EMBEDDEDNESS OF FOREIGN-BORN FACULTY AT COMPREHENSI VE UNIVERSITIES IN THE UNITED STATES By Alexander Nickolas Akulli A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Higher, Adult, and Lifelong Education Œ Doctor of P hilosophy 2015 ABSTRACT EMBEDDEDNESS OF FOREIGN-BORN FACULTY AT COMPREHENSI VE UNIVERSITIES IN THE UNITED STATES By Alexander Nickolas Akulli Globalization and knowledge-driven economies have c reated a transnational landscape of career opportunities for scholars in the Science, T echnology, Engineering, and Mathematics (STEM) disciplines. Attracting and retaining highl y skilled individuals has been an important discussion in academic literature. For several dec ades, foreign-born faculty members have pursued their careers in the American professoriate and make up a significant percentage of STEM disciplines. As competition for talent incre ases across borders, little is understood about foreign-born faculty members in the American profes soriate. The present research investigated the embeddedness of university professors who are foreign-born and employed in the Science, Technolog y, Engineering, and Mathematics (STEM) disciplines at comprehensive universities in the Un ited States. Embeddedness is a positive construct that focuses on why people stay in their jobs. This study utilized online survey methodology to collect data from individuals who we re clustered at randomly selected comprehensive universities. The institution served as a means to find eligible participants; no data was collected on the institution or the local community. The analysis of data collected included calculating an embeddedness index for the study sample; analysis of variance (ANOVA) and t-tests we re used to help explain the variance in embeddedness of foreign-born faculty members includ ed in the present study. Descriptive analysis was used to help understand what describes the foreign-born STEM faculty members in this study. The data from the present study show that foreign-b orn faculty members are highly embedded. Embeddedness in the university was signi ficantly higher than embeddedness in the community. Factors such as gender, homeownership, and geographical region of origin were found influential in the embeddedness of foreign-bo rn faculty members who participated in the present study. In addition, for nearly a quarter o f the sample in the present study, identifying as foreign-born was not synonymous with identifying as an international faculty member. Findings from the present research can inform polic y and practices related to recruitment, socialization and retention of foreign-born faculty members in the American professoriate. University administrators and policy makers should consider these findings and explore ways to increase the embeddedness of foreign-born faculty m embers in the community. Recommendations for future research, policy, and pr actice are included. iv To Mimoza in honor of her unrealized dream to pursu e higher education. Your memory lives on. To Emma and Tessa, may the completion of this disse rtation and PhD degree inspire you to dream bigger and achieve greater successes in your bright future. v ACKNOWLEDGEMENTS This dissertation marks the completion of my PhD jo urney that would not have been possible without the inspiration, support, and enco uragement of so many people. I am forever grateful to my wife, Amy, who made endless sacrific es, taking on the challenges of raising our girls and keeping our household functional so that I could pursue, persevere, and achieve this lifelong dream. She deserves an honorary doctorate . And to my girls, Emma and Tessa, please forgive me for the extensive absences; your smiles and loving hugs made is all better. I owe my greatest gratitude to my academic adviser and dissertation chair, Roger Baldwin, who believed in me even before I was admit ted into the PhD program, challenged my assumptions, shaped this research, reviewed numerou s drafts and answered countless emails and phone calls. Thank you. I also extend my gratitud e to Amita Chudgar who provided guidance with the research methods and data analyses. Her s upport and ability to explain quantitative methods in applied research brought this study to f ruition. I am grateful to Ann Austin, Marilyn Amy, and John Dirkx for their constant support throughout the PhD journey and the completion of th is research. Their genuine care and support, insightful questions, and invitation to engage in c ollaborative scholarship shaped my identity as an emerging foreign-born scholar. My personal, pro fessional, and scholarly work has benefited tremendously from the collaborative work with Ann i n South Africa and John in Vietnam. Special thanks to my classmates, Pamela Roy and Jul ie Sinclair, for their genuine friendship and support. Their endless encouragemen t helped me persevere through the demands of the doctoral program and a full-time job. I als o extend my gratitude to Erik Kyper for his technical assistance with SPSS and Statistica softw are. Thank you. vi TABLE OF CONTENTS LIST OF TABLES .................................... ................................................... .................................. ix LIST OF FIGURES ................................... ................................................... ................................. xi CHAPTER 1: INTRODUCTION ........................... ................................................... ..................... 1 Statement of the Problem .......................... ................................................... ................................... 3 Definitions........................................ ................................................... ............................................ 8 Embeddedness....................................... ................................................... ............... 8 Foreign-born faculty. ............................. ................................................... .............. 9 Personal Attributes. .............................. ................................................... .............. 10 STEM Disciplines. ................................. ................................................... ............ 10 Theoretical Framework ............................. ................................................... ................................. 10 Research Questions and Hypothesis ................. ................................................... ......................... 12 Purpose Statement ................................. ................................................... ..................................... 13 Significance of the Study ......................... ................................................... .................................. 14 Dissertation Structure............................. ................................................... .................................... 16 CHAPTER 2: REVIEW OF LITERATURE ................... ................................................... .......... 17 Overview of International Faculty ................. ................................................... ............................ 17 Employee Mobility ................................. ................................................... ................................... 19 Job Satisfaction .................................. ................................................... ........................................ 21 Gender and Turnover ............................... ................................................... .................................. 23 Intent to Leave ................................... ................................................... ........................................ 23 Embeddedness Theoretical Framework ................ ................................................... ..................... 24 Links to the Organization and Community. .......... ................................................ 25 Fit in Organization and Community. ................ ................................................... . 26 Organization and Community Sacrifice............... ................................................. 2 7 Summary of Embeddedness Research .................. ................................................... ..................... 28 Conclusion ........................................ ................................................... ......................................... 30 CHAPTER 3: RESEARCH METHODS ....................... ................................................... ............ 31 Overview of Research Methodology .................. ................................................... ....................... 31 Research Paradigm.................................. ................................................... ................................... 31 Research Design.................................... ................................................... ..................................... 31 Site Selection and Population of Interest ......... ................................................... .......................... 33 Participant Selection and Recruitment ............. ................................................... .......................... 34 Sample............................................. ................................................... ........................................... 35 Data Collection ................................... ................................................... ....................................... 37 Protecting Human Subjects. ........................ ................................................... ....... 38 Instrumentation. .................................. ................................................... ............... 38 Validity. ......................................... ................................................... .................... 40 Reliability. ...................................... ................................................... .................... 41 Scales and Measures ............................... ................................................... ................................... 42 vii Research Questions, Hypotheses, and Statistical Tes ts ................................................ ................ 43 Methods of Analysis ............................... ................................................... ................................... 45 Descriptive Statistics. ........................... ................................................... .............. 46 T-Tests. .......................................... ................................................... .................... 46 Analyses of Variance (ANOVA). ..................... ................................................... . 47 Pearson Correlation Coefficient. .................. ................................................... ...... 47 Procedures and Statistical Tests. ................. ................................................... ....... 48 Descriptive Statistics. ........................... ................................................... .............. 