reliability was computed using nonlinear SEM reliability coefficient suggested by Green and Yang ( 2009) . In this approach, polychoric correlations are estimated then 169 followed by weighted least square estimation methods in Mplus. Once the model is fit, sample estimate s for factor correlations, factor loadings, thresholds, and polychoric correlation are used as parameters to compute the nonlinear SEM reliability coefficient in SAS (Statistical Analysis Software) program (Green & Yang, 2009) . Based o n the tabl e , several measures demonstrated good dimensionality and reliability (e.g., Project Performance, Communication Performance, Team Performance, Coordination, Credibility, and Goal Alignment). Meanwhile, the higher - order factors were estimated in SEM using ea ch of the full measurement model s (i.e., Individual/Team Performance and TMS). This next step in SEM analysis is to inspect the unconditional and multilevel models. The unconditional model in Mplus did not show significan t variation among case studies or levels of analyses in this study. In addition, SEM in clusive of all higher - order latent variables and sub - factors in the measurement model failed to converge due to sample size. As such, the researcher determined the data in this study was not hierarchical and continued the analyses using multiple regression/correlation analysis (MRC). MRC is a flexible statistical analysis approach when dealing with quantitative variables (Cohen et al., 2013) . It is used to test th e relationships between a dependent variable and multiple independent variables. As with any statistical analyses, several key assumptions must be considered prior to data analysis (Hair, Black, Babin, Anderson, & Tatham, 1998) . 1. A linear relationship exist s between the outcome variable a nd the independent variables. 170 2. The residuals are normally distributed or multivariate normality. 3. Independent variables are not highly correlated with each other or multicollinearity. 4. The v ariance of error terms is similar across the values of independent va riable or homoscedasticity. The assumptions for linearity and normality are validated by evaluating the normality plot and histograms Figure 5 - 4 and Figure 5 - 5. In the normality plot shown in Figure 5 - 4, the data are linearly distributed. A good indication of linearity is the observation that data points are spread ac ross the diagonal line. Figure 5 - 5 shows the histogram and distributions for the sample data. The data in this figure illustrate fairly good in terms of nor mal distribution. Last, Figure 5 - 6 pl ots estimates for individual performance ( x - axis) against its residual terms ( y - axis) to inspect for homoscedasticity. The observations show fairly good distribution above and below the zero point on the y - axis. The data in this study satisfied the assumpt ions and, thus, permitted the researcher the data using MRC. 171 Fi gure 5 - 4 : P - P Plot for individual performance 172 Figure 5 - 6 : Residual plot of individual performance Figure 5 - 5 : Histogram and distribut ion curve for individual performance 173 The descriptive statistics for the surve y data are given next in Table 5 - 4. These data were aggregated from survey responses across each second - order latent variable investigated in the model. These were used to represent a combined average value for higher - order latent variables in the model. Table 5 - 8 : Descriptive statistics for Goal Alignment, TMS, and Individual/Team Performance Goal Alignment TMS Individual/Team Performance Mean * 4.70 4.36 4.21 Standard Error 0.06 0.06 0.08 Median * 4.80 4.50 4.29 Mode * 5.00 4.50 4.47 Standard Deviation 0.38 0.44 0.59 Sample Variance 0.15 0.20 0.35 Kurtosis 0.13 - 0.06 - 0.60 Skewness - 1.10 - 0.80 - 0.60 Range 1.40 1.70 2.16 Minimum 3.60 3.30 2.84 Maximum 5.00 5.00 5.00 Sum 225.60 209.10 202.05 Count ** 48.00 48.00 48.00 Confidence Level (95.0%) 0.11 0.13 0.17 *Based on Likert Sca le ranging from 5 - Strong ly agree to 1 - Strongly disagree (**n=48); Average of all sub - factor scores Results show relatively high mean ratings and low standard deviations across the sample respondents from the online survey (n=48) for the higher - order late nt variables [i.e., goal alignment 4.70 (0.38), TMS 4.36 (0.44), and Individual/team performance 4.21 (0.59). In other words, little variation is present among respondent ratings based on the descriptive statistics. 174 The unconditional model was tested to a ssess whether a hierarchical structure exist for these data (i.e., only included level two or between group variable in the analysis ). The equations below illustrate the unconditional model: Where, The results from this initial inspection f ou nd no significant variation at the between the level of analysis (i.e., 0.36 (0.2 5 ), p= 0.15). T ests at this juncture were conducted using computed factor scores from CFA attached to higher - order latent variables original statistical analyses assumptions. Particularly, test for skewness and non - normality. Factor scores in this study were imputed using CFA thus, are more stringently derived and hold a gainst multiple fit indices. The higher - order latent variables were used to test the full model given in this study. The full model is shown below only examining level one (i.e., individual performance): The researcher tested the full model using multivariate regression. The variab les in this study explained 70 percent of the variance (i.e., Adjusted R Square = 0.70, p < 0.001) in individual performance perceptions for case study project team members. Of the four 175 parameters in the model, one achieved significance at 0.001 (i.e., TMS r = 0.82 , S.E. = 0.14, p = .000). The results are show n in Table 5 - 5. One of three hypotheses is supported as indicated below: Hypothesis 1 : Individual performance in partnered - goal alignment perception . ( Not supported ) Hypothesis 2 : MS moderates the relationship between individuals alignment perceptions and individual/team performance in partnered - projects. ( Not supported ) Table 5 - 9 : Descriptive statistics from m ultivariate regression analysis Regression Statistics Multiple R *0.85 R Square *0.72 Adjusted R Square *0.70 Standard Error 0.45 Observations 48 df SS MS F Regression 3 22.90 7.63 37.034 Residual 44 9.07 0.21 Total 47 31.97 Coefficients Standard Error t Stat P - value Intercept 0.05 0.08 0.64 0.528 Goal Alignment 0.11 0.15 0.69 0.493 Goal Alignment x TMS 0.01 0.12 0.07 0.947 TMS* 0.82 0.14 5.92 *0.000 *Significant p < 0.001 The results from CFA and MR C analyses demonstrate both utility and promise for both practitioners and researchers. The researcher identif ies key attributes underlying goal 176 alignment , TMS , and individual performance . Although only one of the latent constructs (i.e., TMS ) was signific ant, this mild finding paves the way to continue exploring behavioral attributes during collaborative project delivery. A few practical and theoretical implications are listed next. Practical Applications Practitioners should be cognizant during project delivery to the benefits of developing a clear system for coordination and communication in project teams. It is within this knowledge processing system where efficiency and trust are established. Theoretical Implications Support for measures of goal alig nment (Chan et al., 2004; Jap, 1999) , TMS (Kyle Lewis, 2003; Zhang et al., 2015) , and individual performance (Grajek et al., 2000; Gransberg et al ., 1999; Hoegl & Gemuenden, 2001; J. S. - C. Hsu et al., 2012; Yeung et al., 2007) within AEC project teams. perceptions are positively related to transactive memory systems Latent constructs in the model performed well in explaini ng the variance in individual performance perceptions. Future researchers may find these predictors useful to continue exploring relational governance structures that underlie collaborative project delivery practices. 177 Chapter 5 presented the quant itative analyses emanating from this study. Case study data demographics were given. Then, the researcher used confirmatory factor analysis to validate how well the hypothesized model fits the data. The chapter ends with hypothesis testing using multivaria te regression analysis and reports its results. The next chapter summarizes the study findings and discusses practical and theoretical implications. 178 CHAPTER 6 DISCUSSIONS This chapter summarizes key findings, discusses the practical appl ication and theoretical contributions of this research. The chapter begins with qualitative findings from pattern - matching and cross - case synthesis. Next, quantitative findings captured from multivariate regression analysis are presented. This is followed by confirmatory factor analysis findings regarding the utility of study metrics used in quantitative data collection and analyses. The cal implications and contributions are offered. In this section, the first objective was addressed based on qualitative findings. There were three propositions posited and validated using partnered project documents and structured in terview data. The findings for each proposi tion are given below. Proposition 1: Project risk factors and the level of collaborative project delivery practices in partnered - projects are positively related . (Supported) The first proposition was supported in this study as evinced by pattern - matching and cross - case synthesis results. Using case study tactics, the researcher observed a clearly discernable pattern between project risk factors and collaborative project delivery practices. It was shown that four o f six case studies successful ly aligned their practices with project risk factors. When project risk w as perceived as low, it became less important for collaborative project delivery practices to be aligned. However, two case studies (i.e. case studies #3 and #6) did not support this proposition. 179 As a result, these two cases offered a new proposition for future research. Proposition 1a - If collaborative project delivery practice s are not positively aligned with the level of project risk in partnered - pro jects, then project performance will be negatively affected. The new proposition surfaced when holistically examining the structured interview data. In this , it becomes apparent that project performance is negatively affected when collaborative project de livery practices are not aligned with the level of project risk. Proposition 2: perceptions in partnered - projects are positively related . (Limited Support) Findings from proposition 2 ostensibly did not support this proposition. Again, the researcher took a broader look not only on structured interview data but at the case study partnering documents. In doing so, an obvious pattern emerges when collaborative project delivery practice s, goal alignment scores, and goal alignment actions are considered together. Goal alignment actions work in opposite directions. That is, higher collaborative project delivery practices constrained the need for project teams to use many goal alignment act ions or metrics. The following proposition is offered as a result of this finding: Proposition 2a Higher collaborative project delivery practices require fewer goal alignment actions or metrics to hold the project team accountable in partnered - projects. Proposition 3: - projects are positively related . (Supported) 180 The final proposition asserted in this study was also supported. This finding was exemplified when the researcher scr utinized case study partnering documents. Generally, when goal alignment is high project performance is also rated highly. One case study project was inconsistent in the results, and upon further inspection, the researcher discovered the top two projects f ollowed similar performance metrics while the outlier (i.e., case study #6) did not . These two projects , specifically, included goal aligning performance metrics such as ensuring an integrated team response or teamwork was encouraged during problem - solving processes. Additionally, the top two projects sought continuous improvement by including a focal point around increased efficiencies in document management. Given the results, a new proposition developed: Proposition 3a: - Collaborative delivery practices should accommodate continuous improvement in facilitating information exchange among team members in partnered - projects. Considering the propositions above , strong support is demonstrated for the proposed framework posited in this study. In addition, thre e new propositions surfaced to which future research may consider. The quantitative aspects underpinning the first study objective are summarized next. This study sought to understand the relationship between individual/team performa nce and goal alignment. In addition, the researcher hypothesized a moderated effect existed among this relationship. The hypotheses were tested using partnered - project teams as a subset of collaboratively delivered projects. The results are recapitulated i n this section along with the generalizability of findings. 181 Hypothesis 1 : Individual performance in partnered - goal alignment perception . (Not supported) Results from hypothesis 1 did not support the theorized relationship posited. However, results from qualitative analysis help explain and triangulate th ese findings . The results may differ given a larger case study sampling where between level effects are more pronounced. eptions were not related to goal alignment, project performance perceptions showed a clear relationship with TMS . This is explained in more detail at the end of this section . Hypothesis 2 : MS moderates the relationship between individuals goal alignment perceptions and individual/team performance in partnered - projects. ( Not supported ) Results for the theorized relationship in hypothesis 2 was not supported. In essence, little support was found in the data that TMS moderates the relationship goal alignment and individual/team performance perceptions. Table 6 - 1 display s descriptive statistics for the variables investigated in this study. The two hypotheses are useful for AEC projects following collaborative project delive ry approaches, however, generalizing finding is cautioned. Partly, because the sample size is small across case studies and survey respondents (i.e., case studies = 6; N = 48). Despite this, the variables in this study explained 70 percent of the variance (i.e., Adjusted R Square = 0.70, p < 0.001) in individual performance perceptions for case study project team members. Of the four parameters in the model, one achieved significance at 0.001 (i.e., TMS r = 0.82 , S.E. = 0.14, p = .000). 182 Table 6 - 1 : Descriptive statistics for the relationship among study variables Regression Statistics Multiple R *0.85 R Square *0.72 Adjusted R Square *0.70 Standard Error 0.45 Observations 48 Coefficients Standard Er ror t Stat P - value Intercept 0.05 0.08 0.64 0.528 Goal Alignment 0.11 0.15 0.69 0.493 Goal Alignment x TMS 0.01 0.12 0.07 0.947 TMS* 0.82 0.14 5.92 *0.000 *Significant p < 0.001 Taken together, the qualitative findi ngs in this study offer a glimpse into the implications of goal alignment actions found in partnered - projects while quantitative findings begin to pinpoint a vital component for future investigation. The notion of goal alignment is seemingly innocuous, and many would quickly agree that effectively achieving multiple tasks (e.g. constructing a building) requires uniformity amongst the project team. However, research is still burgeoning as to why collaborative project delivery practices (Pishdad - Bozorgi & Beliveau, 2016) are a necessity conduit to bring about trust and goal congruence (Dietrich et al., 2010) . Correctly identifying the project risks serves multiple purposes. First, it helps project teams b etter prepare for transaction hazards such as ambiguity, volatility, uncertainty, and opportunism found during relational exchanges (Poppo & Zenger, 2002) . The use of in formal cues helps to marginalize these hazards during project delivery. For example, partnering workshops bring project teams together to address challenges in an open forum while 183 continuously reminding them of their shared vision and goals for the project . It is within this environment where team members become more empowered to offer solutions when problems are encountered. The increases the level of trust and credibil i ty within the team. Second, it gives the owner and project team the ability to plan f or some measure of uncertainty in advance. As an example, five of the case studies in this study included contingency dollars in the project while case study #6 never reported having any . Ironically, this case study project was noticeabl y impacted by unfor eseen scope straining project resources (i.e., cost and time). Thus, the project pe r formance suffered in this case study. Even with contingencies, however, some case studies reported lower than anticipated pe r formance ratings. This may have presented the perspective. Contingency may give rise to guarded optimism but can quickly change to a more conciliatory tone when change request become s more frequent in projects. Change orders are then sc r utinized from a lens of skeptici sm anticipating inflated pricing. This promotes mistrust o f proj ect performance even in light of hard metrics (e.g., actual cost and schedule data). Despite this, sharing common goals while managing and planning for proj ect risks influences project performance outcomes. Next, a brief discussion on the utility of metrics utilized in data analysis is given. The second objective of this study was to provide empirical support for the quantitative me trics. Alt hough goal alignment and TMS have received vast attention i n organizational (Argote, 2015; Argote, Ingram, Levine, & Moreland, 2000; Stephen & Coote, 2007; Sundaramurthy & Lewis, 2003) and psychology (Hollingshead, 1998a; Park, Spitzmuller, & 184 DeShon, 2013; Wegner, 1987; Wegner et al., 1991) literature, these constructs are still rather novel in AEC literature (Cacamis & El Asmar, 2014; Comu et al., 2013; Suprapto et al., 2015; Zhang et al., 2015) . Given this, this study use d CFA to offer two metrics for future research. Mean while, a measure unique measure of individual performance was posited which moves beyond traditional cost and schedule performance metrics. Findings from CFA confirm the constructs for goal alignment , TMS, individual performance are useful as a measureme nt model to assess team dynamics on partnered projects . The underlying indicators supporting each construct used in the analyses are given below in Table s 6 - 1 and 6 - 2 . Standardized factor loadings, standard errors, and model fit statistics are also display ed in the tables. 185 Table 6 - 2 : Findings from CFA for goal alignment and TMS Latent Constructs and Factors Standardized Factor Scores Standard Errors Goal Alignment ( = 7.61, df = 5.0, p = 0.18, RMSEA = 0.10 , CFI = 0.99) Mutual goals and objectives in the partnering charte r were communicated effectively 0.79 0.08 Clear and compatible partnering goals were established by the project team 0.99 0.06 I generally agreed with project - related goals established by the project team 0.76 0.07 My attitude towards project - related goals established by the project team were similar 0.92 0.10 My goals for the project were in close alignment with the project team 0.78 0.11 TMS ( = 23.62, df = 17.0, p = 0.13, RMSEA = 0.09 , CFI = 0.99). Coordination The project team had very few misunderstandings about what to do during construction 0.94 0.08 I believe we accomplished our task for the project smoothly and efficiently 0.80 0.10 Credibility I was comfortable accepting procedural suggestions from other team members 0.66 0.12 0.99 0.05 I was confident relying on the information that other tea m members brought to the discussion 0.92 0.04 Specialization I understand what skills my team members have and domains they are knowledgeable in 0.42 0.15 The specialized knowledge of several different team members was needed to complete the proje ct 0.40 0.16 0.92 0.16 186 Table 6 - 3 : Findings from CFA for individual performance Latent Constructs and Factors Standardized Factor Scores Standard E rrors Individual Performance ( = 101.02, df = 87.0, p = 0.10, RMSEA = 0.06 , CFI = 0.97 ) Project Performance We adhered to schedule targets for the project 0.62 0.10 We satisfied the scope of work for the project 0.90 0.10 We accomplished our task for the project smoothl y and efficient 0.83 0.09 We achieved project goals established by the project team 0.76 0.09 There was little rework required in this project 0.56 0.10 The project was delivered safely without major safety incidents 0. 74 0.15 Communication Perform ance There was frequent communication within the team 0.91 0.10 Team members communicated often in spontaneous meetings, phone conversations, etc. 0.84 0.08 The team members largely communicated directly and personally with each other 0.79 0.10 T eam Performance All project team members were treated equally in the decision - making process 0.76 0.09 All project team members worked with the same focus on project objectives 0.81 0.07 We worked together to share information across organizational bo undaries 0.73 0.06 We worked towards mutually beneficial outcomes for all participants 0.88 0.07 All project information was readily available to everyone involved in the project 0.73 0.07 We always sought collective identification and resolution of pro blems 0.65 0.10 This study addresse d the third objective by summarizing key attributes case study partnered projects hold in common. Additionally, this section presents the underlying common alities of alignment, and project performance. 187 The data in this study used partnered projects as a subset of collaborative project delivery approaches typified in AEC literature (Rahman & Kumaraswamy, 2005; Suprapto et al., 2015 ; Xue et al., 2010) . Several trends were identified that are consistent with other collaborative approaches. Partnered projects were consistently able to gain efficiencies in the areas of cost and schedule growth. This finding joins recent research which shows team integration and these common performance measures are related ( Franz et al., 2016) . There are o ther practical considerations from this study which are discussed next. When project risks are relaxed collaborative project delivery practices are easily aligned. This fi nding showed up in proposition one . Based on this, it becomes clear that project risks suc h as teams with limited partnering experience or working on projects with high visibility required increased collaborative practices. These may offer early warnings if identified during the initial phase to engage in construction services. Proposition 1 al so provided insight to incentives/disincentives. Owners should a void disincentives such as liquidated damages unless subsequent incentives or risk sharing arrangements are included in contracts (Meng, 2012) . Performance commo n goals and objectives. Otherwise, the informal attributes of flexibility, solidarity, and trust are eroded over time. Proposition two posits that p roject teams should be cautious when many goal aligning action s surf ace during partnering workshops. This may serve as a l eading indicat or that project team s may be strained to align their individual goals with those outlined in the project charter . 188 This becomes problematic when performance is increased because complacency arises and ttribute the success to the goal alignment actions taken (Sundaramurthy & Lewis, 2003) . This is especially true when project risk is low and a limited num ber of collaborative practices are followed during project delivery. Pro position three brought forth the notion that project teams should f ocus on a core set of goal alignment metrics that are detailed around a clear project objective . Particularly, they should avoid competing goal objectives within partnering charters as one can serve two master s Th is internal conflict is a behavioral decision between the individual and team goals when some form of feedback (e.g., goal alignment objectives and actions identified in partnering charters) is present (DeShon et al., 2004) . Meanwhile, project teams should expect goal alignment to deviate over project life - cycle and should make provisions to continuously reinforce them . Again, helps prevent the team from falling into the self - perpetuating trap or cycles of collaboration that are destined for failure and become painfully obvious when project performance indicators drop (Sundaramurthy & Lewis, 2003) . At this point, rigidity builds within the team and only exacerbates the problem to no end. The next section covers the contributions of this study to the knowledge ba se and addresses its limitations. Construction project owners have relied on t raditional construction project delivery methods such as design - bid - build (DBB), construction management (CM), and design - build (DB) for years. 189 Though effective, many of these formal contracting practices work against collaboration and communication across organizations during the construction process. T hese approaches contractors behaviors towards transaction costs while attemptin g to deal with uncertainties involved in project delivery (Li et al., 2013; Ouchi, 1980) . Conversely, relational project delivery methodologies (e.g., Project Partnering, Strategic or Project Alliancing, and Integrated Project Delivery [IPD]) are bent on increasing leve ls of collaboration across organizations and to help mitigate risks (Lahdenperä, 2012) . Relational governance theory which pro mote s norms of flexib ility, solidarity and information exchange spurs these delivery approaches a long (Poppo & Zenger, 2002) . In this arrangement, economic safeguards found in traditional contracts are relaxed as individuals focus on trust to minimize opportunistic behaviors (Zaheer & Venkatraman, 1995) . This study challenged this dynamic wit h the research question centered on how goal alignment affects performance in AEC project teams when collaborative project delivery practices are followed The motivation behind relational or collaborative project delivery approaches is to merge multi - d isciplinary orga nizations into one cohesive unit. R emoving cultural barriers requires strategies such as project team member consistency, colocation, and early involvement of all team members (e.g., prime contractors and specialty subcontractors ) in decisi on - making ( Baiden et al., 2003) . T eam integration strategies help project team members share information open ly and honestly while tapping into a broad range of knowledge and expertise early on when decisions are less costly and more effective (Ospina - Alvarado et al., 2016) . These distinct advantages comprised of experiences, mental models [i.e., an organized 190 structure of shared knowledge among the team (Moh ammed , & Dumville, 2001) ], and motivation bring s about goal alignment within project teams (Dietrich et al., 2010) . De parting from literature, early involvement of key participants ( e.g., designer/contractor/specialty subcontracto rs) during schematic design did not appear significant in this study ( Baiden et al., 2003 ; Cheng & Li, 2002; Pishdad - Bozorgi & Beliveau, 2016) . Moreover, c olocation, a common collaborative project delivery practice, was identified in only one case study despite tremendous benefits asserted in the literature (Baiden et al., 2003; Pishdad - Bozorgi & Beliveau, 2016) . This i s likely due to the constricted sampling of collaborative case study projects . The case studies were primarily public projects which have explicit design and preconstruction processes that are not conducive to early contractor involvement. Perhaps, a large r sample that includes both public and private projects can differentiate between the competing ideas. Relational governance permeates these case studies (Carson et al., 2006; Ouchi, 1980; Williamson, 1979) as ample strategies were formally and informally codified in contract documents bent on flexibility, solidarity, and information exchanges to encourag e trust. Not only did these project teams establish a relational governance structure, but feedback mechanisms (e.g., workshops, scorecards, and partnering meetings) were in place to encourage accountability. The benefits from feedback, benchmarking metric s , and conflict management strategies are to help project teams attain goal alignment. Goal congruence is increased especially during partnering workshops or partnering sessions where a n eutral third - party facilitator is used to provide direct feedback on shared goals and objectives. 191 During partnering sessions , s ocial loafing concerns may arise where project teams or members may take a backseat because the number of goal alignment actions becomes overwhelming (Lam, 2015) . This problem is exacerbated as the number of project team members participating in workshops is increased . Thus, practitioners may find many participants are not fully committing to the partnering process resorting back to TCE theory focusing on individual goals . Recent research points out existing paradoxes in the field of AEC project management and collabo rative approaches as being generalized ideas that, when followed, do not always improve performance (Jacobsson & Roth, 2014) . Conversely, this study uses case study evidence to guide and demonstrate a relationship between collaborative pr actices, goal alignment , and perfor mance outcomes (Bresnen, 2007; Suprapto et al., 2015) . In addition, it fills the gap in literature whereby behavioral underpinnings (i.e., goal alignment and TMS) are illuminated as key links between practices and performance. To summarize, this section described the challenges and continued effort s the AEC industry balances aiming to improve project performance. Collaborative project delivery methods and approaches such as IPD and partnering are closing in on key informal attributes and mechanism s. The chapter summarizes findings from both quantitative and qualitative analysis followed in this study. Important characteristics emanating from the case study evidence is offered for researchers and practitioners. The chapter concludes by providing the oretical implications resulting from this research. 