49 Inferential Statistics. ........................... ................................................... ............... 50 Limitations ....................................... ................................................... .......................................... 54 Participant Selection. ............................ ................................................... ............. 54 Sample Size and Response Rate. .................... ................................................... ... 55 Instrument Design. ................................ ................................................... ............. 55 CHAPTER 4: DATA ANALYSIS AND FINDINGS ............. ................................................... .. 57 Data Overview ..................................... ................................................... ...................................... 58 Results –––––––––––––––––––––––––––––––––––59 Descriptive Analysis. ............................. ................................................... ............ 61 Identify as International. ........................ ................................................... ............ 61 Gender Representation. ............................ ................................................... .......... 62 Family Profiles. .................................. ................................................... ................ 64 Immigration Status Profile. ....................... ................................................... ......... 65 Geographical region of origin. .................... ................................................... ....... 65 Profiles by Rank. ................................. ................................................... ............... 66 Discipline Profiles. .............................. ................................................... ............... 67 Length of Employment. ............................. ................................................... ........ 68 Inferential Analysis .............................. ................................................... ...................................... 72 Embeddedness of Foreign-Born Faculty. ............. ................................................ 73 University and Community Embeddedness. ............ ............................................. 74 Subgroup Differences in Embeddedness. ............. ................................................ 75 Embeddedness by Gender. ........................... ................................................... ...... 75 Embeddedness by Family Status. .................... ................................................... .. 77 Embeddedness by Marital Status. ................... ................................................... ... 77 Embeddedness by School-age Children................ ................................................ 78 Embeddedness by Working Spouse/Partner. ........... ............................................. 79 Embeddedness by Homeownership. .................... ................................................. 8 1 Embeddedness by Immigration Status. ............... .................................................. 81 Embeddedness by Professional Status and Rank. ..... ............................................ 84 Discipline Differences in Embeddedness. ........... ................................................. 8 5 Embeddedness by Length of Employment. ............. ............................................. 86 Summary ........................................... ................................................... ......................................... 88 CHAPTER 5: SUMMARY, CONCLUSIONS, AND DISCUSSION ... ...................................... 92 Impetus for Current Study ......................... ................................................... ................................ 92 Summary of Embeddedness Theoretical Framework ..... ................................................... ........... 93 Summary of Findings ............................... ................................................... .................................. 95 viii Profile of Foreign-Born Faculty.................... ................................................... ............................. 97 Embeddedness of Foreign-Born Faculty .............. ................................................... ................... 103 University and Community Embeddedness ............. ................................................... ................ 105 Relationship of Demographic Factors to Embeddedness .................................................. ......... 109 Relationship of Gender on Embeddedness. ........... ............................................. 110 Homeownership Effect on Embeddedness. ............. ........................................... 112 Origin. ........................................... ................................................... ................... 114 Implications for Theory and Research .............. ................................................... ....................... 116 Implications for Policy and Practice .............. ................................................... .......................... 119 Conclusion ........................................ ................................................... ....................................... 123 APPENDIX .......................................... ................................................... .................................... 126 REFERENCES ........................................ ................................................... ................................ 129 ix LIST OF TABLES Table 1.1 Science and Engineering Doctorates Award ed by US Universities, by Citizenship Status................................. ................................................... ............................ 2 Table 1.2 Percentage of Foreign Nationals Receivin g Science and Engineering Doctorates in 1999 Who Were in the US, 2000-2009 .. ................................................... ... 4 Table 1.3 Percentage of Foreign Students Receiving Science and Engineering Doctorates in 2005 who were in the US, 2006-2007 .. ................................................... ..... 6 Table 2.1 Percentages of international scholars in the United States by country of origin ......... 1 8 Table 3.1 Embeddedness: global-item scale ........ ................................................... ..................... 38 Table 3.2 Variables of Interest .................. ................................................... ................................ 42 Table 3.3 Dependent Variable and sample means ... ................................................... ................ 49 Table 4.1 Scale items measuring embeddedness const ruct .............................................. ........... 59 Table 4.2 Correlations ........................... ................................................... .................................... 60 Table 4.3 Means and Standard Deviation ........... ................................................... ...................... 61 Table 4.4 Gender representation by length of emplo yment ............................................. ............ 63 Table 4.5 Gender representation by discipline .... ................................................... ..................... 64 Table 4.6 Length of employment by having school-ag e children ........................................ ....... 71 Table 4.7 Years of living in the community by havi ng school-age children ............................ ... 72 Table 4.8 T-Tests results comparing male and femal e participants on embeddedness variables ......................................... ................................................... ................................ 76 Table 4.9 T-Tests results comparing married and si ngle participants on embeddedness variables ................................................... ................................................... ..................................... 78 Table 4.10 T-Tests results comparing participants™ embeddedness by having/not having school-age children ............................... ................................................... ......................... 79 Table 4.11 T-Tests results comparing participants™ embeddedness by spouse/domestic partner working outside the home................... ................................................... ............... 80 Table 4.12 Results of T-Tests comparing participan ts™ embeddedness by homeownership ....... 81 Table 4.13 Means and significance of difference fo r immigration status on embeddedness ...... 83 x Table 4.14 Means and significance of difference fo r origin on embeddedness .......................... . 83 Table 4.15 Means and significance of difference fo r rank status on embeddedness ................... 8 4 Table 4.16 Means and significance of difference fo r discipline on embeddedness ..................... 85 Table 4.17 Means and significance of difference fo r length of employment on embeddedness ................................................... ................................................... ................................. –86 Table 4.18 T-Test results comparing embeddedness b y identify as international faculty member ................................................... ................................................... ..................................... 88 Table 4.19 Summary of factors significance on embe ddedness .......................................... ........ 88 Table 4.20 Summary of hypothesis and findings..... ................................................... ................. 90 xi LIST OF FIGURES Figure 1.1 Percentage of Foreign Nationals Receivi ng Science and Engineering Doctorates Who Were in the US Five to 10 years After Receipt of Do ctorate ........................................... .. 5 Figure 1.2 Embeddedness path diagram ............. ................................................... ...................... 11 Figure 4.1 Identify as International Faculty ..... ................................................... ......................... 62 Figure 4.2 Representation by immigration status .. ................................................... ................... 65 Figure 4.3 Representation by geographical region o f origin .......................................... ............. 66 Figure 4.4 Rank distribution ..................... ................................................... ................................ 67 Figure 4.5 Discipline representation ............. ................................................... ............................ 68 Figure 4.6 Gender across discipline .............. ................................................... ............................ 68 Figure 4.7 Years of employment by rank ........... ................................................... ...................... 70 Figure 4.8 Years of employment by gender.......... ................................................... .................... 