192 CHAPTER 7 CONCLUSIONS Chapter 7 b riefly summarize s research goals and objectives, research methods, findings, contributions to the body of knowledge . Then, the chapter finishes with limitations and sugg estions for future research. The aim of this research was to systematically identify the underlying attributes of collaborative project delivery approaches which result in better performance. Focusing on partnered - projects as a type and subset of collaborative AEC project delivery approaches, the specific goals and objectives of this study were to: 1. Develop a framework demonstrating the relationship between project risk factors, collaborative project delivery practi ces, goal alignment, TMS , and performance outcomes by; a. Qualitatively examine the following at partnered - project level: i. The links among project risk factors , collaborative project delivery practices , and project performance . b. Quantitatively examine the follo wing at individual and team level in interorganizational AEC project teams: i. The relationship between individual /team performance , goal alignment , and TMS . 193 2. Test an evaluation metric for goal alignment and utility of TMS metric to investigate AEC collaborati o n during project delivery; 3. Help facilitate collaborative contracting in construction projects by identifying key characteristics individuals in common. 4. Provide theoretical contributions to AEC literature understanding collaborative pro ject delivery methodologies as a form of relational governance. This study used empirical evidence via survey data, structured interviews, and case study project documents to support its goals and objectives. The researcher was able to satisfy these goals and objectives as shown presented in the summary of findings below. A mixed - methods researc h approach was followed in this study building upon an emerging perspective to understand interorganizational project teams in AEC liter ature and analyze subsequent data (Korkmaz, 2007) . The researcher initially asserted a multilevel or nested data structure with individuals embedded within teams (e.g., owner teams, design teams, and contractor teams) and teams are nested within case study projects. According t o Yin (2003), multiple units of analysis offers greater flexibility to inspect data for consistent patterns across units and cases. A mixed - method approach enabled two types of data collection and data analysis methods (i.e., qua l itative, qua nti tative). Th ese data were used to integrate findings and draw inferences from a single study or theoretical perspective (Miller et al., 2011) . Another advantage of mixed - methods is the ability to triangulate the findings of multiple forms of data, quantitative and qualitative (Campbell & Fiske, 1959) . 194 Partnered case study project documents and structured interview data was collected and analyzed qualitatively. Pattern - matching was explored to investigate qualitative data which ostensibly provided case stud y characteristics and objective project performance findings. Next, cross - case synthesis was used to amalgamate findings from each case study and to illustrate support for or against the propositions put forth in this study. Quantitative data was collecte d using an online survey and was analyzed using confirmatory factor analysis (CFA). CFA allowed the researcher to assess the measurement model against various model fit statistics. Next, SEM was conducted regarding the full measurement model. In doing so, limited support for multilevel data was found as such dependent variables in the findings are presented only for individual performance . Th e researcher continued with multivariate regression analyses to validate support for or against hypotheses asserted i n this study . The findings are presented against the five research goals and objectives. 1. Findings from proposition tested in this study were all adequately supported. Those propositions were: 1) Project risk factors and the level of collaborative project delivery practices in partnered - projects are positively related; 2) Collaborative project delivery - projects are positively related; and, 3) t perception and project performance in partnered - projects are positively related. 195 a. Three new propositions were developed as a result of this study. Those propositions are: 1a) If collaborative project delivery practices are not positively aligned with the level of project risk in partnered - projects, then project performance will be negatively affected ; 2a) Higher collaborative project delivery practices require fewer goal alignment actions or metrics to hold the project team accountable in partnered - projec ts ; and, 3a) Collaborative delivery practices should accommodate continuous improvement in facilitating information exchange among team members in partnered - projects. 2. Findings from hypothesis testing demonstrated a relationship between individual performan ce and transactive memory systems (i.e., TMS r = 0.82 , S.E. = 0.14, p = .000 ). The full multivariate model connecting individual performance to goal alignment and transactive memory systems was also supported (i.e., Adjusted R 2 = 0.70, p < 0.001). The vari ables in this study explained 70 percent of the variance (i.e., Adjusted R Square = 0.70) in individual performance perceptions for case study project team members. 3. Study metrics used in quantitative data collection and analysis are innovative in AEC resea rch, thus this study provides validity and posits them for future research. The latent constructs and factors are shown in section 6.3. 4. Co llaborative project delivery and its practices are furthered in this study. This was achieved by summarizing characte ristics of Some of these characteristics are positive while others should be changed to mitigate the impact on performance outcomes. For example, collaborative project delivery 196 practices found in partnering affor cost and schedule growth. Meanwhile, some other key attributes were: a. Easier alignment of collaborative project delivery practices when project risk factors are relaxed; b. Avoid disincentives such as liquidat ed damages unless subsequent incentives or risk sharing arrangements are included in contracts, otherwise gains in solidarity and trust are apt to suffer; c. Focus on a core set of goal alignment metrics that are detailed around a clear project objective to m inimize confusion and social loafing tendencies; and, d. Expect goal alignment to deviate over project life - cycle and should make provisions to continuously reinforce them. 5. The next section details the contributions this research has made to the body of know ledge. There were several theoretical contributions this study made to the body of knowledge. These contributions are listed below. 1. Researchers over the years have alluded to cognitive behaviors and social norms as p otential moderators between performance outcomes and collaborative project delivery approaches (Baiden et al., 2003; Baiden et al., 2006; F ranz et al., 2016; A. Sparkling, 2014; Suprapto et al., 2015; Suprapto, Bakker, Mooi, & Moree, 2014) . This 197 study takes a step further and posits clear metrics to understand behaviors goa l alignment and team dynamics via transactive memory systems . Moreover, it offers an alternative approach to assess performance outcomes reaching beyond traditional cost, schedule, quality, and safety performance measures. It positioned a measurement model for individual performance which picks up on perceptions often masked in other hard performance metrics. 2. This study also directs attention to relational governance theory as explanatory for the unique dynamics research finds in team integration (Poppo & Zenger, 2002; Zaheer & V enkatraman, 1995) . Collaborative project delivery approaches such as partnering and from opportunism . This shifts the foci from TCE theory where individuals are concerned with the costs of the exchange t o one focused on solidarity, flexibility, and trust. Within the mindset, project teams are able to coalesce around shared project goals and objectives anticipating greater benefits than going at it alone. 3. This study also brings in social loafing theory (Lam, 2015) by taking into account the challenges of goal alignment. Maint aining goal alignment becomes problematic as the number of performance measures increases or when competing messages are sent. For instance, codifying disincentives in contracts without subsequent incentives or rewards works against goal congruence. In fac t, it sends projec t team members backward toward opportunistic behaviors. Moreover, some may superficially acquiesce around collaborative project delivery practices rather than embrace the benefits (i.e., social loafing). 198 Despite all the findings and theoretical implications, this study does have limitations. Results were based on a limited number of survey respondents and a small case study sample (i.e., case studies = 6; N = 48) . It is important to point o ut that 5 5% of the survey respondents represented other team members with owners at hand (i.e., internal facilitator s ) . In other words, the quantitative findings are heavily skewed towards their perception s while 25% was that of the contractors and subcontractors. Another limitation of th is study is that does not control for the other characteristics of these projects. As an example, the case studies were not representative of the U.S. as (2) were from the Midwest while (4) were located in the West. A similar issue is found with the project delivery approach followed among case studies. Of the six case studies, five used design - bid - build while only one utilized design - build . This partly offers an explanation for findings which suggest early contractor involvement departs from literature because this is a limitation of DBB. Other s tudy controls future researcher may consider are to include projects following varied project delivery approaches (e.g., tradition al DBB, DB, partnering, and IPD), sizes, and locations. Findings from a larger data set may show disparate relationships than those found in this study. Additionally, a dataset comprised of more survey respondents may lend support for the additional direct (i.e., between individual performance and goal alignment ) and moderated (i.e., TMS moderating goal alignment and individual performance ) relationship suggested by the researcher. Further researchers may also find with a larger data set, unlike this study, a nested or multilevel data structure will allow for full SEM across various levels (e.g., case study projects, project teams, etc.) . 199 As a final thought, this study finds it is important for researchers and practitioners to advance project - based collabo ration as a means to improve the certainty of project performance outcomes . It used a mixed - methods approach to understand survey and case study data in its attempt to address this unique challenge. The contributions from this study are aimed to continue m oving research towards stronger metrics to assess behavioral and social dynamics within AEC project teams. The will allow us to parse out core attributes inhibiting team integration and performance. Although this study did not support multilevel modeling, t he AEC industry is replete with projects yearning for deeper analyses and inspect through this lens. 200 APPENDICES 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 Mplus Code and Output for Confirmatory Factor Analysis of Measurement Model using Sum med Scale Scores INPUT INSTRUCTIONS TITLE: PARTNE RING DATA CONFIRMATORY FACTOR ANALYSIS FULL MODEL DATA: FILE IS Partnering Reduced4 Rev3.dat; VARIABLE: NAMES ARE PROJ GA1 - GA5 CO_SUM CO_AVE CR_SUM CR_AVE SP_SUM SP_AVE PP_SUM PP_AVE CP_SUM CP_AVE TP_SUM TP_AVE ROLE OWNER CONTR OTHER; USEVARIABLES = GA1 GA3 - GA5 CO_SUM CR_SUM SP_SUM PP_SUM CP_SUM TP_SUM; CATEGORICAL IS GA1 GA3 - GA5; MODEL: F4 BY PP_SUM* CP_SUM TP_SUM; F8 BY CO_SUM* CR_SUM SP_SUM; F9 BY GA1* GA3 - GA5; F4@1; F8@1; F9@1; F4 WITH F8@.99; OUTPUT: TECH1 216 INPUT READING TERMINATED NORMALLY PARTNERING DATA CONFIRMATORY FACTOR ANALYSIS FULL MODEL SUMMARY OF ANALYSIS Number of groups 1 Number of observations 48 Number of dependent variables 10 Number of independent variables 0 Number of continuous latent var iables 3 Observed dependent variables Continuous CO_SUM CR_SUM SP_SUM PP_SUM CP_SUM TP_SUM Binary and ordered categorical (ordinal) GA1 GA3 GA4 GA5 Continuous latent variables F4 F8 F9 Estimator WLSMV Maximum number of iterations 1000 Convergence criterion 0.500D - 04 Maximum number of steepest descent iterations 20 Parameterization DELTA Link PROBIT 217 Input data file(s) Partnering Reduced4 Rev3.dat Input data format FREE UNIVARIATE PROPORTIONS AND COUNTS FOR CATEGORICAL VARIABLES GA1 Category 1 0.021 1.000 Category 2 0.229 11.000 Category 3 0.750 36.000 GA3 Category 1 0.250 12.000 Category 2 0.750 36.000 GA4 C ategory 1 0.292 14.000 Category 2 0.708 34.000 GA5 Category 1 0.021 1.000 Category 2 0.354 17.000 Category 3 0.625 30.000 UNIVARIATE SAMPLE STATISTICS UNIV ARIATE HIGHER - ORDER MOMENT DESCRIPTIVE STATISTICS Variable/ Mean/ Skewness/ Minimum/ % with Percentiles Sample Size Variance Kurtosis Maximum Min/Max 20%/60% 40%/80% Median 218 CO_SUM 8.417 - 0.763 5.000 6.25% 7.000 8.000 9.000 48.000 2.285 - 0.377 10.000 31.25% 9.000 10.000 CR_SUM 13.375 - 1.515 6.000 2.08% 12.000 13.000 14.000 48.000 3.401 3.260 15.000 39.58% 14.000 15.000 SP_SUM 13.771 - 0.838 10.000 2.08% 13.000 13.000 14.000 48.000 1.510 0.250 15.000 37.50% 14.000 15.000 PP_SUM 23.521 - 0.285 12.000 2.08% 18.000 21.000 24.000 48.000 29.833 - 1.119 30.000 29.17% 26.000 30.000 CP_SUM 13.833 - 1.653 8.000 2.08% 12.000 14.000 15.000 48.000 2.722 2.478 15.000 52.08% 15.000 15.000 TP_SUM 25.854 - 0.836 17.000 2.08% 23.000 25.000 27.000 48.000 12.500 - 0.252 30.000 10.42% 28.000 29.000 THE MODEL ESTIMATION TERMINATED NORMALLY MODEL FIT INFORMATION Number of Free Parameters 30 Chi - Square Test of Model Fit Val ue 38.458* Degrees of Freedom 33 P - Value 0.2360 * The chi - square value for MLM, MLMV, MLR, ULSMV, WLSM and WLSMV cannot be used for chi - square difference testing in the reg ular way. MLM, MLR and WLSM chi - square difference testing is described on the Mplus website. MLMV, WLSMV, and ULSMV difference testing is done using the DIFFTEST option. 219 RMSEA (Root Mean Square Error Of Approximation) Estimate 0.059 90 Percent C.I. 0.000 0.126 Probability RMSEA <= .05 0.404 CFI/TLI CFI 0.974 TLI 0.965 Chi - Square Tes t of Model Fit for the Baseline Model Value 256.302 Degrees of Freedom 45 P - Value 0.0000 WRMR (Weighted Root Mean Square Residual) Value 0.457 MODEL RESULTS Two - Tailed Estimate S.E. Est./S.E. P - Value F4 BY PP_SUM 3.116 1.021 3.053 0.002 CP_SUM 1.11 3 0.246 4.520 0.000 TP_SUM 2.762 0.604 4.574 0.000 220 F8 BY CO_SUM 1.030 0.258 3.986 0.000 CR_SUM 1.684 0.192 8.758 0.000 SP_SUM 0.726 0.169 4.292 0.000 F9 BY GA1 0.625 0.108 5.798 0.000 GA3 0.847 0.074 11.434 0.000 GA4 0.933 0.071 13.235 0.000 GA5 0.800 0.081 9.906 0.000 F4 WITH F8 0.990 0.000 999.000 999.000 F9 WITH F4 0.908 0.058 15.711 0.000 F8 0.906 0.058 15.581 0.000 Intercepts CO_SUM 8.417 0.273 30.884 0.000 CR_SUM 13.375 0.355 37.719 0.000 SP_SUM 13.771 0.214 64.410 0.000 PP_SUM 23.521 0.827 28.4 32 0.000 CP_SUM 13.833 0.382 36.256 0.000 TP_SUM 25.854 0.658 39.267 0.000 221 Thresholds GA1$1 - 2.037 0.411 - 4.952 0.000 GA1$2 - 0.674 0.197 - 3.429 0.001 GA3$1 - 0.674 0.197 - 3.429 0.001 GA4$1 - 0.549 0.191 - 2.870 0.004 GA5$1 - 2.037 0.411 - 4.952 0.000 GA5$2 - 0.319 0.184 - 1.729 0.084 Variances F4 1.000 0.000 999.000 999.000 F8 1.000 0.000 999.000 999.000 F9 1.000 0.000 999.000 999.000 Residual Variances CO_SUM 1. 224 0.257 4.768 0.000 CR_SUM 0.565 0.257 2.198 0.028 SP_SUM 0.983 0.204 4.815 0.000 PP_SUM 20.119 5.906 3.407 0.001 CP_SUM 1.484 0.323 4.599 0.000 TP_SUM 4.869 0.953 5.110 0.000 R - SQUARE Observed Residual Variable Estimate Variance GA1 0.391 0.609 GA3 0.717 0 .283 GA4 0.871 0.129 222 GA5 0.640 0.360 CO_SUM 0.464 CR_SUM 0.834 SP_SUM 0.349 PP_SUM 0.326 CP_SUM 0.455 TP_SUM 0.6 10 QUALITY OF NUMERICAL RESULTS Condition Number for the Information Matrix 0.343E - 03 (ratio of smallest to largest eigenvalue) TECHNICAL 1 OUTPUT PARAMETER SPECIFICATION TAU GA1$1 GA1$2 GA3$1 GA4$1 GA5$1 ________ ________ ________ ________ ________ 1 2 3 4 5 TAU GA5$2 ________ 6 NU GA1 GA3 GA4 GA5 CO_SUM ________ ________ ________ ________ ________ 223 0 0 0 0 7 NU CR_SUM SP_SUM PP_SUM CP_SUM TP_SUM ________ ________ ________ ________ ________ 8 9 10 11 12 LAMBDA F4 F8 F9 ________ ________ ________ GA1 0 0 13 GA3 0 0 14 GA4 0 0 15 GA5 0 0 16 CO_SUM 0 17 0 CR_SUM 0 18 0 SP_SUM 0 19 0 PP_SUM 20 0 0 CP_SUM 21 0 0 TP_SUM 22 0 0 THETA GA1 GA3 GA4 GA5 CO_SUM ________ ________ _______ _ ________ ________ GA1 0 GA3 0 0 GA4 0 0 0 GA5 0 0 0 0 224 CO_SUM 0 0 0 0 23 CR_SUM 0 0 0 0 0 SP_SUM 0 0 0 0 0 PP_SUM 0 0 0 0 0 CP_SUM 0 0 0 0 0 TP_SUM 0 0 0 0 0 THETA CR_SUM SP_SUM PP_SUM CP_SUM TP_SUM ________ ________ ________ ________ ________ CR_SUM 24 SP_SUM 0 25 PP_SUM 0 0 26 CP_SUM 0 0 0 27 TP_SUM 0 0 0 0 28 ALPHA F4 F8 F9 ________ ________ ________ 0 0 0 BETA F4 F8 F9 ________ ________ ________ F4 0 0 0 F8 0 0 0 F9 0 0 0 225 PSI F4 F8 F9 ________ ________ ________ F4 0 F8 0 0 F9 29 30 0 STARTING VALUES TAU GA1 $1 GA1$2 GA3$1 GA4$1 GA5$1 ________ ________ ________ ________ ________ - 2.037 - 0.674 - 0.674 - 0.549 - 2.037 TAU GA5$2 ________ - 0.319 NU GA1 GA3 GA4 GA5 CO_SUM ________ ________ ________ ________ ________ 0.000 0.000 0.000 0.000 8.417 NU CR_SUM SP_SUM PP_SUM CP_SUM TP_SUM ________ ________ ________ ________ ________ 13.375 13.771 23.521 13.833 25.854 226 LAMBDA F4 F8 F9 ________ ________ ________ GA1 0.000 0.000 1.000 GA3 0.000 0.000 1.000 GA4 0.000 0.000 1.000 GA5 0.000 0.000 1.000 CO_SUM 0.000 1.000 0.000 CR_SUM 0.000 1.000 0.000 SP_SUM 0.000 1.000 0.000 PP_SUM 1.000 0.000 0.000 CP_SUM 1.000 0.000 0.000 TP_SUM 1.000 0.000 0.000 THETA GA1 GA3 GA4 GA5 CO_SUM ________ ____ ____ ________ ________ ________ GA1 1.000 GA3 0.000 1.000 GA4 0.000 0.000 1.000 GA5 0.000 0.000 0.000 1.000 CO_SUM 0.000 0.000 0.000 0.000 1.142 CR_SUM 0.000 0.000 0.000 0.000 0.000 SP_SUM 0.000 0.000 0.000 0.000 0.000 PP_SUM 0.000 0.000 0.000 0 .000 0.000 CP_SUM 0.000 0.000 0.000 0.000 0.000 TP_SUM 0.000 0.000 0.000 0.000 0.000 227 THETA CR_SUM SP_SUM PP_SUM CP_SUM TP_SUM ________ ________ ________ ________ ________ CR_SUM 1.701 SP_SUM 0.000 0.755 PP_SUM 0.000 0.000 14.916 CP_SUM 0.000 0.000 0.000 1.361 TP_SUM 0.000 0.000 0.000 0.000 6.250 ALPHA F4 F8 F9 ________ ________ ________ 0.000 0.000 0.000 BETA F4 F8 F9 ________ ________ ________ F4 0.000 0.000 0.000 F8 0.000 0.000 0.000 F9 0.000 0.000 0.000 PSI F4 F8 F9 ________ ________ ________ F4 1.000 F8 0.990 1.000 F9 0.000 0.000 1.000 228 Mplus Code and Output for Confirmatory Factor Analysis of Goal Alignment INPUT INSTRUCTIONS TITLE: PARTNERING DATA CONFIRMATORY FACTOR ANALYSIS GOAL ALIGNMENT DATA: FILE IS Partnering Reduced4 Rev1.dat; VARIABLE: NAMES ARE PROJ GA1 - GA5 CO1 - CO3 CR1 - CR4 SP1 - SP3 PP1 - PP8 CP1 - CP4 TP1 - TP7 ROLE OWNER CONT OTHER; USEVARIABLES = GA1 - GA5; CATEGORICAL IS GA1 - GA5; MODEL: F9 BY GA1* GA2 GA3 GA4 GA5; F9@1; INPUT READING TERMINATED NORMALLY PARTNERING DATA CONFIRMATORY FACTOR ANALYSIS GOAL ALIGNMENT SUMMARY OF ANALYSIS Number of groups 1 Number of o bservations 48 Number of dependent variables 5 Number of independent variables 0 Number of continuous latent variables 1 Observed dependent variables Binary an d ordered categorical (ordinal) GA1 GA2 GA3 GA4 GA5 229 Continuous latent variables F9 Estimator WLSMV Maximum number of iterations 1000 Convergence cri terion 0.500D - 04 Maximum number of steepest descent iterations 20 Parameterization DELTA Link PROBIT Input data file (s) Partnering Reduced4 Rev1.dat Input data format FREE UNIVARIATE PROPORTIONS AND COUNTS FOR CATEGORICAL VARIABLES GA1 Category 1 0.021 1.000 Category 2 0.229 11.000 Category 3 0.750 36.0 00 GA2 Category 1 0.042 2.000 Category 2 0.208 10.000 Category 3 0.750 36.000 GA3 Category 1 0.250 12.000 230 Category 2 0.750 36.000 GA4 Category 1 0.292 14.000 Category 2 0.708 34.000 GA5 Category 1 0.021 1.000 Category 2 0.354 17.000 Category 3 0.625 30.000 THE MODEL ESTIMATION TERMINATED NORMALLY MODEL FIT INFORMATION Number of Free Parameters 13 Chi - Square Test of Model Fit Value 7.612* Degrees of Freedom 5 P - Value 0.1790 * The chi - square va lue for MLM, MLMV, MLR, ULSMV, WLSM and WLSMV cannot be used for chi - square difference testing in the regular way. MLM, MLR and WLSM chi - square difference testing is described on the Mplus website. MLMV, WLSMV, and ULSMV difference testing is done using the DIFFTEST option. RMSEA (Root Mean Square Error Of Approximation) Estimate 0.104 90 Percent C.I. 0.000 0.244 Probability RMSEA <= .05 0.236 231 CFI/TLI CFI 0.992 TLI 0.984 Chi - Square Test of Model Fit for the Baseline Model Value 327.961 Degrees of Freedom 10 P - Value 0.0000 WRMR (Weighted Root Mean Square Residual) Value 0.499 MODEL RESULTS Two - Tailed Estimate S.E. Est./S.E. P - Value F9 BY GA1 0.791 0.080 9.881 0.000 GA2 0.986 0.060 16.300 0.000 GA3 0.764 0.069 11.093 0.000 GA4 0.920 0 .098 9.436 0.000 GA5 0.782 0.113 6.896 0.000 Thresholds GA1$1 - 2.037 0.411 - 4.952 0.000 GA1$2 - 0.674 0.197 - 3.429 0.001 GA2$1 - 1.732 0.324 - 5.348 0.000 GA2$2 - 0.674 0.197 - 3.429 0.001 232 GA3$1 - 0.674 0.197 - 3.429 0.001 GA4$1 - 0.549 0.191 - 2.870 0.004 GA5$1 - 2.037 0.41 1 - 4.952 0.000 GA5$2 - 0.319 0.184 - 1.729 0.084 Variances F9 1.000 0.000 999.000 999.000 R - SQUARE Observed Residual Variable Estimate Variance G A1 0.626 0.374 GA2 0.972 0.028 GA3 0.583 0.417 GA4 0.847 0.153 GA5 0.612 0.388 QUALITY OF NUMERICAL RESULTS Condition Number for the I nformation Matrix 0.121E - 01 (ratio of smallest to largest eigenvalue) 233 Mplus Code and Output for Confirmatory Factor Analysis of TMS INPUT INSTRUCTIONS TITLE: PARTNERING DATA CONFIRMATORY FACTOR ANALYSIS TMS DATA: FILE IS Partnering Reduced4 Rev1.dat; VARIABLE: NAMES ARE PROJ GA1 - GA5 CO1 - CO3 CR1 - CR4 SP1 - SP3 PP1 - PP8 CP1 - CP4 TP1 - TP7 ROLE OWNER CONT OTHER; USEVARIABLES = CO2 - CO3 CR1 - CR3 SP1 - SP3; CATEGORIC AL IS CO2 - CO3 CR1 - CR3 SP1 - SP3; MODEL: F5 BY CO2* CO3; F6 BY CR1* CR2 CR3; F7 BY SP1* SP2 SP3; F5@1; F6@1; F7@1; INPUT READING TERMINATED NORMALLY PARTNERING DATA CONFIRMATORY FACTOR ANAL YSIS TMS SUMMARY OF ANALYSIS Number of groups 1 Number of observations 48 Number of dependent variables 8 Number of independent variables 0 Number of continuous latent variables 3 234 Observed dependent variables Binary and ordered categorical (ordinal) CO2 CO3 CR1 CR2 CR3 SP1 SP2 SP3 Continuous latent variables F5 F6 F7 Estimator WLSMV Maximum number of iterations 1000 Convergence criterion 0.500D - 04 Maximum number of steepest descent iterations 20 Parameterization DELTA Link PROBIT Input data file(s) Partnering Reduced4 Rev1.dat Input data format FREE UNIVARIATE PROPORTIONS AND COUNTS FOR CATEGO RICAL VARIABLES CO2 Category 1 0.125 6.000 Category 2 0.042 2.000 Category 3 0.417 20.000 Category 4 0.417 20.000 CO3 Category 1 0.146 7.000 235 Cat egory 2 0.417 20.000 Category 3 0.438 21.000 CR1 Category 1 0.021 1.000 Category 2 0.042 2.000 Category 3 0.271 13.000 Category 4 0.667 32.000 CR2 Category 1 0.042 2.000 Category 2 0.042 2.000 Category 3 0.375 18.000 Category 4 0.542 26.000 CR3 Category 1 0.021 1.000 Category 2 0.06 2 3.000 Category 3 0.417 20.000 Category 4 0.500 24.000 SP1 Category 1 0.042 2.000 Category 2 0.375 18.000 Category 3 0.583 28.000 SP2 Category 1 0.021 1.000 Category 2 0.208 10.000 Category 3 0.771 37.000 SP3 Category 1 0.083 4.000 236 Category 2 0.333 16.000 Category 3 0.583 28. 000 THE MODEL ESTIMATION TERMINATED NORMALLY MODEL FIT INFORMATION Number of Fre e Parameters 31 Chi - Square Test of Model Fit Value 23.618* Degrees of Freedom 17 P - Value 0.1302 * The chi - square value for MLM, MLMV, MLR, ULSMV, WLSM and WLSMV cannot be used for chi - square difference testing in the regular way. MLM, MLR and WLSM chi - square difference testing is described on the Mplus website. M LMV, WLSMV, and ULSMV difference testing is done using the DIFFTEST option. RMSEA (Root Mean Square Error Of Approximation) Estimate 0.090 90 Percent C.I. 0.000 0.170 Probability RMSEA <= .05 0.222 CFI/TLI CFI 0.986 TLI 0.977 Chi - Square Test of Model Fit for the Baseline Model 237 Value 509.782 D egrees of Freedom 28 P - Value 0.0000 WRMR (Weighted Root Mean Square Residual) Value 0.515 MODEL RESULTS Two - Tailed Estimate S.E. Est./S.E. P - Value F5 BY CO2 0.939 0.081 11.618 0.000 CO3 0.804 0.099 8.104 0.000 F6 BY CR1 0.664 0.120 5.523 0.000 CR2 0.990 0.045 22.220 0.000 CR3 0.915 0.044 20.798 0.000 F7 BY SP1 0.419 0.147 2.845 0.004 SP2 0.399 0.155 2.583 0.010 SP3 0.923 0.157 5.861 0.000 F6 WITH F5 0.790 0.101 7.838 0.000 238 F7 WITH F5 0.715 0.147 4.865 0.000 F6 0.761 0.145 5.236 0.000 Thresholds CO2$1 - 1.150 0.232 - 4.961 0.000 CO2$2 - 0.967 0.215 - 4.493 0.000 CO2$3 0.210 0.182 1.154 0.249 CO3$1 - 1.054 0.223 - 4.736 0.000 CO3$2 0.157 0.182 0.866 0.387 CR1$1 - 2.037 0.411 - 4.952 0.000 CR1$2 - 1.534 0.284 - 5.400 0.000 CR1$3 - 0 .431 0.187 - 2.302 0.021 CR2$1 - 1.732 0.324 - 5.348 0.000 CR2$2 - 1.383 0.260 - 5.315 0.000 CR2$3 - 0.105 0.181 - 0.577 0.564 CR3$1 - 2.037 0.411 - 4.952 0.000 CR3$2 - 1.383 0.260 - 5.315 0.000 CR3$3 0.000 0.181 0.000 1.000 SP1$1 - 1.732 0.324 - 5.348 0.000 SP1$2 - 0.210 0.182 - 1.154 0.249 SP2$1 - 2.037 0.411 - 4.952 0.000 SP2$2 - 0.742 0.200 - 3.704 0.000 SP3$1 - 1.383 0.260 - 5.315 0.000 SP3$2 - 0.210 0.182 - 1.1 54 0.249 Variances 239 F5 1.000 0.000 999.000 999.000 F6 1.000 0.000 999.000 999.000 F7 1.000 0.000 999.000 999.000 R - SQUARE Observed Res idual Variable Estimate Variance CO2 0.882 0.118 CO3 0.646 0.354 CR1 0.441 0.559 CR2 0.981 0.019 CR3 0.838 0.162 SP1 0.176 0.824 SP2 0.159 0.841 SP3 0.852 0.148 QUALITY OF NUMERICAL RESULTS Condition Number for the Information Matrix 0.547E - 02 (ratio of smallest to largest eigenv alue) 240 Mplus Code and Output for Confirmatory Factor Analysis of Individual/Team Performance INPUT INSTRUCTIONS TITLE: PARTNERING DATA CONFIRMATORY FACTOR ANALYSIS PERFORMANCE DATA: FILE IS Partnering Reduced4 Rev1.dat; VARIABLE: NAME S ARE PROJ GA1 - GA5 CO1 - CO3 CR1 - CR4 SP1 - SP3 PP1 - PP8 CP1 - CP4 TP1 - TP7 ROLE OWNER CONT OTHER; USEVARIABLES = PP2 - PP5 PP7 PP8 CP2 - CP4 TP1 - TP6; CATEGORICAL IS PP2 - PP5 PP7 PP8 CP2 - CP4 TP1 - TP6; MODEL: F1 BY PP2* PP3 PP4 PP5 PP7 PP8; F2 BY CP2* CP3 CP4; F3 BY TP1* TP2 TP3 TP4 TP5 TP6; F1@1; F2@1; F3@1; INPUT READING TERMINATED NORMALLY PARTNERING DATA CONFIRMATORY FACTOR ANALYSIS PERFORMANCE SUMMARY OF ANALYSIS Number of groups 1 Number of observations 48 Number of dependent variables 15 Number of independent variables 0 Numbe r of continuous latent variables 3 241 Observed dependent variables Binary and ordered categorical (ordinal) PP2 PP3 PP4 PP5 PP7 PP8 CP2 CP3 CP4 TP1 TP2 TP3 TP4 TP5 TP6 Continuous latent variables F1 F2 F3 Estimator WLSMV Maximum number of iterations 1000 Convergence criterion 0. 500D - 04 Maximum number of steepest descent iterations 20 Parameterization DELTA Link PROBIT Input data file(s) Partnering Reduced4 Rev1.dat Input dat a format FREE UNIVARIATE PROPORTIONS AND COUNTS FOR CATEGORICAL VARIABLES PP2 Category 1 0.042 2.000 Category 2 0.208 10.000 Category 3 0.146 7.000 Category 4 0.042 2.000 Category 5 0.562 27.000 242 PP3 Category 1 0.188 9.000 Category 2 0.042 2.000 Category 3 0.771 37.000 PP4 Category 1 0.396 19.000 Category 2 0.104 5.000 Category 3 0.021 1.000 Category 4 0.479 23.000 PP5 Category 1 0.021 1.000 Category 2 0.396 19.000 Category 3 0.104 5.000 Category 4 0.479 23.000 PP7 Category 1 0.021 1.000 Category 2 0.271 13.000 Category 3 0.167 8.000 Category 4 0.062 3.000 Category 5 0.479 23.000 P P8 Category 1 0.167 8.000 Category 2 0.021 1.000 Category 3 0.812 39.000 CP2 Category 1 0.021 1.000 Category 2 0.208 10.000 243 Category 3 0.771 37.000 CP3 Category 1 0.042 2.000 Category 2 0.250 12.000 Category 3 0.708 34.000 CP4 Category 1 0.042 2.000 Category 2 0.042 2.000 Catego ry 3 0.312 15.000 Category 4 0.604 29.000 TP1 Category 1 0.062 3.000 Category 2 0.167 8.000 Category 3 0.479 23.000 Category 4 0.292 14.000 TP2 Category 1 0.021 1.000 Category 2 0.042 2.000 Category 3 0.042 2.000 Category 4 0.396 19.000 Category 5 0.500 24.000 TP3 Category 1 0.042 2.000 Category 2 0.083 4.000 Category 3 0.333 16.000 Category 4 0.542 26.000 TP4 244 Category 1 0.021 1.000 Category 2 0.062 3.000 Category 3 0.312 15.000 Category 4 0.604 29.000 TP5 Category 1 0.042 2.000 Category 2 0.083 4.000 Category 3 0.312 15.000 Category 4 0.562 27.000 TP6 Category 1 0.188 9.000 Category 2 0.354 17.000 Category 3 0.458 22.000 THE MODEL ESTIMATION TERMINATED NORMALLY MODEL FIT INFORMATION Number of Free Parameters 61 Chi - Square Test of Model Fit Value 104.016* Degrees of Freedom 87 P - Value 0.1031 * The chi - square value for MLM, MLMV, MLR, ULSMV, WLSM and WLSMV cannot be used for chi - squar e difference testing in the regular way. MLM, MLR and WLSM chi - square difference testing is described on the Mplus website. MLMV, WLSMV, and ULSMV difference testing is done using the DIFFTEST option. 245 RMSEA (Root Mean Square Error Of Approximati on) Estimate 0.064 90 Percent C.I. 0.000 0.106 Probability RMSEA <= .05 0.313 CFI/TLI CFI 0.972 TLI 0.966 Chi - Square Test of Model Fit for the Baseline Model Value 702.098 Degrees of Freedom 105 P - Value 0.0000 WRMR (Weighted Root Mean Square Residual) Value 0.706 MODEL RESULTS Two - Tailed Estimate S.E. Est./S.E. P - Value F1 BY PP2 0.623 0.104 5.979 0. 