71 Figure 4.9 Distribution of embeddedness data ..... ................................................... .................... 73 Figure 4.10 Distribution of university embeddednes s data ............................................ ............. 74 Figure 4.11 Distribution of community embeddedness data ............................................. .......... 74 Figure 4.12 Embeddedness across gender ........... ................................................... ..................... 76 Figure 4.13 Effect of working spouse/partner on em beddedness ........................................ ........ 80 Figure 4.14 Effect of immigration status on embedd edness ............................................ ............ 82 Figure 4.15 Effect of length of employmnet on embe ddedness .......................................... ........ 87 1 CHAPTER 1: INTRODUCTION For several decades, employee retention has been th e focus of human resource managers and top level administrators because organizational success depends on retaining highly talented workers (Ehrenberg, Kasper, & Rees, 1991; Mitchell et al., 2001; Zhou & Volkwein, 2004). Retention is especially important in academe where fi[t]he excellence of higher education is a function of the kind of people it is able to enlist and retain on its facultiesfl (Bowen & Shuster, 1986, p. 3). Put differently, the fiquality and ach ievements of [faculty]–determine the quality of the institutionfl (Duderstadt, 2000, p. 54). Facult y turnover presents several significant negative impacts at the university, including, but not limit ed to, disruption of course offerings and student learning, fragmented integration within and across disciplines, increased recruitment and start-up costs, and weakened morale within the institution ( Mobley 1982; Price 1997). Historically, American higher education has attract ed growing numbers of foreign-born scholars as an appealing destination characterized by academic freedom, state-of-the-art research facilities, opportunities for professional flexibil ity and career advancement within the professoriate (Schuster and Finkelstein, 2006; Thel in 2004). The recent growth in foreign-born faculty employment at American universities can be attributed to at least three factors, (1) the graying of the American professoriate (Doyle, 208), (2) the exponential growth in awarding graduate degrees in Science, Technology, Engineerin g, and Mathematics (STEM) disciplines to foreign-born students (Finn, 2012), and (3) the ste ady, but slow, growth in awarding such degrees to American-born students (Wasem, 2012; Con nolly, 2010). Table 1.1 captures the shift in science and engineering doctorate degrees awarde d over time to US-born and foreign-born students. 2 Table 1.1 Science and Engineering Doctorates Award ed by US Universities, by Citizenship Status 1 Furthermore, Doyle (2008) observed that the number of faculty, ages 55-64 and 71 or above, grew by 2 percent, respectively, from 1998 t o 2004. As larger numbers of faculty in the United States are approaching retirement (Schuster and Finkelstein, 2006; Thelin 2004), and a limited number of STEM degrees are being awarded to American-born students compared to international students (Kuenzi, Matthews, and Magna n, 2006), more than 30 percent of new faculty hires are now foreign-born (Kim, Wolf-Wende ll, & Towmbly, 2011c; Skachkova, 2007). This trend suggests that STEM disciplines depend si gnificantly on foreign-born scholars (Nelson & Rogers, 2005; NSB, 2003). Yet, the embeddedness of foreign-born faculty in the American professoriate remains unexamined. This study expands our understanding of foreign-bor n faculty members by examining their embeddedness in the American professoriate. I focus on the embeddedness construct because it helps us better understand why people st ay in their jobs and helps to predict employee attrition (Mitchell, Holtom, & Lee, 2001). Embedde dness informs employee retention practices, which is very important at a time when significant changes are occurring in the labor and employment markets and increasing globalization. O ne of the derivatives of globalization is the expanding landscape of career opportunities across borders and the growing competition for foreign talent. In a transnational landscape of jo b opportunities, embeddedness helps us 1 Data from Finn (2012) Table 1 3 understand if, and why foreign-born STEM scholars m ight stay at their current university rather than leave for better career opportunities elsewher e. Until this study, the higher education literature has focused mainly on why foreign-born f aculty members have pursued their careers in the American professoriate, but studies have not ex amined why these faculty members might opt to stay or leave. Statement of the Problem To a large extent, the growing numbers of foreign-b orn faculty in American academe are a positive trend, especially when considering that foreign-born faculty are highly productive and that they enrich the academic profession with their diverse perspectives. However, in a transnational landscape of career opportunities, ex panding globalization and internationalization, scholars in STEM disciplines have a wider array of employment options. Scholarship has demonstrated that institutions and governments acro ss the world are promoting greater internationalization and that the global competitio n for international talent is intensifying (Douglas & Edelstein, 2009; Hawthorne, 2005; Hercog , 2008; The Economist, 2006; Batalova, 2007; Iredale, 2000). Moreover, in recent years th e departure of foreign-born scholars from the United States has been on the rise (Finn, 2010; Kim et al., 2011a). Generally, universities recruit and employ foreign-born scholars from graduates who completed their doctorate degrees in the United States; only a very small percentage of such hires originate directly from overseas. Table 1.2 below, reports data from Finn™s (2012) st udy and demonstrates the growing rate of departure among foreign-born scholars who e arned their doctorate degrees in STEM fields from universities in the United States (the table includes students on temporary and permanent visas). 4 Table 1.2 Percentage of Foreign Nationals Receiving Science and Engineering Doctorates in 1999 Who Were in the US, 2000-2009 Table 1.2 above indicates that fewer foreign-born s cholars are staying in the United States. This trend is observed longitudinally (Fin n, 2012), and can be problematic for American universities because an organization™s competitive advantage depends, to a large extent, on the human capital it attracts and retains (Pfeffer, 199 4). Figure 1.1 captures the increased rate of departure of foreign-born scholars from the United States over a five to ten year period following their receipt of the doctorate degree from American universities. 5 Figure 1.1 Percentage of Foreign Nationals Receivi ng Science and Engineering Doctorates Who Were in the US Five to 10 years After Receipt of Do ctorate (Doctorates Awarded in 1991, 1993, 199 5, 1997, and 1999 2) It is also important to note that the percentage fo reign-born scholars who stay in the United States also varies by discipline. Table 1.3 is reproduced from Finn (2010) and captures this variance. 2 Data from Finn (2012), Figure 1. 6 Table 1.3 Percentage of Foreign Students Receiving Science and Engineering Doctorates in 2005 who were in the US, 2006-2007 3 The increase in departure of foreign-born scholars from the United States is not clearly understood. Current literature suggests that that competition for foreign talent is intensifying in an increasingly global marketplace. Additionally, supply and demand for talent could be another possible explanation for this trend. While visa re gulations in the United States make it very difficult for foreign-born scholars to stay, other countries are recruiting and hiring the highly skilled foreign-born graduates from American univer sities. Recently, the U.S. House of Representatives passed legislation that creates a p athway to permanent residency for international graduates of American universities wh o receive advanced degrees in STEM disciplines (Stratford, 2012). Nevertheless, it is disconcerting that it is not known to what degree foreign-born scholars are embedded in American univ ersities. Although in an increasingly global and knowledge-dr iven economy, attracting and retaining foreign-born scholars has become a high p riority for universities, industry, and governments around the world, little is known in hi gher education literature about the embeddedness of foreign-born faculty. This researc h gap is particularly problematic for several 3 Data from Finn (2010), Table 1. 7 reasons. Specifically, many STEM disciplines in Am erican higher education rely significantly on foreign-born scholars (Nelson & Rogers, 2005; NS B, 2003who, though highly productive, are least satisfied, among all faculty, with their expe rience in the American professoriate (Sabharwal, 2011; Corley & Sabhawal, 2007; Mamiseis hvili & Rosser, 2011, Kim et al., 2011a). Additionally, many foreign-born faculty feel they d o not belong in American academe (Wei, 2007). It is important for university administrato rs in the United States to better understand their foreign-born faculty who have already demonstrated a willingness to relocate to a new country. However, it is not known how embedded foreign-born faculty members are in their organization and the community, in particular, and the American professoriate, in general. This study begins to fill this gap in the prevailin g literature by assessing the embeddedness of foreign-born faculty in the America n professoriate. According to the embeddedness theory, highly embedded employees are less likely to voluntarily leave their organization (Lee, Mitchell, Sablynski, Burton, & H oltom, 2004). In examining their embeddedness, this study will focus on the percepti ons of the foreign-born faculty member about their perceived fit and links within the university where they work as well as within the community in which they live, and examine their per ceptions about perceived sacrifices if they were to leave their job and community. This study will contribute to existing scholarship on faculty recruitment, retention, and turnover. In this study I adopt the idea: fithink globally and act locallyfl to deepen and broaden our understanding of foreign-born faculty within the co ntext of their individual experiences both on- and-off the job. I rely on existing scholarship ab out foreign-born faculty to bring in the macro perspectives to illuminate our understanding of the ir experiences at the local or micro level within the American professoriate. To achieve this goal, I employ the embeddedness theory 8 because it provides a fibroad constellation of influ ences on employee retentionfl (Mitchell at al., 2001, p. 1104), and ficaptures variance in turnover above and beyond job attitudes, job alternatives, and job searchfl (Ramesh & Gelfand, 20 10, p. 809). Existing embeddedness research (Lee et al., 2004; Mitchell et al., 2001; Ng & Feldman, 2007) suggests that employees become enmeshed in their professional and community life to such an extent that it becomes too difficult (costly) to leave. This study assumes th at the more embedded foreign-born faculty members are in their organization and their communi ty, the more likely it is that they will stay in their jobs, and thus remain in the American profess oriate. Moreover, this study is proposing that embeddedness is dependent upon the perceptions of f oreign-born members™ fit and links in their organization and community, as well as what they pe rceive they would sacrifice if they were to leave. Definitions This research study examines the embeddedness of f oreign-born members in the American professoriate. It uses terms that may be defined differently across existing literature. The following definitions are central to this study and further elaborated throughout this manuscript. Embeddedness. Embeddedness is a construct that conceptualizes t he influences of social relations on economic behavior (Granovetter, 1985). Mitchell, Holton, and Lee (2001) developed this construct and associated it with fi–a net or a web in perception [in] which an individual can become [deeply enmeshed]fl (p. 1104). According to Mitchell and associates (2001), the individual stays on the job because too much would be sacrificed by leaving. Furthermore, embeddedness is the outcome of perceiv ed level and value added to links or attachments a person has to other people and / or a ctivities in the job and the community; the 9 extent to which the person perceives congruency bet ween his or her aspects of life and his or her job and the community in which they live; and the s acrifices a person would make in the process of leaving the employment and/ or the community (Ta nova and Holton, 2008; Mitchell et al., 2001). Foreign-born faculty. Existing literature uses interchangeably terms su ch as foreign- born and international faculty. Although initially these terms may be per ceived as synonymous, it is important to acknowledge the slight, but impo rtant distinction. Wells (2007) defined international faculty as fiall faculty members other than American-born U.S. citizensfl (p. 77). International faculty often refers to professors wh o are on a visa status, not permanent residents or naturalized citizens in the United States. Well s™ (2007) definition appears to include faculty members in temporary non-immigrant visa status and does not distinguish them from faculty members who may have immigrated. Additionally, fiin ternational facultyfl is also used in literature referring to American-born professors wh o are engaged in international work, such as those leading study abroad programs and conducting research abroad. The Merriam-Webster Dictionary offers the term diaspora referring to fipeople settled far from their ancestral homelands.fl This is a very use ful definition because it offers a tangible description about the sentimental and social status of international faculty. However, diaspora does not appear to resonate strongly with the popul ation of interest in this study. In a pilot survey of foreign-born faculty that was informally conducted by the author at a liberal arts college in the United States, seven out of ten resp ondents self-identified as fiinternational faculty,fl two preferred fidiaspora,fl and one respond ent favored the fiforeign-bornfl definition. This study adopted Wells™ (2007) definition of inte rnational faculty, fiall faculty members other 10 than American-born U.S. citizensfl (p. 77), because this definition seems to resonate better with the population of interest. Personal Attributes. Included in the data analysis are the faculty membe rs' demographic attributes (gender, family status, immigration stat us, and geographical region of origin), and professional attributes (rank, disciplinary field, and tenure at the institution and in the community). Geographic region of origin is determi ned from a pre-established list of world regions with country composition adopted from the U nited Nations Statistics Division (see Appendix C). STEM Disciplines. A wide range of disciplines are included in Scien ce, Technology, Engineering, and Mathematics (STEM); there is no ge nerally accepted list of STEM fields (Wasem, 2012; Kuenzi et al., 2006). Consistent wit h other studies (Wasem, 2012; Kuenzi et al., 2006), this study adopts the U.S. Immigration and C ustoms Enforcement (ICE) list of designated STEM degree programs, which includes six discipline s: (1) computer and information sciences; (2) engineering and engineering technologies; (3) b iological and biomedical sciences; (4) mathematics and statistics; (5) physical sciences; and (6) science technologies. A complete and updated list (14 pages) of ICE designated STEM fiel ds is available online 4. Theoretical Framework This study is informed by the embeddedness theory, which was developed by Mitchell and associates (2001). While other theoretical mod els have focused on employee turnover, the embeddedness construct focuses on why people stay i n their organizations. Embeddedness represents a broad array of influences on an employ ee™s decision to stay. Embeddedness consists of three dimensions: (a) links, (b) fit, and (c) sacrifices that explain the degree to which an 4 Complete list of ICE designated STEM fields is ava ilable online from http://www.ice.gov/sevis/stemlis t.htm 11 employee is enmeshed in his or her organization and community. Mitchell and associates (2001) postulated that embeddedness is a multidimensional causal aggregate that influences the employee to stay in the organization. They further argue that, in a path diagram, all the indicators would point towards each of the dimensio ns (links, fit and sacrifices), and from the dimensions to the aggregate construct (embeddedness ). Figure 1.2 below describes the causality of the embeddedness construct as theorized by Mitch ell and associates (2001). Figure 1.2 Embeddedness path diagram According to Mitchell et al. (2001), links refers to the perceived attachment the person has developed over time to people and activities in the organization and community. Fit refers to the individual™s perceived congruency or similariti es with his/her life spaces. Sacrifices refer to the individual™s perceived costs or the ease with w hich links can be broken when considering leaving the organization and/ or community. Chapte r Two provides more detailed descriptions of each of these dimensions of the embeddedness con struct. Embeddedness theory was selected for this study bec ause it is believed to be a stronger predictor of important organizational outcomes Œ re tention Œ than the conventional psychological explanations such as job satisfaction and organizat ional commitment (Mitchell et al., 2001). A growing body of research has demonstrated empirical support for the job embeddedness model ˘ 12 (Ramesh & Gelfand, 2010; Ramesh, 2007; Crossley, Be nnett, Jex, & Burnfield, 2007; Cunningham, Fink, & Sagas, 2005; Lee et al., 2004; Mallol, Holtom, & Lee, 2007; Mitchell et al., 2001). This study applies the embeddedness co nstruct to examine the foreign-born member™s embeddedness (intent to stay) and also determine th e degree to which this emerging theoretical construct can help us better understand foreign-bor n faculty. Research Questions and Hypothesis This study proposes that embeddedness is dependent upon the perceptions the foreign- born member has about his or her links and fit in t he organization and the community in which s/he lives, as well as the sacrifices (costs) s/he associates with leaving. The purpose of the proposed study is to investigate the embeddedness o f foreign-born in American universities, describe the foreign-born as a group, compare diffe rences between subgroup experiences, and identify patterns that may exist among them. The f ollowing are research questions that guide this study and the proposed hypothesis. 1. How embedded are foreign-born faculty members who a re employed in STEM disciplines at Master™s-level universities in the United States? 2. What personal and professional characteristics desc ribe the foreign-born faculty members who are employed in STEM disciplines at Master™s-le vel universities in the United States? 3. How are the demographic characteristics of these fo reign-born faculty members related to their embeddedness in their university and communit y? 3.1 How is the embeddedness of female foreign-born facu lty different than the embeddedness of male foreign-born faculty? 13 3.2 Does family status (single versus married/live with domestic partner, have/not have children enrolled in local K-12 school system, home ownership) matter in the embeddedness of foreign-born faculty in the America n Professoriate? 3.3 Does immigration status (work visa, US citizenship/ permanent resident) matter in the embeddedness of foreign-born faculty? 3.4 Does geographic region of origin matter in the embe ddedness of foreign-born faculty? 4. How are professional attributes related to embedded ness of foreign-born faculty members in STEM disciplines in Master™s-level universities? 4.1 Does length of employment matter in the embeddednes s of foreign-born faculty? 4.2 Does rank matter in the embeddedness of foreign-bor n faculty? 4.3 Does professional discipline matter in the embedded ness of foreign-born faculty? 5. What is the relationship between identifying as an international faculty member and their embeddedness at comprehensive universities in the U nited States? Purpose Statement This study recognizes that, in a transnational land scape of career opportunities and increasing global competition for talent in the STE M disciplines, international scholars have a wide array of employment options. My interest was in understanding what influences foreign- born faculty to stay at their universities and the communities in which they live. An emerging construct, called embeddedness, was developed by Mi tchell and associates (2001) to explain factors that influence employees to stay in their o rganizations and communities. Through quantitative and non-experimental survey research, this study sought to discover how select variables might help us understand the embeddedness of foreign-born faculty who are employed in the STEM disciplines at American universities. 