000 PP3 0.904 0.103 8.736 0.000 PP4 0.832 0.088 9.446 0.000 PP5 0.745 0.086 8.629 0.000 PP7 0.559 0.099 5.676 0.000 PP8 0.737 0.148 4.982 0.000 246 F2 BY CP2 0.914 0.100 9.126 0.000 CP3 0.838 0.083 10.113 0.000 CP4 0.785 0.104 7.573 0.00 0 F3 BY TP1 0.763 0.091 8.358 0.000 TP2 0.805 0.068 11.806 0.000 TP3 0.734 0.062 11.912 0.000 TP4 0.876 0.069 12.776 0.000 TP5 0.732 0.069 10.585 0.000 TP6 0.645 0.098 6.577 0.000 F2 WITH F1 0.517 0.126 4.097 0.000 F3 WITH F1 0.652 0.111 5.860 0.000 F2 0.695 0.086 8.050 0.000 Thresholds PP2$1 - 1.732 0.324 - 5.348 0.000 PP2$2 - 0.674 0.197 - 3.429 0.001 PP2$3 - 0. 264 0.183 - 1.442 0.149 PP2$4 - 0.157 0.182 - 0.866 0.387 PP3$1 - 0.887 0.209 - 4.239 0.000 247 PP3$2 - 0.742 0.200 - 3.704 0.000 PP4$1 - 0.264 0.183 - 1.442 0.149 PP4$2 0.000 0.181 0.000 1.000 PP4$3 0.052 0.181 0.289 0.773 PP5$1 - 2.037 0.411 - 4.952 0.000 PP5$2 - 0.210 0.182 - 1.154 0.249 PP5$3 0.052 0.181 0.289 0.773 PP7$1 - 2.037 0.411 - 4.952 0.000 PP7$2 - 0.549 0.191 - 2.870 0.004 PP7$3 - 0.105 0.181 - 0.57 7 0.564 PP7$4 0.052 0.181 0.289 0.773 PP8$1 - 0.967 0.215 - 4.493 0.000 PP8$2 - 0.887 0.209 - 4.239 0.000 CP2$1 - 2.037 0.411 - 4.952 0 .000 CP2$2 - 0.742 0.200 - 3.704 0.000 CP3$1 - 1.732 0.324 - 5.348 0.000 CP3$2 - 0.549 0.191 - 2.870 0.004 CP4$1 - 1.732 0.324 - 5.348 0.000 CP4$2 - 1.383 0.260 - 5.315 0.000 CP4$3 - 0.264 0.183 - 1.442 0.149 TP1$1 - 1.534 0.284 - 5.400 0.000 TP1$2 - 0.742 0.200 - 3.704 0.000 TP1$3 0.549 0.191 2.870 0.004 TP2$1 - 2.037 0.411 - 4.952 0.000 TP2$2 - 1.534 0.284 - 5.400 0.000 TP2$3 - 1.258 0.244 - 5.159 0.000 TP2$4 0.000 0.181 0.000 1.000 248 TP3$1 - 1.732 0.324 - 5.348 0.000 TP3$2 - 1.150 0.232 - 4.961 0.000 TP3$3 - 0.105 0.181 - 0.577 0.564 TP4$1 - 2.03 7 0.411 - 4.952 0.000 TP4$2 - 1.383 0.260 - 5.315 0.000 TP4$3 - 0.264 0.183 - 1.442 0.149 TP5$1 - 1.732 0.324 - 5.348 0.000 TP5$2 - 1.150 0 .232 - 4.961 0.000 TP5$3 - 0.157 0.182 - 0.866 0.387 TP6$1 - 0.887 0.209 - 4.239 0.000 TP6$2 0.105 0.181 0.577 0.564 Variances F1 1.000 0.000 999.000 999.000 F2 1.000 0.000 999.000 999.000 F3 1.000 0.000 999.000 999.000 R - SQUARE Observed Residual Variable Estimate Variance PP2 0.388 0.612 PP3 0.817 0.183 PP4 0.692 0.308 PP5 0.555 0.445 PP7 0.313 0.687 PP8 0.544 0.456 249 CP2 0.8 35 0.165 CP3 0.703 0.297 CP4 0.617 0.383 TP1 0.582 0.418 TP2 0.648 0.352 TP3 0.539 0.461 TP4 0.768 0.232 TP5 0.536 0.464 TP6 0.417 0.583 QUALITY OF NUMERICAL RESULTS Condition Number for the Information Matrix 0.596E - 02 (ratio of smallest to largest eigenvalue) 250 Mplus Input Code and Output for SEM Reliability of Study Latent Variable TMS INPUT INSTRUCTIONS TITLE: PARTNERING DATA CONFIRMATORY FACTOR ANALYSIS TMS DATA: FILE IS Partnering Redu ced4 Rev1.dat; VARIABLE: NAMES ARE PROJ GA1 - GA5 CO1 - CO3 CR1 - CR4 SP1 - SP3 PP1 - PP8 CP1 - CP4 TP1 - TP7 ROLE OWNER CONT OTHER; USEVARIABLES = CO2 - CO3 CR1 - CR3 SP1 - SP3; CATEGORICAL IS CO2 - CO3 CR1 - CR3 SP1 - SP3; MODEL: F5 BY CO2 CO3; F6 BY CR1 CR2 CR3; F7 BY SP1 SP2 SP3; F5@1; F6@1; F7@1; F8 BY F5* (P1) F6 F7 (P2 - P3); F5 - F7 (P4 - P6); F8@1; MODEL CONSTRAINT: NEW(COMP_REL); COMP_REL =(P1+P2+P3)**2/ ((P1+P2+P3)**2+P4+P5+P6); OUTPUT: CINTERVAL; 251 INPUT READING TERMINATED NORMALLY PARTNERING DATA CONFIRMATORY FACTOR ANALYSIS TMS SUMMARY OF ANALYSIS Number o f groups 1 Number of observations 48 Number of dependent variables 8 Number of independent variables 0 Number of continuous l atent variables 4 Observed dependent variables Binary and ordered categorical (ordinal) CO2 CO3 CR1 CR2 CR3 SP1 SP2 SP3 Continuous latent variables F5 F6 F7 F8 Estimator WLSMV Maximum number of iterations 1000 Convergence criterion 0.500D - 04 Maximum number of steepest descent iterations 20 Parameteri zation DELTA Link PROBIT Input data file(s) Partnering Reduced4 Rev1.dat 252 Input data format FREE UNIVARIATE PROPORTIONS AND COUNTS FOR CATEGORICAL VARIABLES CO2 Category 1 0.125 6.000 Category 2 0.042 2.000 Category 3 0.417 20.000 Category 4 0.417 20.000 CO3 Category 1 0.146 7.000 Category 2 0.417 20.000 Category 3 0.438 21.000 CR1 Category 1 0.021 1.000 Category 2 0.042 2.000 Category 3 0.271 13.000 Category 4 0.667 32.000 CR2 Category 1 0.042 2.000 Category 2 0.042 2.000 Category 3 0.375 18.000 Category 4 0.542 26.000 CR3 Category 1 0.021 1.000 Category 2 0.062 3.000 Category 3 0.417 20.000 Category 4 0.500 24.000 253 SP1 Category 1 0.042 2.000 Category 2 0.375 18.000 Category 3 0.583 28.000 SP2 Category 1 0.021 1.000 Category 2 0.208 10.000 Category 3 0.771 37.000 SP3 Category 1 0.083 4.000 Category 2 0.333 16.000 Category 3 0.583 28.000 THE MODEL ESTIMATION TERMINATED NORMALLY MODEL FIT INFORMATION Number of Free Parameters 31 Chi - Square Test of Model Fit Value 23.618* Degrees of Freedom 17 P - Value 0 .1302 * The chi - square value for MLM, MLMV, MLR, ULSMV, WLSM and WLSMV cannot be used for chi - square difference testing in the regular way. MLM, MLR and WLSM chi - square difference testing is described on the Mplus website. MLMV, WLSMV, and ULSMV difference testing is done using the DIFFTEST option. 254 RMSEA (Root Mean Square Error Of Approximation) Estimate 0.090 90 Percent C.I. 0.000 0.170 Probability RMSEA <= .05 0.222 CFI/TLI CFI 0.986 TLI 0.977 Chi - Square Test of Model Fit for the Baseline Model Value 509.782 Degrees of Freedo m 28 P - Value 0.0000 WRMR (Weighted Root Mean Square Residual) Value 0.515 MODEL RESULTS Two - Tailed Estim ate S.E. Est./S.E. P - Value F5 BY CO2 1.000 0.000 999.000 999.000 CO3 0.855 0.130 6.575 0.000 F6 BY CR1 1.000 0.000 999.000 999.000 255 C R2 1.491 0.266 5.607 0.000 CR3 1.378 0.251 5.497 0.000 F7 BY SP1 1.000 0.000 999.000 999.000 SP2 0.952 0.529 1.800 0.072 SP3 2.202 0.851 2.587 0.010 F8 BY F5 0.810 0.091 8.859 0.000 F6 0.609 0.111 5.469 0.000 F7 0.348 0.138 2.519 0.012 Thresholds CO2$1 - 1.150 0.232 - 4.961 0.000 CO2$2 - 0.967 0.215 - 4.493 0.000 CO2$3 0.210 0.182 1.154 0.249 CO3$1 - 1.054 0.223 - 4.736 0.000 CO3$2 0.157 0.182 0.866 0.387 CR1$1 - 2.037 0.411 - 4.952 0.000 CR1$2 - 1.534 0.284 - 5.400 0.000 CR1$3 - 0.431 0.187 - 2.302 0. 021 CR2$1 - 1.732 0.324 - 5.348 0.000 CR2$2 - 1.383 0.260 - 5.315 0.000 CR2$3 - 0.105 0.181 - 0.577 0.564 CR3$1 - 2.037 0.411 - 4.952 0.000 CR3$2 - 1.383 0.260 - 5.315 0.000 256 CR3$3 0.000 0.181 0.000 1.000 SP1$1 - 1.732 0.324 - 5.348 0.000 SP1$2 - 0.210 0.182 - 1.154 0.249 SP2$1 - 2.037 0.411 - 4.952 0.000 SP2$2 - 0.742 0.200 - 3.704 0.000 SP3$1 - 1.383 0.260 - 5.315 0.000 SP3$2 - 0.210 0.182 - 1.154 0.249 Variances F8 1.000 0.000 999.000 999.000 Residual Variances F5 0.227 0.176 1.286 0.198 F6 0.070 0.076 0.921 0.357 F7 0.055 0.054 1.005 0 .315 New/Additional Parameters COMP_REL 0.899 0.053 17.008 0.000 R - SQUARE Observed Residual Variable Estimate Variance CO2 0.882 0.118 CO3 0.646 0.354 CR1 0.441 0.559 CR2 0.981 0.019 257 CR3 0.838 0.162 SP1 0.176 0.824 SP2 0.159 0.841 SP3 0.852 0.148 L atent Variable Estimate F5 0.743 F6 0.841 F7 0.688 QUAL ITY OF NUMERICAL RESULTS Condition Number for the Information Matrix 0.532E - 03 (ratio of smallest to la rgest eigenvalue) CONFID ENCE INTERVALS OF MODEL RESULTS Lower .5% Lower 2.5% Lower 5% Estimate U pper 5% Upper 2.5% Upper .5% F5 BY CO2 1.000 1.000 1.000 1.000 1.000 1.000 1.000 CO3 0.520 0.600 0.641 0.855 1.069 1.110 1.191 F6 BY CR1 1.000 1.000 1.000 1.000 1.000 1.000 1.000 CR2 0.806 0.969 1.053 1.491 1.928 2.012 2.175 CR3 0.732 0.886 0.965 1.378 1.790 1.869 2.023 258 F7 BY SP1 1.000 1.000 1.000 1.000 1.000 1.000 1.000 SP2 - 0.410 - 0.085 0.082 0.952 1.822 1.989 2.315 SP3 0.010 0.534 0.802 2.202 3.601 3.869 4.393 F8 BY F5 0.574 0.631 0.659 0.810 0.960 0.989 1.045 F6 0.322 0.391 0.426 0.609 0.792 0.828 0.896 F7 - 0.008 0.077 0.121 0.348 0.575 0.618 0.703 Thresholds CO2$1 - 1.748 - 1.605 - 1.532 - 1.150 - 0.769 - 0.696 - 0.553 CO2$2 - 1.522 - 1.389 - 1.322 - 0.967 - 0.613 - 0.545 - 0.413 CO2$3 - 0.2 59 - 0.147 - 0.090 0.210 0.510 0.568 0.680 CO3$1 - 1.628 - 1.491 - 1.421 - 1.054 - 0.688 - 0.618 - 0.481 CO3$2 - 0.311 - 0.199 - 0.142 0.157 0.456 0 .513 0.625 CR1$1 - 3.096 - 2.843 - 2.713 - 2.037 - 1.360 - 1.231 - 0.977 CR1$2 - 2.266 - 2.091 - 2.001 - 1.534 - 1.067 - 0.977 - 0.802 CR1$3 - 0.913 - 0.798 - 0.739 - 0.431 - 0.123 - 0.064 0.051 CR2$1 - 2.566 - 2.366 - 2.264 - 1.732 - 1.199 - 1.097 - 0.898 CR2$2 - 2.053 - 1.893 - 1.811 - 1.383 - 0.955 - 0.873 - 0.713 CR2$3 - 0.572 - 0.460 - 0.403 - 0.105 0.194 0.251 0.362 CR3$1 - 3.096 - 2.843 - 2.713 - 2.037 - 1.360 - 1.231 - 0.977 CR3$2 - 2.053 - 1.893 - 1.811 - 1 .383 - 0.955 - 0.873 - 0.713 CR3$3 - 0.466 - 0.355 - 0.298 0.000 0.298 0.355 0.466 SP1$1 - 2.566 - 2.366 - 2.264 - 1.732 - 1.199 - 1.097 - 0.898 SP1$2 - 0.680 - 0.568 - 0.510 - 0.210 0.090 0.147 0.259 259 SP2$1 - 3.096 - 2.843 - 2.713 - 2.037 - 1.360 - 1.231 - 0.977 SP2$2 - 1.257 - 1.134 - 1.071 - 0.742 - 0.412 - 0.349 - 0.226 SP3$1 - 2.053 - 1.893 - 1.811 - 1.383 - 0.955 - 0.873 - 0.713 SP3$2 - 0.680 - 0.568 - 0.510 - 0.210 0.090 0.147 0.259 Variances F8 1 .000 1.000 1.000 1.000 1.000 1.000 1.000 Residual Variances F5 - 0.227 - 0.119 - 0.063 0.227 0.517 0.572 0.681 F6 - 0.126 - 0.079 - 0.055 0. 070 0.196 0.220 0.267 F7 - 0.086 - 0.052 - 0.035 0.055 0.144 0.161 0.195 New/Additional Parameters COMP_REL 0.763 0.795 0.812 0.899 0.986 1.002 1.035 260 Mplus Input Code and Output for SEM Reliability of Study Latent Variable Individual/Team Performance INPUT INSTRUCTIONS TITLE: PARTNERING DATA CONFIRMATORY FACTOR ANALYSIS PERFORMANCE DATA: FILE IS Partnering Reduced4 Rev1.dat; VARIABLE: NAMES ARE PROJ GA1 - GA5 CO1 - CO3 CR1 - CR4 SP1 - SP3 PP1 - PP8 CP1 - CP4 TP1 - TP7 ROLE OWNER CONT OTHER; USEVARIABLES = PP2 - PP5 PP7 PP8 CP2 - CP4 TP1 - TP6; CATEGORICAL IS PP2 - PP5 PP7 PP8 CP2 - CP4 TP1 - TP6; MODEL: F1 BY PP2 PP3 PP4 PP5 PP7 PP8; F2 BY CP2 CP3 CP4; F3 BY TP1 TP2 TP3 TP4 TP5 TP6; F1@1; F2@1; F3@1; F4 BY F1* (P1) F2 F3 (P2 - P3); F1 - F3 (P4 - P6); F4@1; MODEL CONSTRAINT: NEW(COMP_REL); COMP_REL =(P1+P2+P3)**2/ ((P1+P2+P3)**2+P4+P5+P6); OUTPUT: CINTERVAL; 261 INPUT READING TERMINATED NORMALLY PARTNERING DATA CONFIRMATORY FACTOR ANALYSIS PERFORMANCE SUMMARY OF ANALYSIS Number of groups 1 Number of observations 48 Number of dependent variables 15 Number of independent variables 0 Numbe r of continuous latent variables 4 Observed dependent variables Binary and ordered categorical (ordinal) PP2 PP3 PP4 PP5 PP7 PP8 CP2 CP3 CP4 TP1 TP2 TP3 TP4 TP5 TP6 Continuous latent variables F1 F2 F3 F4 Estimator WLSMV Maximum number of iterations 1000 Convergence criterion 0.500D - 04 Maximum number of steepest descent iterations 20 Parameterization DELTA Link PROBIT Input data file(s) 262 Partnering Reduced4 Rev1.da t Input data format FREE UNIVARIATE PROPORTIONS AND COUNTS FOR CATEGORICAL VARIABLES PP2 Category 1 0.042 2.000 Category 2 0.208 10.000 Category 3 0.146 7.000 Category 4 0.042 2.000 Category 5 0.562 27.000 PP3 Category 1 0.188 9.000 Category 2 0.042 2.000 Category 3 0.771 37.000 PP4 Category 1 0.396 19.000 Category 2 0.104 5.000 Category 3 0.021 1.000 Category 4 0.479 23.000 PP5 Category 1 0.021 1.000 Category 2 0.396 19.000 Category 3 0.104 5.000 Category 4 0.479 23.000 PP7 Category 1 0.021 1.000 Category 2 0.271 13.000 263 Category 3 0.167 8.000 Category 4 0.062 3.000 Category 5 0.479 2 3.000 PP8 Category 1 0.167 8.000 Category 2 0.021 1.000 Category 3 0.812 39.000 CP2 Category 1 0.021 1.000 Category 2 0.208 10.000 Category 3 0.771 37.000 CP3 Category 1 0.042 2.000 Category 2 0.250 12.000 Category 3 0.708 34.000 CP4 Category 1 0.042 2.000 Category 2 0.042 2.000 Category 3 0.312 15.000 Category 4 0.604 29.000 TP1 Category 1 0.062 3.000 Category 2 0.167 8.000 Category 3 0.479 23.000 Category 4 0.292 14.000 TP2 Category 1 0.021 1.000 264 Category 2 0.042 2.000 Category 3 0.042 2.000 Category 4 0.396 19.000 Category 5 0.500 24.000 TP3 Category 1 0.042 2.000 Category 2 0.083 4.000 Category 3 0.333 16.000 Category 4 0.542 26.000 TP4 Category 1 0.021 1.000 Category 2 0.062 3.000 Category 3 0.312 15.000 Category 4 0.604 29.000 TP5 Category 1 0.042 2.000 Category 2 0.083 4.000 Category 3 0.312 15.000 Category 4 0.562 27. 000 TP6 Category 1 0.188 9.000 Category 2 0.354 17.000 Category 3 0.458 22.000 THE MODEL ESTIMATION TERMINATED NORMALLY MODEL FIT INFORMATION 265 Number of Free Parameters 61 Chi - Square Test of Model Fit Value 104.016* Degrees of Freedom 87 P - Value 0.1031 * The chi - square value for MLM, MLMV, MLR, ULSMV, WLSM and WLSMV cannot be used for chi - square difference testing in the regular way. MLM, MLR and WLSM chi - square difference testing is described on the Mplus website. MLMV, WLSMV, and ULSMV difference testing is done using the DIFFTEST option. RMSEA (Root Mean Square Error Of Approximation) Estimate 0.064 90 Percent C.I. 0.000 0.106 Probability RMSEA <= .05 0.313 CFI/TLI CFI 0.972 TLI 0.966 Chi - Square Test of Model Fit for the Baseline Model Value 702.098 Degrees of Freedom 105 P - Value 0.0000 266 WRMR (Weighted Root Mean Square Residual) Value 0.706 MODEL RESULTS Two - Tailed Estimate S.E. Est./S.E. P - Value F1 BY PP2 1.000 0.000 999.000 999.000 PP3 1.451 0.280 5.182 0.000 PP4 1.335 0.279 4.785 0.000 PP5 1.196 0.230 5.209 0.000 PP7 0.898 0.210 4.274 0.000 PP8 1.184 0.262 4.509 0.000 F2 BY CP2 1.000 0.000 999.000 999.000 CP3 0.917 0.154 5.955 0.000 CP4 0.859 0.171 5.026 0.000 F3 BY TP1 1.000 0.000 999.000 999.000 TP2 1.055 0.170 6.187 0.000 TP3 0.962 0.163 5.902 0.000 TP4 1.148 0.15 0 7.659 0.000 TP5 0.959 0.122 7.834 0.000 TP6 0.846 0.150 5.646 0.000 267 F4 BY F1 0.434 0.104 4.162 0.000 F2 0.678 0.134 5.050 0.000 F3 0.715 0.108 6.642 0.000 Thresholds PP2$1 - 1.732 0.324 - 5.348 0.000 PP2$2 - 0.674 0.197 - 3.429 0.001 PP2$3 - 0.26 4 0.183 - 1.442 0.149 PP2$4 - 0.157 0.182 - 0.866 0.387 PP3$1 - 0.887 0.209 - 4.239 0.000 PP3$2 - 0.742 0.200 - 3.704 0.000 PP4$1 - 0.264 0 .183 - 1.442 0.149 PP4$2 0.000 0.181 0.000 1.000 PP4$3 0.052 0.181 0.289 0.773 PP5$1 - 2.037 0.411 - 4.952 0.000 PP5$2 - 0.210 0.182 - 1.154 0.249 PP5$3 0.052 0.181 0.289 0.773 PP7$1 - 2.037 0.411 - 4.952 0.000 PP7$2 - 0.549 0.191 - 2.870 0.004 PP7$3 - 0.105 0.181 - 0.577 0.564 PP7$4 0.052 0.181 0.289 0.773 PP8$1 - 0.967 0.215 - 4.493 0.000 PP8$2 - 0.887 0.209 - 4.239 0.000 CP2$1 - 2.037 0.411 - 4.952 0.0 00 CP2$2 - 0.742 0.200 - 3.704 0.000 CP3$1 - 1.732 0.324 - 5.348 0.000 268 CP3$2 - 0.549 0.191 - 2.870 0.004 CP4$1 - 1.732 0.324 - 5.348 0.000 C P4$2 - 1.383 0.260 - 5.315 0.000 CP4$3 - 0.264 0.183 - 1.442 0.149 TP1$1 - 1.534 0.284 - 5.400 0.000 TP1$2 - 0.742 0.200 - 3.704 0.000 TP1$3 0.549 0.191 2.870 0.004 TP2$1 - 2.037 0.411 - 4.952 0.000 TP2$2 - 1.534 0.284 - 5.400 0.000 TP2$3 - 1.258 0.244 - 5.159 0.000 TP2$4 0.000 0.181 0.000 1.000 TP3$1 - 1.732 0.324 - 5.348 0.000 TP3$2 - 1.150 0.232 - 4.961 0.000 TP3$3 - 0.105 0.181 - 0.577 0.564 TP4$1 - 2.037 0.411 - 4.952 0.000 TP4$2 - 1.383 0.260 - 5.315 0.000 TP4$3 - 0.264 0.183 - 1.442 0.149 TP5$1 - 1.732 0.324 - 5.348 0.000 TP5$2 - 1.150 0.2 32 - 4.961 0.000 TP5$3 - 0.157 0.182 - 0.866 0.387 TP6$1 - 0.887 0.209 - 4.239 0.000 TP6$2 0.105 0.181 0.577 0.564 Variances F4 1.000 0.000 999.000 999.000 269 Residual Variances F1 0.200 0.089 2.247 0.025 F2 0.375 0.146 2.575 0.010 F3 0.072 0.115 0.626 0.531 New/Additional Par ameters COMP_REL 0.838 0.045 18.483 0.000 R - SQUARE Observed Residual Variable Estimate Variance PP2 0.388 0.612 PP3 0.817 0.183 PP4 0.692 0.308 PP5 0.555 0.445 PP7 0.313 0.687 PP8 0.544 0.456 CP2 0.835 0.165 CP3 0.703 0.297 CP4 0.617 0.383 TP1 0.582 0.418 TP2 0.648 0.352 TP3 0.539 0.461 TP4 0.768 0.232 TP5 0.536 0.464 TP6 0.417 0.583 270 La tent Variable Estimate F1 0.485 F2 0.551 F3 0.877 QUALITY OF NUMERICAL RESULTS Condition Number for the Information Matrix 0.946E - 03 (ratio of smallest to lar gest eigenvalue) CONFIDENCE INTERVALS OF MODEL RESULTS Lower .5% Lower 2.5% Lower 5% Estimate Upper 5% Upper 2.5% Upper .5% F1 BY PP2 1.000 1.000 1.000 1.000 1.000 1.000 1.000 PP3 0.730 0.902 0.990 1.451 1.912 2.000 2.172 PP4 0.616 0.788 0.876 1.335 1.794 1.882 2.054 PP5 0.605 0.746 0 .818 1.196 1.574 1.646 1.788 PP7 0.357 0.486 0.552 0.898 1.243 1.309 1.438 PP8 0.507 0.669 0.752 1.184 1.615 1.698 1.860 F2 BY CP2 1.000 1.000 1.000 1.000 1.000 1.000 1.000 CP3 0.521 0.615 0.664 0.917 1.171 1.219 1.314 CP4 0.419 0.524 0.57 8 0.859 1.141 1.195 1.300 F3 BY 271 TP1 1.000 1.000 1.000 1.000 1.000 1.000 1.000 TP2 0.616 0.721 0.774 1.055 1.335 1.389 1.494 TP3 0.542 0.643 0.694 0.962 1.231 1.282 1.382 TP4 0.762 0.854 0.902 1.148 1.395 1.442 1.534 TP5 0.644 0.719 0.758 0.959 1.160 1.199 1.274 TP6 0.460 0.552 0.599 0.846 1.092 1.139 1.232 F4 BY F1 0.165 0.230 0.262 0.434 0.606 0.638 0.7 03 F2 0.332 0.415 0.457 0.678 0.899 0.942 1.024 F3 0.437 0.504 0.538 0.715 0.892 0.925 0.992 Thresholds PP2$1 - 2.566 - 2.366 - 2.264 - 1.732 - 1.199 - 1.097 - 0.898 PP2$2 - 1.181 - 1.060 - 0.998 - 0.674 - 0.351 - 0.289 - 0.168 PP2$3 - 0.736 - 0.623 - 0.566 - 0.264 0.037 0.095 0.208 PP2$4 - 0.625 - 0.513 - 0.456 - 0.157 0.142 0.199 0.311 PP3$1 - 1.426 - 1.297 - 1.231 - 0.887 - 0.543 - 0.477 - 0.348 PP3$2 - 1.257 - 1.134 - 1.071 - 0. 