14 Quantitative research methodology was employed in t his study because it allows for the greater possibility of drawing inferences and gener alizing findings beyond the selected sites. I utilized an online survey of participants clustered at randomly selected master™s level universities in the United States. Adopted from Mitchell and as sociates (2001), the survey instrument was designed to collect data on three dimensions of the embeddedness construct (links, fit, and sacrifices) and participant attributes. This study makes valuable contributions to emerging organizational literature on embeddedness theory an d the limited higher education knowledge about international faculty in the American profess oriate. Significance of the Study The significance of the current study is multidime nsional. This study is significant for economic reasons in relation to expanding competiti on for foreign talent across the globe and retention of quality faculty. Recruitment, retenti on, and turnover of highly skilled workers can be very costly for any organization. According to Kim and associates (2011c), fiHigh faculty turnover rates can be directly transl ated into lost income due to the investment the institutions have made in the facult y Œ this is particularly significant in the Science, Technology, Engineering, and Math (STEM) f ields for whom universities invest significant amounts of start-up funds, ranging from an average of $390,000 to $490,000 at the assistant professor level and about $700,000 to $1.44 million at the senior faculty levelfl (p. 1). The graying of the American professoriate (Doyle, 2 008), and the growing demand for scholars in STEM disciplines across the world, elev ates the necessity to better understand foreign-born faculty in order to improve retention and prevent voluntary turnover. Furthermore, this study is significant for practical and ideolog ical reasons, namely, foreign-born faculty constitute an expanding subgroup in the American pr ofessoriate. In addition to being highly skilled and productive scholars, foreign-born facul ty members also bring international 15 perspectives to colleges and universities that ende avor to prepare interculturally literate students, however defined. This research study is also significant for policy reasons in relation to the growing trend for talent-mobility and the internationalization of higher education. The study further contributes to the growing body of literature on campus interna tionalization as well as faculty and organizational development. It broadens existing l iterature on job embeddedness by applying this new construct to higher education (Biddle, 201 0; Hardcastle, 2010), in general, and understanding foreign-born faculty, in particular. Lastly, the proposed research study is significant from a humanistic perspective. At the time when this study was undertaken, very little was currently known about the experiences of foreign-born faculty. Although research shows that this subgroup is the most productive in research output (Mamiseishvili & Rosser, 2011, Kim et al., 2011b), foreign-born faculty are reportedly the least satisfied, among all faculty, with their jobs as compared to their Ameri can-born counterparts. The present research was the first study to examine the embeddedness of foreign-born faculty in the American professoriate. Other studies have examined job sat isfaction and productivity (Corley & Sabhawal, 2007; Sabharwal, 2011; Mamiseishvili & Ro sser, 2011), mobility (Kim at al., 2011b; 2011c), and sense of belonging (Wei, 2007) of forei gn-born faculty in American professoriate. In summary, the present study contributes to existi ng research by (1) broadening and deepening our understanding of foreign-born faculty and factors that influence their intent to stay at American universities, (2) examines the aggregat e dimensions of organizational and community embeddedness, (3) informs faculty retenti on policies and practices in American higher education, and (4) informs future research s tudies examining the international professoriate. 16 Dissertation Structure This manuscript is organized into five chapters. C hapter One sets the context for the study by way of introduction, the problem statement and research questions, while highlighting the potential relevance of the research. The secon d chapter provides a review of literature that is pertinent to the research questions in the present study and the theoretical construct (embeddedness) that will be used for interpreting f indings. Chapter Three describes the research design and data analysis methodology. In Chapter F our I provide a report of findings from the analysis of data and hypothesis testing. The last chapter provides an interpretation of findings and discussion of results through the theoretical f ramework and prior literature. The dissertation concludes with recommendations for future research, policy and practice. 17 CHAPTER 2: REVIEW OF LITERATURE The primary focus of this study was to examine the embeddedness of international faculty. It examined how embedded international f aculty are in the American professoriate, and in particular, how international faculty perceive t heir attachment ( links ) and congruency ( fit ) within their university and community, as well as h ow costly ( sacrifices ) they perceive relocating elsewhere to be. This chapter is organized as foll ows: first, I set the context by providing a literature-based overview of what describes interna tional faculty; next, I summarize relevant literature on employee mobility (voluntary turnover ). The chapter progresses with a review of literature relevant to the conceptual framework tha t is used in the design of this dissertation. The conceptual framework places the perspectives of the international faculty member in relation to the three dimensions of the embeddedness construct (links, fit, and sacrifices) and adapts the embeddedness construct to the intersecting contexts of work and community. In the final section I describe in more detail the embeddedness construc ts and synthesize research studies that have used the embeddedness model to study faculty in hig her education settings. Overview of International Faculty As the professoriate in the American higher educat ion is becoming increasingly diverse, international faculty have also become more diverse (Skachkova, 2007). During the mid-20 th century, the majority of international scholars in the United States came from Europe (Schuster & Finkelstein, 2006). Since the late 20 th century, the majority of international faculty in the American professoriate originate from Asian countri es, are increasingly female-entrants, younger in age, and are concentrated in the science and eng ineering disciplines (Manrique & Manrique, 1999; Schuster & Finkelstein, 2006). Moreover, sch olarship demonstrates that by the end of the 20 th century the fiproportion of foreign-born faculty wh o are women has more than doubled–and 18 the proportion of Asians among all foreign-born fac ulty has nearly trebledfl (Schuster & Finkelstein, 2006, p. 58). Table 2.1 below provide s percentages of international scholars in the United States. Table 2.1 Percentages of international scholars in the United States by country of origin Data from Open Doors Report 2008 5 Country of Origin China India South Korea Japan Germany Canada Percentage 22 9.4 9.3 5.4 5 4.5 Existing literature also indicates that a large maj ority of the international faculty in the American professoriate completed their education and trainin g in the United States. Specifically, according to Kim et al. (2001c), fiabout 64% of the faculty wh o were born and completed their education abroad (except for doctoral education) are naturali zed [US] citizensfl (p. 4). Existing scholarship has also demonstrated that int ernational faculty are predominately early-career and non-tenured (Nelson & Rogers, 2005 ; Theobald, 2007). Additionally, according to Kim and associates (2011b), the majori ty of international faculty members are unmarried, and for those who are married, their spo uses often do not work outside the home. According to embeddedness theory, these internation al faculty members, as described by Kim and associates (2011b), might be loosely embedded i n the American professoriate. According to Theobald (2007), conditions of departu re from the home country inform the likelihood or possibility of returning (after a few years) and the objectives of the immigrant in moving. Theobald posits that resources immigrants bring with them provide a measure of flexibility or constraint, pertaining to both emplo yment opportunities as well as personal situations. The social and economic contexts that greet immigrants create a framework in which adjustment to a new place and culture occurs, encou raging or discouraging immigrants from 5 Data from Kim et al., 2011c. 19 participating in society. This assertion is partic ularly significant for the community embeddedness dimension of the embeddedness construc t that informs this study. Theobald (2007) further postulates that internation al faculty members follow different paths as they determine how deeply to insert themse lves in their new situations, but these issues of assimilation affect everyone at some point along the path to adjustment into a new country; in other words, embeddedness of international faculty on-and-off the job in their new country is partially influenced by their assimilation paths. For example, it has been the experience of the author that international faculty who married U.S. born spouses inserted themselves deeply into their new community. In such instances, family tie s become important factors when considering leaving the newfound community and moving elsewhere . Thus, mobility becomes more constrained. In the next section I review relevant literature on employee mobility. Employee Mobility Employee mobility refers to both voluntary and invo luntary turnover. Voluntary mobility may occur due to life changing events such as child birth, disability, illness, retirement; or when the employee perceives more rewarding employment al ternatives. Porter and Steers (1973) posited that employee turnover is a function of unm et expectations. Involuntary mobility refers to when the organization terminates or does not ren ew one™s employment. Employee mobility, whether voluntary or involuntary, has the potential to be both functional and dysfunctional for the organization. On one hand, functional mobility or turnover, whether voluntary or involuntary, occurs when the termination does not d isrupt the organization (Griffeth and Hom, 2001). Mallol (2003) argues that in some instances , the departure of an employee may be beneficial to the firm because it may facilitate op portunities for change and innovation through 20 acquiring a new employee, or the possibility of rai sing productivity by letting go of a non- performing employee. On the other hand, dysfunctional mobility or turnov er refers to an organization losing a valuable member that it would rather retain. Accor ding to Mallol, Holtom and Lee (2007), fivoluntary turnover research has traditionally foll owed two main paths: (1) the study of turnover from the attitudinal perspective, considering job s atisfaction and organizational commitment, and (2) the influence of the labor market on turnover t hrough the perceived ease of movement and availability of alternative employmentfl (p. 35). March and Simon (1958) theorized that employee mobi lity hinges largely on the perceived ease of leaving the job and anticipated s atisfactory alternative employment; in other words, employee mobility occurs when there is diseq uilibrium in the respective employee- employer contribution-reward exchange (March and Si mon 1957). March and Simon posited that when employee contributions outweigh employee rewards, the employee transitions into a mobile state where desirability of the job and ease of movement are evaluated juxtaposed with associated costs with turnover. Thus, mobility bec omes a function of the severity of dissatisfaction with the job, the perceived impact of losing benefits associated with current employment, and the perceived probability of findin g compatible, if not, better employment (Mobley, 1977). Although research has demonstrated a positive relat ionship between employee mobility and job satisfaction (Hom & Griffeth, 1995; Carrell , Kuzmits and Elbert, 1992; Igbaria and Siegel, 1992; Labatmediene et al., 2007), job searc h intentions have not been found to be a strong predictor of mobility (Miller, Katerberg and Hulin, 1979). March and Simon (1958) postulated that voluntary turnover is based on ease and desirability of leaving one™s organization. 21 According to Brown (1967), faculty mobility is some times to be expected in higher education fi–because loyalty to discipline transcends loyalty to school and because teaching-research skills are readily transferable among schoolsfl (p. 25). Kim and associates (2011c) argued that internationa l faculty mobility is constrained by their immigration or visa regulations that tie the immigrant to the sponsoring institution. In other words, an international faculty member on a sponsor ed work visa (i.e., H-1b category) is obligated to maintain employment at the university that petitioned his/her visa status. In contrast, however, an international faculty member who has a permanent resident status or is a US citizen may opt to change employers and not sacr ifice losing his/her immigration status. The current study assumes a significant and positive re lationship between immigration status and embeddedness of international faculty. Moreover, international faculty mobility is strongl y related to their fiprofessional experiences and the institutions where they workedfl (Kim et al., 2011c, p. 4). Kim and associates also contended that international facult y mobility is best predicted by variables such as satisfaction with work conditions and compensation, tenure status, and length of time in the academic position. In the next section I review re levant literature on job satisfaction. Job Satisfaction Boswell, Boudreau and Tichy (2005) observed that a lmost every theory and study of turnover (i.e., March & Simon, 1958; Mobley, 1977; Steers & Mowday, 1981) included job satisfaction as an important antecedent. There is general agreement in the literature that job satisfaction is related to voluntary turnover (Grif feth, Hom & Gaertner, 2000; Mobley 1977; Porter & Steers, 1973). Hom and Kinicki (2001) pos ited that job satisfaction is the most commonly studied variable predicting both intention and actual mobility (turnover). 22 Gaylor (2004) studied organizational commitment and embeddedness of faculty members at public and private (Christian-based) colleges an d universities, and found that job satisfaction was perceived to be a stronger motivator than job e mbeddedness or organizational commitment. Gaylor also found that employees remain with an org anization because they are linked with the community in which they live, or believe they are a fit within the organization. This supports other findings that suggest employees remain in a p osition for non-work related factors (Hagedorn 2000; Mitchel et al. 2001; Drysdale 2005) . Scholarship has also demonstrated a positive relati onship between job satisfaction and interpersonal relationships, sense of community, co llaboration, teamwork, and congruence with personal values, compensation, and workplace climat e (Cohen 1974; Blank 1993; Baily 1997; Olanrewaju 2002). Steers and Mowday (1981) posited that job satisfaction, among other factors (i.e., organizational commitment and involvement), influenced employee™s intent to leave or stay at the organization. According to Hagedorn (2000), job satisfaction depends on antecedents of work that include institutional culture, quality of relationships with administration, quality of students and the relationships with them. Research has shown that job satisfaction is essential to faculty retention (Tack and Patitu 1992). Accordin g to Sanderson et al. (2000), more than forty percent of faculty members in the United States hav e seriously considered leaving their institutions and switching careers because of low j ob satisfaction. Expounding on job satisfaction as an antecedent to employ turnover, M itchell and associates (2001), posited that: fi–many people who leave (1) are relatively satisfie d with their jobs, (2) don™t search for other jobs before leaving, and (3) leave because of some sort of precipitating event [which, Lee and colleagues call a shock] rather tha n because of a negative attitude. In addition, the content or issues involved with shock frequently occur off the job; a spouse relocating is an examplefl (p. 1104). 23 Gender and Turnover Scholarship has also demonstrated that employee tur nover varies by gender, rank, and social status. Cropsey and associates (2008) analy zed data from a four-year study of reasons for attrition of women and minority faculty from a medi cal school and found that over 25 percent of women who left their jobs did so because of low sal aries, more than 22 percent reported issues with their supervisor, and nearly 30 percent of the participants felt they could not advance in their career at the same employer. The same study r eported these issues to be more significant for minority women who fiwere at lower academic rank s at the time they leftfl (p. 1111). Other studies (Menges & Exum, 1983, Rothblum, 1988) have observed higher attrition rates among women faculty compared to men, regardle ss of their tenure status. According to Gardner (2012) and Xu (2008), women faculty left th eir university because of lack of resources, inconsistent leadership practices, poor work-life b alance, and negative work climate. In 2008, Yonghong Xu analyzed the 1999 data from the Nationa l Study of Postsecondary Faculty to examine the attrition and turnover intention of mal e and female faculty in STEM disciplines and found no difference, based on gender, in their fiint ention to depart from academia, but women faculty had a significantly higher likelihood to ch ange positions within academiafl (p. 607). In the next section I review literature on employee in tent to leave their organization. Similar to job satisfaction, intent to leave informs the developme nt of embeddedness theory. Intent to Leave Intent to leave or stay at the organization has ser ved as a reliable proxy indicator for actual turnover (Mitchell et al., 2001; Holland, 19 92; Bluerdorn, 1982; Lee & Mowday, 1982; Steers & Mowday, 1981). Bluerdorn (1982) and Lee a nd Mowday (1987) argued that intent to leave was the best predictor of actual turnover, ab ove and beyond perceived alternative job 24 opportunities variable. Intent to leave, however, is influenced by several factors such as individual and professional attributes, as well as satisfaction with organizational and career dimensions such as compensation and morale (Johnsru d & Rosser 2002; Smart, 1990). Specifically, Johnsrud and Rosser (2002), posited t hat, fi–how individuals perceive the quality of their work lives has a direct impact on their mo rale, and their morale has a direct impact on their intention to leave their institutionfl (p. 536 ). Others factors are also associated with intent to l eave. Quality of life and relationships with colleagues are two significant factor influenc ing faculty members™ intention to leave (Weiler, 1985). Other research (Barnes, Agago, and Combs 1998) has demonstrated that fithe two most important predictors of faculty members™ i ntent to leave were a sense of frustration due to time constraints and a lack of sense of communit y at their institutionfl (as cited in Johnsrud & Rosser 2002, p. 521). This study argues that the s ense of community is very important to the population of interest in this study. Scholarship has also demonstrated that demographic variables and perceptual variables about work-life in the institution were significant in differentiating between those faculty members who stayed and those who left the organizat ion (Smart 1990; Olsen, 1993; Johnsrud and Heck, 1994). The embeddedness theory accounts for on-and-off the job factors that influence employee mobility; it focuses on why peop le stay on their jobs. In the next section I provide a more detailed description of the embedded ness construct and its dimensions. Embeddedness Theoretical Framework A growing body of research has demonstrated that hi ghly embedded employees are less likely to leave their jobs if they are attached to their organization (Allen, 2006; Bergiel et al., 2009; Crossley et al., 2007; Felps et al., 2009; Ho ltom & O'Neill, 2004), and especially if they 25 are enmeshed in their community (Lee et al., 2004; Mitchell et al., 2001). Embeddedness is closely connected/related to employee retention; it explains what influences the employee™s decision to stay (Mitchell et al., 2001). The embe ddedness construct focuses on why people stay in their jobs. It accounts for the individual™s pe rceptions about (1) her links to other people, teams, and groups, (2) her perceptions of fit with the job, organization, and the community, and (3) what she perceives she would sacrifice if she left her job and perhaps also the community in which she lives (Mitchell et al., 2001). Next, I f urther describe the embeddedness constructs links, fit, and sacrifice . Links to the Organization and Community. Mitchell et al. (2001) theorized that links to the organization and the community describe the extent to which an individual is attached to other people, activities, and the number of people who depend on her within the organization and community. Links can be both formal and informal t ies. The number of colleagues and advisees who depend on the faculty member serves as one exam ple of links to the organization. Owning a home, having children in the school system or famil y members living in the same area, serve as examples of links to the community. Mitchell and a ssociates (2001) theorized that the greater and stronger the links between the individual and t he people in the organization and the community, the more attached or enmeshed the employ ee feels; thus, less likely to leave. Research studies have demonstrated that the likelih ood a person will leave an organization decreases by the number of links she h as in the organization (Burt, 2001; Maertz & Griffeth, 2004), largely because links to people wi thin an organization increase organizational attachment (Mossholder, Setton and Henagan 2005; Fr iedman and Holtom 2002). Similarly, empirical research has demonstrated that links to t he community, by way of community involvement, decrease employee mobility (Cohen 1995 ). Mitchell and associates (2001) theorize 26 that having a spouse and children increases links t o the community. Their theory is grounded on earlier empirical research by Lee and Maurer (1999) , who found a positive relationship between higher retention rates and community links such as having a spouse and the number of children employees had. The proposed study theorizes that in ternational faculty members who have children enrolled in the local K-12 education syste m will demonstrate greater links to the