742 - 0.412 - 0.349 - 0.226 PP4$1 - 0.736 - 0.623 - 0.566 - 0.264 0.037 0.095 0.208 PP4$2 - 0.466 - 0.355 - 0.298 0.000 0.298 0.355 0.466 PP4$3 - 0.414 - 0.302 - 0.245 0.052 0.350 0.407 0.518 PP5$1 - 3.096 - 2.843 - 2.713 - 2.037 - 1.360 - 1.231 - 0.977 PP5$2 - 0.680 - 0.568 - 0.510 - 0.210 0.090 0.147 0.259 PP5$3 - 0.414 - 0.302 - 0.245 0.052 0.350 0.407 0.518 PP7$1 - 3.096 - 2.843 - 2.713 - 2.037 - 1.360 - 1.231 - 0.977 PP7$2 - 1.041 - 0. 923 - 0.863 - 0.549 - 0.234 - 0.174 - 0.056 272 PP7$3 - 0.572 - 0.460 - 0.403 - 0.105 0.194 0.251 0.362 PP7$4 - 0.414 - 0.302 - 0.245 0.052 0.350 0.407 0.518 PP8$1 - 1.522 - 1.389 - 1.322 - 0.967 - 0.613 - 0.545 - 0.413 PP8$2 - 1.426 - 1.297 - 1.231 - 0.887 - 0.543 - 0.477 - 0.348 CP2$1 - 3.096 - 2.843 - 2.713 - 2.037 - 1.360 - 1.231 - 0.977 CP2$2 - 1.257 - 1.134 - 1.071 - 0.742 - 0.412 - 0.349 - 0.226 CP3$1 - 2.566 - 2.366 - 2.264 - 1.732 - 1.199 - 1.097 - 0.898 CP3$2 - 1.041 - 0.923 - 0.863 - 0.549 - 0.234 - 0.174 - 0.056 CP4$1 - 2.566 - 2.366 - 2.264 - 1.732 - 1.199 - 1.097 - 0.898 CP4$2 - 2.053 - 1.893 - 1.811 - 1.383 - 0.955 - 0.873 - 0.713 CP4$3 - 0.736 - 0.623 - 0.566 - 0.264 0.037 0.095 0.208 TP1$1 - 2.266 - 2.091 - 2.001 - 1.534 - 1.067 - 0.977 - 0.802 TP1$2 - 1.257 - 1.134 - 1.071 - 0.742 - 0.412 - 0.349 - 0.226 TP1$3 0.056 0.174 0.234 0.549 0.863 0.923 1.041 TP2$1 - 3.096 - 2.843 - 2.713 - 2.037 - 1.360 - 1.231 - 0.977 TP2$2 - 2.266 - 2.091 - 2.001 - 1.534 - 1.067 - 0.977 - 0.802 TP2$3 - 1.886 - 1.736 - 1.659 - 1.258 - 0.857 - 0.780 - 0.630 TP2$4 - 0.466 - 0.355 - 0.298 0.000 0.298 0.355 0.466 TP3$1 - 2.566 - 2.366 - 2.264 - 1.732 - 1.199 - 1.097 - 0.898 TP3$2 - 1.748 - 1.605 - 1.532 - 1.150 - 0.769 - 0.696 - 0.553 TP3$3 - 0.572 - 0.460 - 0.403 - 0.105 0.194 0.251 0.362 TP4$1 - 3.096 - 2.843 - 2.713 - 2.037 - 1.360 - 1.231 - 0.977 TP4$2 - 2.053 - 1.893 - 1.811 - 1.383 - 0.955 - 0.873 - 0.713 TP4$3 - 0.736 - 0.623 - 0.566 - 0.264 0.037 0.095 0.208 TP5$1 - 2.566 - 2.366 - 2.264 - 1.732 - 1.199 - 1.097 - 0.898 TP5$2 - 1.748 - 1.605 - 1.532 - 1.150 - 0.769 - 0.696 - 0.553 TP5$3 - 0.625 - 0.513 - 0.456 - 0.157 0.142 0.199 0.311 273 TP6$1 - 1.426 - 1.297 - 1.231 - 0.887 - 0.543 - 0.477 - 0.348 TP6$2 - 0.362 - 0.251 - 0.194 0.105 0.403 0.460 0.572 Variances F4 1.000 1.000 1.000 1.000 1.000 1.000 1.000 Residual Variances F 1 - 0.029 0.025 0.053 0.200 0.346 0.374 0.429 F2 0.000 0.090 0.135 0.375 0.615 0.660 0.750 F3 - 0.224 - 0.153 - 0.117 0.072 0.260 0.297 0.367 New/Additional Parameters COMP_REL 0.721 0.749 0.763 0.838 0.912 0.927 0.954 274 SAS Input Code for Nonlinearity SEM Reliability of Study Latent Variable Goal Alignment proc iml; RESET fuzz; THRESH={ - 2.037 - .610, - 1.732 - .674, - .674 0, - 2.307 - .610, - 2.037 - .431}; LOAD={.837,.975,.798,.895,.698}; FACCOR={1}; POLY={1 .812 .442 .601 .669,.81 2 1 .773 .821 .754, .442 .773 1 .812 .451, .601 .821 .812 1 .780, .669 .754 .451 .780 1}; NTHRESH=Ncol(thresh); NCAT=NTHRESH+1; NITEM=Nrow(LOAD); NFACT=Ncol(LOAD); POLYR=LOAD*FACCOR*T(LOAD); do j=1 to NITEM; POLYR[j,j]=1; end; DIFFPOLY=POLY - POLYR; Print NT HRESH[label="Number of Thresholds"], NITEM[label="Number of items"], NCAT[label="Number of response categories"], NFACT[label="Number of factors"], THRESH[label="Response Thresholds"],LOAD[label="Factor Loadings"], FACCOR[label="Factor Correlation Matrix"] , POLY[label="Polychoric Correlation Matrix among Continuous Items"] ; print "The matrix below is the difference between polychoric 275 correlation matrix generated by factors and inputted polychoric correlation matrix. Nonzero values should represent the esti mated correlated errors, as specified by the user, or an error in inputted data."; print DIFFPOLY[label=" "]; sumnum=0; addden=0; do j=1 to NITEM; do jp=1 to NITEM; sumprobn2=0; addprobn2=0; do c=1 to NTHRESH; do cp=1 to NTHRESH; sumrvstar=0; do k=1 to NFA CT; do kp=1 to NFACT; sumrvstar=sumrvstar+LOAD[j,k]*LOAD[jp,kp]*FACCOR[k,kp]; end; end; sumprobn2=sumprobn2+probbnrm(THRESH[j,c],THRESH[jp,cp],sumrvstar); addprobn2=addprobn2+probbnrm(THRESH[j,c],THRESH[jp,cp],POLY[j,jp]); end; end; sumprobn1=0; sumprobn1p =0; do cc=1 to NTHRESH; sumprobn1=sumprobn1+CDF('NORMAL',THRESH[j,cc]); 276 sumprobn1p=sumprobn1p+CDF('NORMAL',THRESH[jp,cc]); end; sumnum=sumnum+(sumprobn2 - sumprobn1*sumprobn1p); addden=addden+(addprobn2 - sumprobn1*sumprobn1p); end; end; reliab=sumnum/addden; print sumnum[label="Numerator of Eq. (21)"], addden[label="Denominator of Eq. (21)"], reliab[label="Nonlinear SEM Reliability Coefficient"]; quit; 277 SAS Nonlinear SEM Reliability Results for Study Latent Variable Goal Alignment 278 SAS Nonline ar SEM Reliability Results for Study Latent Variable Goal Alignment (C ) 279 SAS Input Code for Nonlinearity SEM Reliability of Study Sub - factor Coordination proc iml; RESET fuzz; THRESH={ - 2.037 - .610, - 1.732 - .674, - .674 0, - 2.307 - .610, - 2.037 - .431}; LOA D={.837,.975,.798,.895,.698}; FACCOR={1}; POLY={1 .812 .442 .601 .669,.812 1 .773 .821 .754, .442 .773 1 .812 .451, .601 .821 .812 1 .780, .669 .754 .451 .780 1}; NTHRESH=Ncol(thresh); NCAT=NTHRESH+1; NITEM=Nrow(LOAD); NFACT=Ncol(LOAD); POLYR=LOAD*FACCOR*T (LOAD); do j=1 to NITEM; POLYR[j,j]=1; end; DIFFPOLY=POLY - POLYR; Print NTHRESH[label="Number of Thresholds"], NITEM[label="Number of items"], NCAT[label="Number of response categories"], NFACT[label="Number of factors"], THRESH[label="Response Thresholds"] ,LOAD[label="Factor Loadings"], FACCOR[label="Factor Correlation Matrix"], POLY[label="Polychoric Correlation Matrix among Continuous Items"] ; print "The matrix below is the difference between polychoric correlation matrix generated by factors and inputte d polychoric correlation matrix. Nonzero values should represent the estimated 280 correlated errors, as specified by the user, or an error in inputted data."; print DIFFPOLY[label=" "]; sumnum=0; addden=0; do j=1 to NITEM; do jp=1 to NITEM; sumprobn2=0; addpr obn2=0; do c=1 to NTHRESH; do cp=1 to NTHRESH; sumrvstar=0; do k=1 to NFACT; do kp=1 to NFACT; sumrvstar=sumrvstar+LOAD[j,k]*LOAD[jp,kp]*FACCOR[k,kp]; end; end; sumprobn2=sumprobn2+probbnrm(THRESH[j,c],THRESH[jp,cp],sumrvstar); addprobn2=addprobn2+probbnrm (THRESH[j,c],THRESH[jp,cp],POLY[j,jp]); end; end; sumprobn1=0; sumprobn1p=0; do cc=1 to NTHRESH; sumprobn1=sumprobn1+CDF('NORMAL',THRESH[j,cc]); sumprobn1p=sumprobn1p+CDF('NORMAL',THRESH[jp,cc]); end; 281 sumnum=sumnum+(sumprobn2 - sumprobn1*sumprobn1p); addden= addden+(addprobn2 - sumprobn1*sumprobn1p); end; end; reliab=sumnum/addden; print sumnum[label="Numerator of Eq. (21)"], addden[label="Denominator of Eq. (21)"], reliab[label="Nonlinear SEM Reliability Coefficient"]; quit; 282 SAS Nonlinear SEM Reliability Resu lts for Study Sub - factor Coordination 283 SAS Input Code for Nonlinearity SEM Reliability of Study Latent Variable Credibility proc iml; RESET fuzz; THRESH={ - 2.037 - 1.534 - .431, - 1.732 - 1.383 - .105, - 2.037 - 1.383 0}; LOAD={.550,.985,.936}; FACCOR={1}; POLY={ 1 .542 .515,.542 1 .923,.515 .923 1}; NTHRESH=Ncol(thresh); NCAT=NTHRESH+1; NITEM=Nrow(LOAD); NFACT=Ncol(LOAD); POLYR=LOAD*FACCOR*T(LOAD); do j=1 to NITEM; POLYR[j,j]=1; end; DIFFPOLY=POLY - POLYR; Print NTHRESH[label="Number of Thresholds"], NITEM[label="Nu mber of items"], NCAT[label="Number of response categories"], NFACT[label="Number of factors"], THRESH[label="Response Thresholds"],LOAD[label="Factor Loadings"], FACCOR[label="Factor Correlation Matrix"], POLY[label="Polychoric Correlation Matrix among Co ntinuous Items"] ; print "The matrix below is the difference between polychoric correlation matrix generated by factors and inputted polychoric correlation matrix. Nonzero values should represent the estimated correlated errors, as specified by the user, o r an error in inputted 284 data."; print DIFFPOLY[label=" "]; sumnum=0; addden=0; do j=1 to NITEM; do jp=1 to NITEM; sumprobn2=0; addprobn2=0; do c=1 to NTHRESH; do cp=1 to NTHRESH; sumrvstar=0; do k=1 to NFACT; do kp=1 to NFACT; sumrvstar=sumrvstar+LOAD[j,k]* LOAD[jp,kp]*FACCOR[k,kp]; end; end; sumprobn2=sumprobn2+probbnrm(THRESH[j,c],THRESH[jp,cp],sumrvstar); addprobn2=addprobn2+probbnrm(THRESH[j,c],THRESH[jp,cp],POLY[j,jp]); end; end; sumprobn1=0; sumprobn1p=0; do cc=1 to NTHRESH; sumprobn1=sumprobn1+CDF('NOR MAL',THRESH[j,cc]); sumprobn1p=sumprobn1p+CDF('NORMAL',THRESH[jp,cc]); end; sumnum=sumnum+(sumprobn2 - sumprobn1*sumprobn1p); 285 addden=addden+(addprobn2 - sumprobn1*sumprobn1p); end; end; reliab=sumnum/addden; print sumnum[label="Numerator of Eq. (21)"], addden[ label="Denominator of Eq. (21)"], reliab[label="Nonlinear SEM Reliability Coefficient"]; quit; 286 SAS Nonlinear SEM Reliability Results for Study Sub - factor Credibility 287 SAS Input Code for Nonlinearity SEM Reliability of Study Latent Variable Specializati on proc iml; RESET fuzz; THRESH={ - 1.732 - .210, - 2.037 - .742, - 1.383 - .210}; LOAD={.427,.512,.779}; FACCOR={1}; POLY={1 .219 .333,.219 1 .399,.333 .399 1}; NTHRESH=Ncol(thresh); NCAT=NTHRESH+1; NITEM=Nrow(LOAD); NFACT=Ncol(LOAD); POLYR=LOAD*FACCOR*T(LOAD); do j=1 to NITEM; POLYR[j,j]=1; end; DIFFPOLY=POLY - POLYR; Print NTHRESH[label="Number of Thresholds"], NITEM[label="Number of items"], NCAT[label="Number of response categories"], NFACT[label="Number of factors"], THRESH[label="Response Thresholds"],LOAD[la bel="Factor Loadings"], FACCOR[label="Factor Correlation Matrix"], POLY[label="Polychoric Correlation Matrix among Continuous Items"] ; print "The matrix below is the difference between polychoric correlation matrix generated by factors and inputted polych oric correlation matrix. Nonzero values should represent the estimated correlated errors, as specified by the user, or an error in inputted 288 data."; print DIFFPOLY[label=" "]; sumnum=0; addden=0; do j=1 to NITEM; do jp=1 to NITEM; sumprobn2=0; addprobn2=0; do c=1 to NTHRESH; do cp=1 to NTHRESH; sumrvstar=0; do k=1 to NFACT; do kp=1 to NFACT; sumrvstar=sumrvstar+LOAD[j,k]*LOAD[jp,kp]*FACCOR[k,kp]; end; end; sumprobn2=sumprobn2+probbnrm(THRESH[j,c],THRESH[jp,cp],sumrvstar); addprobn2=addprobn2+probbnrm(THRESH[ j,c],THRESH[jp,cp],POLY[j,jp]); end; end; sumprobn1=0; sumprobn1p=0; do cc=1 to NTHRESH; sumprobn1=sumprobn1+CDF('NORMAL',THRESH[j,cc]); sumprobn1p=sumprobn1p+CDF('NORMAL',THRESH[jp,cc]); end; sumnum=sumnum+(sumprobn2 - sumprobn1*sumprobn1p); 289 addden=addden+( addprobn2 - sumprobn1*sumprobn1p); end; end; reliab=sumnum/addden; print sumnum[label="Numerator of Eq. (21)"], addden[label="Denominator of Eq. (21)"], reliab[label="Nonlinear SEM Reliability Coefficient"]; quit; 290 SAS Nonlinear SEM Reliability Results for Study Sub - factor Specialization 291 SAS Input Code for Nonlinearity SEM Reliability of Study Latent Variable Project Performance proc iml; RESET fuzz; THRESH={ - 1.732 - .674 - .264 - .157, - .887 - .742 0 0, - .264 0 .052 0, - 2.037 - .210 .052 0, - 2.037 - .549 - .105 . 052, - .967 - .887 0 0}; LOAD={.663,.718,.845,.858,.597,.610}; FACCOR={1}; POLY={1 .406 .310 .704 .230 .338,.406 1 .681 .565 .514 .416,.310 .681 1 .648 .575 .663,.704 .565 .648 1 .369 .393,.230 .514 .575 .369 1 .440,.338 .416 .663 .393 .440 1}; NTHRESH=Ncol (thresh); NCAT=NTHRESH+1; NITEM=Nrow(LOAD); NFACT=Ncol(LOAD); POLYR=LOAD*FACCOR*T(LOAD); do j=1 to NITEM; POLYR[j,j]=1; end; DIFFPOLY=POLY - POLYR; Print NTHRESH[label="Number of Thresholds"], NITEM[label="Number of items"], NCAT[label="Number of response ca tegories"], NFACT[label="Number of factors"], THRESH[label="Response Thresholds"],LOAD[label="Factor Loadings"], FACCOR[label="Factor Correlation Matrix"], POLY[label="Polychoric Correlation Matrix among Continuous Items"] ; print "The matrix below is the difference between polychoric 292 correlation matrix generated by factors and inputted polychoric correlation matrix. Nonzero values should represent the estimated correlated errors, as specified by the user, or an error in inputted data."; print DIFFPOLY[labe l=" "]; sumnum=0; addden=0; do j=1 to NITEM; do jp=1 to NITEM; sumprobn2=0; addprobn2=0; do c=1 to NTHRESH; do cp=1 to NTHRESH; sumrvstar=0; do k=1 to NFACT; do kp=1 to NFACT; sumrvstar=sumrvstar+LOAD[j,k]*LOAD[jp,kp]*FACCOR[k,kp]; end; end; sumprobn2=sump robn2+probbnrm(THRESH[j,c],THRESH[jp,cp],sumrvstar); addprobn2=addprobn2+probbnrm(THRESH[j,c],THRESH[jp,cp],POLY[j,jp]); end; end; sumprobn1=0; sumprobn1p=0; do cc=1 to NTHRESH; sumprobn1=sumprobn1+CDF('NORMAL',THRESH[j,cc]); 293 sumprobn1p=sumprobn1p+CDF('NOR MAL',THRESH[jp,cc]); end; sumnum=sumnum+(sumprobn2 - sumprobn1*sumprobn1p); addden=addden+(addprobn2 - sumprobn1*sumprobn1p); end; end; reliab=sumnum/addden; print sumnum[label="Numerator of Eq. (21)"], addden[label="Denominator of Eq. (21)"], reliab[label="No nlinear SEM Reliability Coefficient"]; quit; 294 SAS Nonlinear SEM Reliability Results for Study Sub - factor Project Performance 2 95 SAS Nonlinear SEM Reliability Results for Study Sub - 296 SAS Input Code for Nonlinearity SEM Reliability of Study Latent Variable Communication Performance proc iml; RESET fuzz; THRESH={ - 2.037 - .742 0, - 1.732 - .549 0, - 1.732 - 1.383 - .264}; LOAD={.852,.940,.698}; FACCOR={1}; POLY={1 .801 .595,.801 1 .656,.595 .656 1}; NTHRESH=Ncol(thresh); NCAT=NTHR ESH+1; NITEM=Nrow(LOAD); NFACT=Ncol(LOAD); POLYR=LOAD*FACCOR*T(LOAD); do j=1 to NITEM; POLYR[j,j]=1; end; DIFFPOLY=POLY - POLYR; Print NTHRESH[label="Number of Thresholds"], NITEM[label="Number of items"], NCAT[label="Number of response categories"], NFACT[l abel="Number of factors"], THRESH[label="Response Thresholds"],LOAD[label="Factor Loadings"], FACCOR[label="Factor Correlation Matrix"], POLY[label="Polychoric Correlation Matrix among Continuous Items"] ; print "The matrix below is the difference between polychoric correlation matrix generated by factors and inputted polychoric correlation matrix. Nonzero values should represent the estimated correlated errors, as specified by the user, or an error in inputted 297 data."; print DIFFPOLY[label=" "]; sumnum=0; a ddden=0; do j=1 to NITEM; do jp=1 to NITEM; sumprobn2=0; addprobn2=0; do c=1 to NTHRESH; do cp=1 to NTHRESH; sumrvstar=0; do k=1 to NFACT; do kp=1 to NFACT; sumrvstar=sumrvstar+LOAD[j,k]*LOAD[jp,kp]*FACCOR[k,kp]; end; end; sumprobn2=sumprobn2+probbnrm(THRE SH[j,c],THRESH[jp,cp],sumrvstar); addprobn2=addprobn2+probbnrm(THRESH[j,c],THRESH[jp,cp],POLY[j,jp]); end; end; sumprobn1=0; sumprobn1p=0; do cc=1 to NTHRESH; sumprobn1=sumprobn1+CDF('NORMAL',THRESH[j,cc]); sumprobn1p=sumprobn1p+CDF('NORMAL',THRESH[jp,cc]) ; end; sumnum=sumnum+(sumprobn2 - sumprobn1*sumprobn1p); 298 addden=addden+(addprobn2 - sumprobn1*sumprobn1p); end; end; reliab=sumnum/addden; print sumnum[label="Numerator of Eq. (21)"], addden[label="Denominator of Eq. (21)"], reliab[label="Nonlinear SEM Reliabi lity Coefficient"]; quit; 299 SAS Nonlinear SEM Reliability Results for Study Sub - factor Communication Performance 300 SAS Input Code for Nonlinearity SEM Reliability of Study Latent Variable Team Performance proc iml; RESET fuzz; THRESH={ - 1.534 - .742 .549 0, - 2.037 - 1.534 - 1.258 0, - 1.732 - 1.150 - .105 0, - 2.037 - 1.383 - .264 0, - 1.732 - 1.150 - .157 0, - .887 .105 0 0}; LOAD={.750,.812,.741,.869,.744,.634}; FACCOR={1}; POLY={1 .582 .533 .695 .411 .577,.582 1 .704 .659 .597 .460,.533 .704 1 .582 .549 .460,.695 .659 . 582 1 .705 .589,.411 .597 .549 .705 1 .422,.577 .460 .460 .589 .422 1}; NTHRESH=Ncol(thresh); NCAT=NTHRESH+1; NITEM=Nrow(LOAD); NFACT=Ncol(LOAD); POLYR=LOAD*FACCOR*T(LOAD); do j=1 to NITEM; POLYR[j,j]=1; end; DIFFPOLY=POLY - POLYR; Print NTHRESH[label="Numbe r of Thresholds"], NITEM[label="Number of items"], NCAT[label="Number of response categories"], NFACT[label="Number of factors"], THRESH[label="Response Thresholds"],LOAD[label="Factor Loadings"], FACCOR[label="Factor Correlation Matrix"], POLY[label="Poly choric Correlation Matrix among Continuous Items"] ; print "The matrix below is the difference between polychoric correlation matrix generated by factors and inputted polychoric correlation matrix. Nonzero values should represent the estimated 301 correlated e rrors, as specified by the user, or an error in inputted data."; print DIFFPOLY[label=" "]; sumnum=0; addden=0; do j=1 to NITEM; do jp=1 to NITEM; sumprobn2=0; addprobn2=0; do c=1 to NTHRESH; do cp=1 to NTHRESH; sumrvstar=0; do k=1 to NFACT; do kp=1 to NFA CT; sumrvstar=sumrvstar+LOAD[j,k]*LOAD[jp,kp]*FACCOR[k,kp]; end; end; sumprobn2=sumprobn2+probbnrm(THRESH[j,c],THRESH[jp,cp],sumrvstar); addprobn2=addprobn2+probbnrm(THRESH[j,c],THRESH[jp,cp],POLY[j,jp]); end; end; sumprobn1=0; sumprobn1p=0; do cc=1 to NTH RESH; sumprobn1=sumprobn1+CDF('NORMAL',THRESH[j,cc]); sumprobn1p=sumprobn1p+CDF('NORMAL',THRESH[jp,cc]); end; 302 sumnum=sumnum+(sumprobn2 - sumprobn1*sumprobn1p); addden=addden+(addprobn2 - sumprobn1*sumprobn1p); end; end; reliab=sumnum/addden; print sumnum[label ="Numerator of Eq. (21)"], addden[label="Denominator of Eq. (21)"], reliab[label="Nonlinear SEM Reliability Coefficient"]; quit; 303 SAS Nonlinear SEM Reliability Results for Study Sub - factor Team Performance 304 SAS Nonlinear SEM Reliability Results for Stud y Sub - 305 REFERENCES 306 REFERENCES AIA. (2007). . Integrated Project Delivery: A Guide . San Francisco. Retrieved from ht tp://www.cmhc.ca Anderson, L. L., & Polkinghorn, B. D. (2011). Efficacy of Partnering on the Woodrow Wilson Bridge Project: Empirical Evidence of Collaborative Problem - Solving Benefits. Journal of Legal Affairs and Dispute Resolution in Engineering and Co nstruction , 3 , 17 27. Argote, L. (2015). An Opportunity for Mutual Learning between Organizational Learning and Global Strategy Researchers: Transactive Memory Systems. Global Strategy Journal . doi:10.1002/gsj.1096 Argote, L., Ingram, P., Levine, J. M., & Moreland, R. L. (2000). Knowledge Transfer in Organizations: Learning from the Experience of Others. Organizational Behavior and Human Decision Processes , 82 , 1 8. Baiden, B. K., & Price, A. D. F. (2011). The effect of integration on project delivery te am effectiveness. International Journal of Project Management , 29 , 129 136. Baiden, B. K., Price, A. D. F., & Dainty, A. R. J. (2006). The extent of team integration within construction projects. International Journal of Project Management , 24 , 13 23. Ba iden, B., Price, A., & Dainty, A. (2003). Looking beyond processes: Human factors in team integration. In D. J. Greenwood, ed (Vol. 1, pp. 3 5). Bennett, J., & Peace, S. (2007). Partnering in the Construction Industry . Routledge. Retrieved from https://bo oks.google.com/books?id=FNn1woL42r8C&pgis=1 Betts, M., Fischer, M. a., & Koskela, L. (1995). The purpose and definition of integration. Integrated Construction Information , 3 18. Black, C., Akintoye, A., & Fitzgerald, E. (2000). Analysis of success facto rs and benefits of partnering in construction. International Journal of Project Management , 18 , 423 434. Bresnen, M. (2007). Deconstructing partnering in project - based organisation: Seven pillars, seven paradoxes and seven deadly sins. International Journ al of Project Management , 25 , 365 374. Bresnen, M. (2009). Living the dream? Understanding partnering as emergent practice. Construction Management and Economics , 27 , 923 933. 307 Bresnen, M., & Marshall, N. (2002). The engineering or evolution of co - operat ion? A tale of two partnering projects. International Journal of Project Management , 20 , 497 505. Cost Engineering , 43/No.4 , 32 37. Bureau of Economic Analysis , B. (2018). Gross Domestic Product by Industry: Fourth Quarter and Annual 2017 Real GDP and Real Value Added by Sector. Bygballe, L. E., Jahre, M., & Swärd, A. (2010). Partnering relationships in construction: A literature review. Journal of Purchasing a nd Supply Management , 16 , 239 253. Cacamis, M., & El Asmar, M. (2014). Improving project performance through partnering and emotional intelligence. Practice Periodical on Structural Design and Construction , 19 , 50 56. Campbell, D. T., & Fiske, D. W. (195 9). Convergent and Discriminant Validity by the Mutitrait - Multimethod Matrix. Psychological Bulletin , 56 , 81 105. Carson, S. J., Madhok, A., & Wu, T. (2006). UNCERTAINTY , OPPORTUNISM , AND ND RELATIONAL CONTRACTING. The Academy of Management Journal , 49 , 1058 1077. Chan, A. P. C., Chan, D. W. M., Chiang, Y. H., Tang, B. S., Chan, E. H. W., & Ho, K. S. K. (2004). Exploring Critical Success Factors for Partnering in Construction Projects. Jou rnal of Construction Engineering and Management , 130 , 188 198. Chan, A. P. C., Chan, D. W. M., & Ho, K. S. K. (2003). Partnering in Construction: Critical Study of Problems for Implementation. Journal of Management in Engineering , 19 , 126 135. Che Ibrahi m, C. K. I., Costello, S. B., & Wilkinson, S. (2015). A Fuzzy Approach to Developing Scales for Performance Levels of Alliance Team Integration Assessment. Journal of Construction Engineering and Management , 141 , 04014094. Chen, L., & Manley, K. (2014). V alidation of an Instrument to Measure Governance and Performance on Collaborative Infrastructure Projects. Journal of Construction Engineering and Management , 140 . doi:10.1061/(ASCE)CO.1943 - 7862.0000834. Cheng, E., & Li, H. (2001). Development of a concep tual model of construction partnering. Engineering, Construction and Architectural Management , 8 , 292 303. Cheng, E. W. L., & Li, H. (2002). Construction Partnering Process and Associated Critical Success Factors: Quantitative Investigation. Journal of Ma nagement in Engineering , 18 , 194 202. 308 Cheung, S. O., Ng, T. S. T., Wong, S. P., & Suen, H. C. H. (2003). Behavioral aspects in construction partnering. International Journal of Project Management , 21 , 333 343. Cheung, S. O., Suen, H. C. ., & Cheung, K. K . . (2003). An automated partnering monitoring system Partnering Temperature Index. Automation in Construction , 12 , 331 345. Chiocchio, F., & Essiembre, H. (2009). Cohesion and Performance . Small Group Research I , 40 , 382 - 420. Claro, D. P., Hagelaar, G., & Omta, O. (2003). The determinants of relational governance and performance: How to manage business relationships? Industrial Marketing Management , 32 , 703 716. Cohen, J., Cohen, P., West, S., & Aiken, L. (2013). Applied multiple regression/correlation analysis for the behavioral sciences. Routledge. Comu, S., Iorio, J., Taylor, J. E., & Dossick, C. S. (2013). Quantifying the Impact of Facilitation on Transactive Memory System Formation in Global Virtual Project Networks. Journal of Construction Enginee ring and Management , 139 , 294 303. Construction Industry Institute [CII]. (1989). Partnering: meeting the challenges of the future, Partnering Task Force Interim Report . Texas A&M University. Crespin - Mazet, F., Havenvid, M. I., & Linne, A. (2015). Antece dents of project partnering in the construction industry - The impact of relationship history. Industrial Marketing Management , 50 , 4 15. of Phenomenology. Philo sophy of the Social Sciences , 1 , 309 344. Davis, P., & Love, P. (2011). Alliance contracting: adding value through relationship development. Engineering, Construction and Architectural Management , 18 , 444 461. DeChurch, L. a., & Mesmer - Magnus, J. R. (201 0). Measuring shared team mental models: A meta - analysis. Group Dynamics: Theory, Research, and Practice , 14 , 1 14. DeShon, R. P., Kozlowski, S. W. J., Schmidt, A. M., Milner, K. R., & Wiechmann, D. (2004). A multiple - goal, multilevel model of feedback ef fects on the regulation of individual and team performance. The Journal of Applied Psychology , 89 , 1035 1056. Dewulf, G., & Kadefors, A. (2012). Collaboration in public construction contractual incentives, partnering schemes and trust. Engineering Project Organization Journal , 2 , 240 250 . 309 Diamantopoulos, A., & Winklhofer, H. M. (2001). Index construction with formative indicators: An alternative to scale development. Journal of Marketing Research , 38 , 269 277. Dietrich, P., Eskerod, P., Dalcher, D., & S andhawalia, B. (2010). The Dynamics of Collaboration in Multipartner Projects. Project Management Journal , 41 , 59 78. Considerations for the applied researcher. Practica l Assessment, Research & Evaluation , 14 , 1 11. Doloi, H. (2009). Relational partnerships: the importance of communication, trust and confidence and joint risk management in achieving project success. Construction Management and Economics , 27 , 1099 1109. Drexler Jr., J. A., & Larson, E. W. (2000). Partnering: Why Project Owner - Contractor Relationships Change. Engineering , 293 297. Earley, P. C. (1990). Impact of Process and Outcome Feedback on the Relation of Goal Setting to Task Performance. Academy of M anagement Journal , 33 , 87 105. AND CHALLENGES diverse. Academy of Management Journal , 50 , 25 32. Eriksson, P. E. (2010). Partnering: what is it, when should it be use d, and how should it be implemented? Construction Management and Economics , 28 , 905 917. Eriksson, P. E., Atkin, B., & Nilsson, T. (2009). Overcoming barriers to partnering through cooperative procurement procedures. Engineering, Construction and Architec tural Management , 16 , 598 611. Fellows, R., & Liu, A. M. M. (2012). Managing organizational interfaces in engineering construction projects: Addressing fragmentation and boundary issues across multiple interfaces. Construction Management and Economics , 30 , 653 671. Franz, B., Leicht, R., Molenaar, K., & Messner, J. (2016). Impact of Team Integration and Group Cohesion on Project Delivery Performance. Journal of Construction Engineering and Management , 04016088 - 6 , 1 12. Franz, B. W., & Leicht, R. M. (2016 ). An alternative classification of project delivery methods used in the United States building construction industry. Construction Management and Economics , 34 , 160 173. Grajek, K. M., Gibson Jr., G. E., & Tucker, R. L. (2000). PARTNERED PROJECT PERFORMA NCE IN TEXAS DEPARTMENT OF TRANSPORTATION. Journal of Infrastructure Systems , 6 , 73 79. 310 Granovetter, M. (1985). Economic Action and Social Structure: The Problem of Embeddedness. American Journal of Sociology , 91 , 481 510. Gransberg, D. D., Dillon, W. D., Reynolds, L., & Boyd, J. (1999). Quantitative analysis of partnered project performance. Journal of Construction Engineering and Management , 125 , 161 167. Gransberg, D. D., Reynolds, H. L., & Boyd, J. (1997). Quanitative Analysis of the TxDOT Partnering Plus Program . Gransberg, D. D., & Scheepbouwer, E. (2015). US Partnering Programs and International Partnering Contracts and Alliances: Comparative Analysis. Transportation Research Record: Journal of the Transportation Research Board , 73 77. Green, S. B ., & Yang, Y. (2009). Reliability of summed item scores using structural equation modeling: An alternative to coefficient alpha. Psychometrika , 74 , 155 167. Hair, J. F., Black, W. C., Babin, B. J., Anderson, R. E., & Tatham, R. L. (1998). Multivariate dat a analysis (Volume 5,). Upper Saddle River, NJ: Prentice Hall. Harkins, S. G. (1987). Social loafing and social facilitation. Journal of Experimental Social Psychology , 23 , 1 18. Hoegl, M., & Gemuenden, H. G. (2001). Teamwork Quality and the Success of I nnovative Projects. Organization Science , 12 , 435 449. Hollingshead, A. B. (1998a). Communication, Learning, and Retrieval in Transactive Memory Systems. Journal of Experimental Social Psychology , 34 , 423 442. Hollingshead, A. B. (1998b). Retrieval proce sses in transactive memory systems. Journal of Personality and Social Psychology , 74 , 659 671. Hong, Y., Chan, D. W. M., Chan, A. P. C., & Yeung, J. F. Y. (2012). Critical Analysis of Partnering Research Trend in Construction Journals. Journal of Manageme nt in Engineering , 28 , 82 95. H fit. Electronic Journal of Business Research Methods , 6 , 53 60. Hox, J. J. (2010). Multilevel Analysis Techniques and Applications Second Edition (Second). New: Routledge. Retrieved from http://www.researchmethodsarena.com/multilevel - analysis - 9781848728462 311 Hsu, J. S. - C., Shih, S. - P., Chiang, J. C., & Liu, J. Y. - C. (2012). The im pact of transactive memory International Journal of Project Management , 30 , 329 340. Hsu, J. S. C., Chang, J. Y. T., Klein, G., & Jiang, J. J. (2011). Exploring the impact of t eam mental models on information utilization and project performance in system development. International Journal of Project Management , 29 , 1 12. Hsu, J. S., Liang, T. P., Wu, S. P. J., Klein, G., & Jiang, J. J. (2011). Promoting the integration of users and developers to achieve a collective mind through the screening of information system projects. International Journal of Project Management , 29 , 514 524. Hughes, D., Williams, T., & Ren, Z. (2012). Differing perspectives on collaboration in constructio n. Construction Innovation , 12 , 355 368. Hughes, D., Williams, T., & Ren, Z. (2012). Is incentivisation significant in ensuring successful partnered projects? Engineering, Construction and Architectural Management , 19 , 306 319. Hunter, J. E., Gerbing, D. W., & Boster, F. J. (1982). Machiavellian beliefs and personality: Construct invalidity of the Machiavellianism dimension. Journal of Personality and Social Psychology , 43 , 1293 1305. Ilgen, D. R., & Moore, C. F. (1987). Types and Choices of Performance Feedback. Journal of Applied Psychology , 72 , 401 406. IPI. (2017a). IPI Horizontal Construction Project Partnering Scalability Matrix . Livermore, CA. Retrieved from https://partneringinstitute.org/owners - toolbox/ IPI. (2017b). IPI Vertical Construction P roject Partnering Scalability Matrix . Livermore, CA. Retrieved from https://partneringinstitute.org/owners - toolbox/ Jaafari, A., & Manivong, K. (1999). The need for life - cycle integration of project processes. Engineering Construction & Architectural Mana gement (Wiley - Blackwell) , 6 , 235 255. Jackson, D. L., Gillaspy, J. A., & Purc - Stephenson, R. (2009). Reporting Practices in Confirmatory Factor Analysis: An Overview and Some Recommendations. Psychological Methods , 14 , 6 23. Jacobsson, M., & Roth, P. (20 14). Towards a shift in mindset: partnering projects as engagement platforms. Construction Management and Economics , 32 , 419 432. James, L. R., Demaree, R. G., & Wolf, G. (1984). Estimating within - group interrater reliability with and without response bia s. Journal of Applied Psychology , 69 , 85 98. 