community, and thus higher embeddedness. Fit in Organization and Community. According to Mitchell at all (2001), fit refers t o the congruency between the individual™s needs and e xpectations and the organizational provisions and the community™s amenities. These tw o dimensions, organizational fit and community fit, describe the extent to which the ind ividual perceives the organization and community to be compatible with her preferences and dispositions. An example of high organization fit would be a faculty member who has a passion for teaching and works at a teaching university. An example of high community fit for an international faculty would be living in an ethnically diverse area that is welcom ing to foreigners. The higher the congruency between the employee and the workplace and communit y, the more likely it is for the employee to stay (Mitchell et al., 2001). Chatman (1989) po sited that organizational fit mediates the employee™s socialization experiences, productivity, and job satisfaction. Olsen and associates (1995) caution, however, that organizational fit is more complex than employee™s perception; it fimay be confounded by misperceptions of the individ ual™s professional values and strengthsfl (p. 272). Furthermore, scholarship has demonstrated that the employee™s perceived congruency with the organization™s norms and values is positiv ely related to tenure with the organization (Bretz and Judge, 1994). In addition to individual ™s perceptions of organizational fit, or 27 congruency, another variable is the perception of c ongruency with job roles and responsibilities, which has been found to be significantly correlated to intent to leave (Cable & Judge, 1996; Lauver & Kristof-Brown, 2001). Similarly, research has demonstrated a positive correlation between employee mobility and perceived congruency with community amenities (Feldman and Bolino 1998; Shaffer and Harrison 1998). Organization and Community Sacrifice. The sacrifice dimension of job embeddedness describes the individual™s perception of what she must give up if she were to leave the organization, and perhaps also the community in whi ch she lives. In other words, if the employee were to voluntarily leave the organization , she would give up associated benefits such as status and potential promotion, as well as certa in perks associated with job seniority. Other examples of job-related sacrifices may also include accumulated vacation time, retirement and insurance benefits, established sources of support and collaborative networks, and balanced teaching and research assignments (Mitchell et al., 2001). Community sacrifices are also of significant import ance in the embeddedness construct. Examples of community sacrifice could mean removing children from their schools, one™s spouse having to also resign from a job, leaving a support network and an area that feels familiar and safe. These variables could also weigh heavily on an employee™s decision to stay. Mitchell et al. (2001) theorized that the more an employee p erceives she would have to give up when leaving, the more enmeshed she would feel in the or ganization and the community. Empirical research has demonstrated a positive corr elation between perceived sacrifices and organizational embeddedness (Mitchell et al., 2 001; Feldman and Bolino 1998; Shaw, Delery, Jenkins and Gupta, 1998). Mitchell and ass ociates (2001) found that community sacrifice, such as leaving a safe community or leav ing a community in which one feels respected, 28 was negatively related to voluntary turnover. Howe ver, research is limited, and not parsimonious (findings do not appear to be easily g eneralized across populations), regarding community sacrifice and employee mobility (Ramesh, 2007). This is especially of concern regarding the limited understanding of internationa l faculty in American professoriate. The proposed study has the potential to bridge this gap in the higher education literature. In summary, research studies have confirmed the inf luence of embeddedness on employee mobility (turnover) among samples of facul ty at Christian colleges (Biddle, 2010; Hardcastle, 2010), information technology professio nals (Berrios-Ortiz, 2012; Dinger, 2011), state correctional officers (Bergiel et al., 2009), financial professionals (Allen, 2006; Felps et al. , 2009; Lee et al., 2004), and healthcare professiona ls (Crossley, Bennett, Jex, & Burnfield, 2007; Holtom & O'Neill, 2004; Mitchell et al., 2001). Th e decision to stay or leave is affected by non- work influences such as having a spouse and childre n (Lee and Maurer 1999), and prioritizing family needs (Maertz & Campion, 1998; Steers and Mo wday, 1981; Mobley, 1979). Summary of Embeddedness Research Empirical research using the embeddedness theory in higher education is very limited. In this section I synthesize the few studies that have used the embeddedness theory to study university professors™ intent to stay (Biddle, 2010 ), organizational commitment (Hardcastle, 2010), and job performance (Emmerik & Sanders, 2004 ), at select types of institutions. Biddle (2010) examined the impact of job embeddedness, sat isfaction, and job search behavior on faculty members (N-385) intent to stay at Christian colleges. He found that job embeddedness was a positive predictor of intent to stay (p = 0.0 02); but community embeddedness was not a significant predictor of intent to stay (p=0.548). Further, Biddle found that job satisfaction was 29 also a positive predictor of intent to stay (p=0.00 0) at Christian colleges. Job search, however, was a negative predictor of intent to stay (p=0.000 ) in the Biddle (2010) study. Hardcastle (2010) also used the embeddedness theor y to examine the institutional commitment of faculty (n = 652) teaching in select institutions that are members of the Council for Christian Colleges and Universities (CCCU). In his study, Hardcastle compared embeddedness results between ministerial and non-mi nisterial faculty and found that the former group demonstrated higher embeddedness than the lat ter group; calling was observed as a positive and significant variable in commitment, sa tisfaction, and embeddedness scales. Further, the author posited that affective organizational co mmitment was observed to be significantly related to organizational fit. Additionally, facul ty who owned their homes and were married demonstrated a higher institutional commitment (low er intent to leave) than their counterparts, though the difference was reported as not statistic ally significant (owning a home p-value 0.094; and marital status p-value 0.061). Another study by Emmerik and Sanders (2004) employ ed embeddedness theory to examine how different employment relationships in a cademic settings (tenured versus non- tenured appointments) are associated with different types of job performance efforts. (n=1342). Authors proposed that there would be a positive association between temporal embeddedness and: (a) compliance, (b) contextual pe rformance, and (c) this relationship would be stronger for non-tenured than tenured faculty. Using regression analysis, Emmerik and Sanders (2004) found support for hypothesis 1a and 1b, but not for 1c. Tenured faculty members™ intention to stay was significantly relate d to compliance ( =.09, p<.05), and contextual performance ( =.13, p<.01). Further, the authors were interested in investigating whether a positive association between network embe ddedness and (a) compliance, (b) 30 contextual performance, and (c) this relationship w ould be stronger for non-tenured than tenured faculty. The authors demonstrated through their fi ndings that there was a positive association between network embeddedness and work compliance ( =.11, p<.01) and contextual performance ( =.16, p<.01) for non-tenured faculty, but the same was not true for job performance of tenured faculty. Emmerik and Sanders (2004) also measured the associ ation between institutional embeddedness and (a) compliance, (b) contextual per formance, and (c) this relationship was expected to be stronger for non-tenured than tenure d faculty. Results from statistical analysis showed a positive relationship between embeddedness and compliance ( =.26, p<.01), and embeddedness and contextual performance ( =25, p<.01). However, institutional embeddedness was less important in explaining contextual perform ance of non-tenured faculty than of tenured faculty members. Conclusion Although the embeddedness construct has not been s tudied extensively in the higher education context, this emerging theory presents si gnificant potential to inform our understanding of the experiences of international f aculty members in the U.S professoriate. While, historically, research has focused on why pe ople leave their organizations, this study adopts the embeddedness construct because it focuse s on why people choose to stay. Studying the embeddedness of international faculty in the U. S. professoriate is especially important for several reasons, such as: (1) international faculty are a growing and very important subgroup in the U.S.professoriate; (2) international faculty ar e very productive, yet less satisfied than their counterparts who were born in the U.S.; and (3) int ernational faculty may be willing to be mobile, especially given the rising demand for STEM talent in the global marketplace. 31 CHAPTER 3: RESEARCH METHODS This study examined the embeddedness of foreign-bor n faculty who are employed in the STEM disciplines at comprehensive universities in t he United States. This chapter is organized in the following sections: (a) overview of the rese arch design; (b) participant selection; (c) human subject privacy and confidentiality; (d) data collection; (e) and data analysis procedures. I conclude with an overview of anticipated limitati ons. Overview of Research Methodology This study utilized quantitative research methodolo gy and analyzes empirical data from an online survey of a sample of foreign-born facult y members clustered within randomly selected public universities in the United States. The unit of analysis in the present study is the individual faculty member. The population of inter est is university professors who were born outside United States and are employed in STEM disc iplines at randomly selected public universities. Research Paradigm The present study was conducted from a post-positiv ist epistemology. This approach informed understanding of complex relationships bet ween variables (Scotland, 2012) by systematically analyzing the perspectives of partic ipants from their own context (Creswell, 2009). The goal of the present study was to learn about the embeddedness of foreign-born faculty members in the American professoriate. The research was exploratory. The empirical data collected was self-reported. Research Design The present study employed quantitative, non-experi mental, cross-sectional survey research methods. According to Creswell (2003), th e quantitative design is appropriate when 32 fiexamining the relationship between and among varia bles is central to answering questionsfl (p. 153). The cross-sectional study approach was adopt ed because data were