312 Jap, S. D. (1999). Pie - expansion efforts: Collaboration processes in buyer - supplier relationships. Journal of Marketing Research , 36 , 461. Johnson, M. D., Hollenbeck, J. R., Scott DeRue, D., Barnes, C. M., & Ju ndt, D. (2013). Functional versus dysfunctional team change: Problem diagnosis and structural feedback for self - managed teams. Organizational Behavior and Human Decision Processes , 122 , 1 11. Katz, N., Lazer, D., Arrow, H., & Contractor, N. (2004). Networ k Theory and Small Groups. Small Group Research , 35 , 307 332. Kerlinger, F. N., & Lee, H. B. (1999). Chapter 17: Research Design: Purpose and Principles. In Foundations of behavioral research . Kluger, A. N., & DeNisi, A. (1996). The effects of feedback i nterventions on performance: A historical review, a meta - analysis, and a preliminary feedback intervention theory. Psychological Bulletin , 119 , 254 284. Korkmaz, S. (2007). Piloting Evaluation Metrics For High Performance Green Building Project Delivery . The Pennsylvania State University. Kotlarsky, J., van den Hooff, B., & Houtman, L. (2015). Are We on the Same Page? Knowledge Boundaries and Transactive Memory System Development in Cross - Functional Teams. Communication Research , 42 , 319 344. Kozlowski, S. W. J., & Ilgen, D. R. (2006). Enhancing the effectiveness of work groups and teams. Psychological Science , 7 , 77 124. Kumaraswamy, M. M., Asce, M., Rahman, ; M Motiar, Yean, F., Ling, Y., & Phng, S. T. (2005). Reconstructing Cultures for Relational Co ntracting. Journal of Construction Engineering and Management , 131 , 1065 1075. Lahdenperä, P. (2012). Making sense of the multi - party contractual arrangements of project partnering, project alliancing and integrated project delivery. Construction Manageme nt and Economics , 30 , 57 79. Lam, C. (2015). The Role of Communication and Cohesion in Reducing Social Loafing in Group Projects. Business and Professional Communication Quarterly , 78 , 454 475. Larson, E. (1997). Partnering on construction projects: A st udy of the relationship between partnering activities and project success. IEEE Transactions on Engineering Management , 44 , 188 195. 313 Le - Hoai, L., Lee, Y. D., & Son, J. J. (2010). Partnering in Construction - Investigation of Problematic Issues for Imple mentation in Vietnam. KSCE Journal of Civil Engineering , 14 , 731 741. Lewis, K. (2003). Measuring transactive memory systems in the field: Scale development and validation. Journal of Applied Psychology , 88 , 587 604. Lewis, K., & Herndon, B. (2011). Tran sactive Memory Systems: Current Issues and Future Research Directions. Organization Science , 22 , 1254 1265. Li, H., Arditi, D., & Wang, Z. (2013). Factors That Affect Transaction Costs in Construction Projects. Journal of Construction Engineering & Manage ment , 139 , 60 68. Lu, P., Guo, S., Qian, L., He, P., & Xu, X. (2015). The effectiveness of contractual and relational governances in construction projects in China. International Journal of Project Management , 33 , 212 222. Manley, K., & Chen, L. (2015). Collaborative learning model of infrastructure construction: a capability perspective. Construction Innovation , 15 , 355 377. Manley, K., Mcfallan, S., & Kajewski, S. (2009). Relationship between construction firm strategies and innovation outcomes. Journa l of Construction Engineering and Management , 135 , 764 771. Mathieu, J., Maynard, M. T., Rapp, T., & Gilson, L. (2008). Team Effectiveness 1997 - 2007: A Review of Recent Advancements and a Glimpse Into the Future. Journal of Management , 34 , 410 476. Matur ana, S., Alarcón, L. F., Gazmuri, P., & Vrsalovic, M. (2007). On - Site Subcontractor Evaluation Method Based on Lean Principles and Partnering Practices. Journal of Management in Engineering , 23 , 67 74. McKim, R. A. (1992). Risk Management - Back to Basics . Cost Engineering , 34 , 7 12. Meng, X. (2012). The effect of relationship management on project performance in construction. International Journal of Project Management , 30 , 188 198. (2011). Advancing Research in Organizational Communication Through Quantitative Methodology. Management Communication Quarterly , 25 , 4 58. Mohammed, S., & Dum ville, B. C. (2001). Team mental models in a team knowledge framework: Expanding theory and measurement across disciplinary boundaries. Journal of Organizational Behavior , 22 , 89 106. 314 Mohammed, S., Ferzandi, L., & Hamilton, K. (2010). Metaphor no more: A 1 5 - year review of the team mental model construct. Journal of Management , 36 , 876 910. Mollaoglu - Korkmaz, S., Swarup, L., & Riley, D. (2013). Delivering Sustainable, High Performance Buildings: Influence of Project Delivery Methods on Integration and Proje ct Outcomes. Journal of Management in Engineering , 29 , 71 78. Mollaoglu, S., & Sparkling, A. (2015). A Meta - Analytic Synthesis of Partnering Literature in the Architecture, Engineering, and Construction Industry . Livermore, CA. Mollaoglu, S., Sparkling, A., & Thomas, S. (2015). An Inquiry to Move an Underutilized Best Practice Forward: Barriers to Partnering in the Architecture, Engineering, and Construction Industry. Project Management Journal , 46 , 69 83. Muthén, L. K., & Muthén, B. O. (2017). ide Manual (Eighth). Los Angeles, CA: Muthén & Muthén. Nadler, D. a. (1979). The effects of feedback on task group behavior: A review of the experimental research. Organizational Behavior and Human Performance , 23 , 309 338. Nadler, D. a., & Tushman, M. L. (1980). A model for diagnosing organizational behavior. Organizational Dynamics , 9 , 35 51. Nadler, D., Mirvis, P., & Cammann, C. (1976). The ongoing feedback system. Organizational Dynamics , 4 , 63 80. Naoum, S. (2003). An overview into the concept of partnering. International Journal of Project Management , 21 , 71 76. Nezlek, J. B. (2008). An Introduction to Multilevel Modeling for Social and Personality Psychology. Social and Personality Psychology Compass , 2 , 842 860. Ng, S. T., Rose, T. M., Mak, M. , & Chen, S. E. (2002). Problematic issues associated with project partnering - the contractor perspective. International Journal of Project Management , 20 , 437 449. Nunnally, J. C. (1978). Chapter 7: Assessment of reliability. Psychometric Theory . Ospin a - Alvarado, A., Castro - Lacouture, D., & Roberts, J. S. (2016). Unified Framework for Construction Project Integration. Journal of Construction Engineering and Management , 142 . Ouchi, W. G. (1980). Markets, Bureaucracies, and Clans. Administrative Science Quarterly , 25 , 129 142. 315 Park, G., Spitzmuller, M., & DeShon, R. P. (2013). Advancing Our Understanding of Team Motivation: Integrating Conceptual Approaches and Content Areas. Journal of Management , 39 , 1339 1379. Pishdad - bozorgi, P., & Beliveau, Y. J. ( 2016). Symbiotic Relationships between Integrated Project Delivery ( IPD ) and Trust. International Journal of Construction Education and Research , 12 , 179 192. Pishdad - Bozorgi, P., & Beliveau, Y. J. (2016). A Schema of Trust Building Attributes and Their Corresponding Integrated Project Delivery Traits. International Journal of Construction Education and Research , 12 , 142 160. Poppo, L., & Zenger, T. (2002). Do formal contracts and relational governance function as substitutes or complements? Strategic M anagement Journal , 23 , 707 725. Puddicombe, M. S. (1997). Designers and Contractors: Impediments to Integration. Journal of Construction Engineering and Management , 123 , 245 252. Rahman, M. M., & Kumaraswamy, M. M. (2002). Joint risk management through t ransactionally efficient relational contracting. Construction Management and Economics , 20 , 45 54. Rahman, M. M., & Kumaraswamy, M. M. (2004). Contracting relationship trends and transitions. Journal of Management in Engineering , 20 , 147 161. Rahman, M. M., & Kumaraswamy, M. M. (2005). Relational Selection for Collaborative Working Arrangements. Journal of Construction Engineering and Management , 131 , 1087 1099. Robinson, W. S. (1950). Ecological Correlations and the Behavior of Individuals. American Soc iological Review , 15 , 337 341. Roehrich, J. K., & Lewis, M. A. (2010). Towards a model of governance in complex ( product service ) inter - organizational systems. Construction Management and Economics , 28 , 1155 1164. Salas, E., Cooke, N. J., & Rosen, M. a. (2008). On teams, teamwork, and team performance: discoveries and developments. Human Factors , 50 , 540 7. Salas, E., Dickinson, T. L., Converse, S. A., & Tannenbaum, S. I. (1992). Toward an understanding of team performance and training. Salas, E., G rossman, R., Hughes, A. M., & Coultas, C. W. (2015). Measuring Team Cohesion Observations from the Science. Human Factors: The Journal of the Human Factors and Ergonomics Society , 57 , 365 374. 316 Shaffer, J. A., DeGeest, D., & Li, A. (2016). Tackling the Pro blem of Construct Proliferation: A Guide to Assessing the Discriminant Validity of Conceptually Related Constructs. Organizational Research Methods , 19 , 80 110. Smith, R., Mossman, A., & Emmitt, S. (2011). Editorial: Lean and Integrated Project Delivery. Lean Construction Journal , 1 16. Snijders, T., & Bosker. (2012). Multilevel Analysis: An Introduction to Basic and Advanced Multilevel Modeling . Thousand Oaks. Retrieved from http://www.amazon.com/Multilevel - Analysis - Introduction - Advanced - Modeling - ebook/d p/B00HWAQHAI Sohani, S. (2016). In - Depth Case Study of a Partnered Project Delivery . Michigan State University. Sparkling, A. (2014). A Research Synthesis of Key Partnering Drivers and Performance Outcomes in Architecture, Engineering, and Construction R esearch . Michigan State University. Sparkling, A. E., Mollaoglu, S., & Kirca, A. (2016). Research Synthesis Connecting Trends in Architecture, Engineering, and Construction Project Partnering. Journal of Management in Engineering , 04016033 - 1 , 1 12. Step hen, A. T., & Coote, L. V. (2007). Interfirm behavior and goal alignment in relational exchanges. Journal of Business Research , 60 , 285 295. Sundaramurthy, C., & Lewis, M. (2003). Control and collaboration: Paradoxes of governance. Academy of Management R eview , 28 , 397 415. Suprapto, M., Bakker, H. L. M., & Mooi, H. G. (2015). Relational factors in owner contractor collaboration: The mediating role of teamworking. International Journal of Project Management , 33 , 1347 1363. Suprapto, M., Bakker, H. L. M., Mooi, H. G., & Moree, W. (2014). Sorting out the essence of owner - contractor collaboration in capital project delivery. International Journal of Project Management , 33 , 664 683. Teicholz, P., & Fischer, M. (1994). Strategy for Computer Integrated Constru ction Technology. Journal of Construction Engineering and Management , 120 , 117 131. Touran, A., Gransberg, D. D., Molenaar, K. R., & Ghavamifar, K. (2011). Selection of project delivery method in transit: Drivers and objectives. Journal of Management in E ngineering , 27 , 21 27. Tuuli, M. M., & Rowlinson, S. (2009). Performance Consequences of Psychological Empowerment. Journal of Construction Engineering and Ma nagement , 135 , 1334 1347. 317 Wegner, D. M. (1987). Transactive Memory: A Contemporary Analysis of th e Group Mind. In Theories of Group Behavior (pp. 185 208). New York: Springer. Wegner, D. M., Erber, R., & Raymond, P. (1991). Transactive memory in close relationships. Journal of Personality and Social Psychology , 61 , 923 929. Williamson, O. E. (1979). Transaction - cost economics: the governance of contractual relations. Journal of Law and Economics , 22 , 233 261. Xue, X., Shen, Q., & Ren, Z. (2010). Critical Review of Collaborative Working in Construction Projects: Business environment and human behavio urs. Journal of Management in Engineering , 26 , 196 208. Yeung, J. F. Y., Chan, A. P. C., & Chan, D. W. M. (2012). Defining relational contracting from the Wittgenstein family - resemblance philosophy. International Journal of Project Management , 30 , 225 239 . Yeung, J. F. Y., Chan, A. P. C., Chan, D. W. M., & Li, L. K . (2007). Development of a Partnering Construction Management and Economics , 25 , 1219 1237. Yin, R. K. (2003). Case study research: Design and methods (Third). Thousand Oaks: Sage Publications, Inc. Zaheer, A., McEvily, B., & Perrone, V. (1998). Does Trust Matter? Exploring the Effects of Interorganizational and Interpersonal Trust on Performance. Organization Science , 9 , 141 159. Zaheer, A., & Venkatraman, N. (1995). Relati onal Governance as an Interorganizational . Strategic Management Journal , 16 , 373 392. Zhang, L., Cheng, J., & Fan, W. (2015). Party Selection for Integrated Project Delivery Based on I nterorganizational Transactive Memory System. Journal of Construction Engineering and Management , 04015089. Zuo, J., Chan, A. P. C., Zhao, Z., Zillante, G., & Xia, B. (2013). Supporting and impeding factors for partnering in construction: a China study. F acilities , 31 , 468 488.