gathered from the population of interest across different attributes (Creswell, 2003). This study also describes foreign-born faculty as a group; comparing differences between highly embedded and not-as-highly embedded responde nts. I sought to examine relationships between the variables and explain variation observe d based on participant personal characteristics (gender, family status, immigration status, and geographical region of origin) and professional attributes (tenure at their institutio n, rank, and discipline). Using an online survey for the present research stu dy allowed for including a larger number of participants more efficiently. Accordin g to Creswell (2003), the survey method allows the researcher to collect and report data in a numerical format on the same variables from a larger sample and to report on characteristics of subgroups within the sample. Furthermore, the survey method allowed for collecting and analyzing data that looks for variation in specific variables across a participant sample (DeVaus, 2002 ). Before administering the instrument, I piloted the survey with a small sample of foreign- born faculty. The goals for piloting the instrumen t included: (1) to test and ensure the clarity of instructions, (2) to assess the time needed to comp lete the survey so that the researcher does not lose credibility or subjects give up and discontinu e their participation, (3) to evaluate the clarity of the questions asked, and (4) to increase the con fidence level for the overall design of the study (Pallant, 2011). Feedback received from the pilot survey indicated that completing the survey required an average of seven minutes. In addition, participant s in the pilot survey commented that the questions asked were clear. One participant sugges ted rewording one scale item measuring 33 community links. The decision was made to keep the original wording in order to ensure instrument validity. Further, two participants not ed the repeated scale item measuring fisacrifice- organizationfl dimension of embeddedness construct. The repeated scale item resulted from combining the six single-item scale developed by Cu nningham et al. (2005) with the original expanded scale developed by Mitchell and associates (2001). Based on this feedback, I removed the duplicated question. Site Selection and Population of Interest The sites (universities) selected for this study s erved as a means for identifying the population of interest; no institutional data was c ollected. The population of interest in this study was the foreign-born faculty who are clustered with in STEM disciplines at randomly selected institutions. The number of foreign-born faculty i n the United States is unknown. Because rosters of faculty identifying individuals as fifore ign-bornfl are nonexistent, to make the study feasible, I applied the following site selection cr iteria. First, I downloaded a customized report of fiMaster™sfl classified colleges and universities from the Carnegie Classification of Institutions of Higher Education website 6; the initial report consisted of 727 institutions. This included institutions awarding at a minimum fi50 master™s deg rees and fewer than 20 doctoral degrees per yearfl (The Carnegie Foundation, 2010). Next, I ref ined my search by filtering out private colleges and universities and narrowed the list to 271 public universities. I decided to select public universities because there is a significant variation in the mission and values between private and public institutions. For example, many private universities are religiously affiliated, serve a different student population, tend to have smaller STEM discipline departments, and often draw fewer foreign-born scholars. In contras t, public institutions tend to have more clearly 6 Source: http://carnegieclassifications.iu.edu/look up_listings/custom.php 34 recognizable STEM discipline departments and draw a larger number of foreign-born STEM scholars. Further, using a random number generating formula in Microsoft Excel, I extracted from the master excel worksheet a list of 20 such public universities. Next, I created a list of STEM departments and then visited the public webpages of such departments at the randomly selected institutions and reviewed faculty pages to identify potentially eligible participants. In cases where faculty curriculum vitae or biographical info rmation were not publically available on university webpages, I skipped that institution and selected additional sites. As previously stated, each university in this study served only a s a means to identify potentially eligible participants. No institutional data were collected . Participant Selection and Recruitment To identify potentially eligible participants, I r eviewed each faculty member™s biography and curriculum vitae searching for the following id entifiers: (1) completed undergraduate degree in a country other than the United States, and / or (2) commonly used references (for example, grew up in x-country, moved or immigrated from y-co untry) indicating that a faculty member may have been born outside United States. I copied and pasted into an Excel Spreadsheet file the names and email addresses of all potentially eligib le participants. To protect participants™ privacy, the names and email addresses of subjects were recorded in a password protected Excel file and stored on a secure password-protected comp uter in a locked office. One assumption in participant selection included t hat faculty members who completed their undergraduate degree outside the United State s were likely born outside United States. To control for participant selection error, the online survey required participants to respond to a fiyesfl or finofl question that asked fiwere you born in the United States?fl This question had a 35 skip logic feature; a positive response disqualifie d the participant from this study. Subjects were recruited through an email invitatio n administered through SurveyMonkey.com Œ the online data collection surve y system utilized for this study. The email address of each potentially eligible subject was en tered into SurveyMonkey™s list of recipients. Follow-up email reminders were initially sent in a 15-day interval. Understanding that faculty members receive large volumes of emails, weekly sub sequent email reminders were sent to encourage participation. Only one subject replied to the email invitation asking not to be contacted again. In addition to potential requests from subjects asking to be removed from the study, the names and addresses from the list of fipe rmanently failed deliveryfl emails were also removed from the follow-up sample list. Sample The population sample for the present study was 63 0 potentially eligible participants. Only 598 of these received the survey email invitat ion, 32 invitations did not get delivered due to inactive email address and potential participants h aving opted out from SurveyMonkey. A total of 149 subjects (n=149) responded to the invitation for participation, yielding an overall response rate of 25 percent. Of the 149 respondents, 29 sel f-identified as being born in the United States, and were not eligible to participate in the current study. Throughout the conceptualization of the current st udy and during the data collection phase, the researcher was cognizant of the importan ce of securing an adequate sample size for reliable statistical analysis. According to Cohen (1977), sample size is very important in quantitative research because reliability of sample results depends on sample size; fiincreases in sample size increase statistical powerfl (p. 7). W ithout doubt, sample size is very important for reliability of sample results. However, literature in this area is inconsistent about the 36 recommended number of cases needed for quantitative research. Stevens (2001) posits that no less than 100 participants are needed in order to u phold the power of statistical tests. Other authors (Sapnas & Zeller, 2002) postulate that fiund er some circumstances 100 - or even 50 - cases are sufficientfl (Tabachnick and Fidell, 2007, p. 613). On the other hand, there is general consensus in t he prevailing literature that the larger the sample size, the better. And, although sample size recommendations in the literature are both general and broad; the ratio of observations per va riable measured appears to be more important than the actual sample size (Pallant 2011). Some a uthors have posited that a ratio of as few as five-to-one observations per variable measured may be sufficient; others have argued that a ten- to-one observations ratio is adequate in most cases (Nunnally, 1978). The goal for the current study was to have a large enough sample size so tha t hypotheses would be tested at the 95% confidence level and a 5% significance level. Keeping in mind the difficulties of identifying po tentially eligible subjects, the sample size for the present study was estimated to be grea ter than 100. Given the general rule of thumb in sample sizing, the minimum number of cases or ob servations (completed surveys) needed to measure the proposed variables in this study was 10 0. That is ten observations per independent variable. Specifically, this study measured the ef fect of ten independent variables on one dependent variable. The independent variables were professional characteristics (rank, length of tenure, and discipline), and demographic variables (gender, visa or immigration status, geographic region of origin), and family status Œ w hich is a categorical variable that includes marital status, having children in the K-12 system, and homeownership. The dependent variable was embeddedness, which is a continuous categorical variable; a construct that includes six dimensions (fit-university, fit-community, links-un iversity, links-community, sacrifice- 37 university, and sacrifice-community). Anticipating potential issues associated with the sample size and to ensure an adequate number of observations (completed surveys) for stat istical analysis, I adopted the following methods for addressing sample size issues: (a) I co mmitted to increase the number of sites (universities) in order to grow the population and sample size; or (b) reduce the number of independent (dummy) variables (geographical region of origin, discipline) from the statistical equation. Further, it is important to note that a smaller sample does not change how embeddedness is measured on individual cases; but, it does mean that the statistical analysis cannot test as many independent variables for stati stical significance (in other words, their statistical impact on embeddedness). Acknowledging that faculty members already receive many requests to fill out surveys, it was understood that a low survey response rate woul d be possible. The response rate for the present study was 25 percent. The literature sugge sts that a 30 percent response rate for online survey research is adequate for generalizing of fin dings from the sample to the population (Hamilton, 2003; Kittleson, 1997). However, the li terature also suggests that the ratio of observations per variable measured is more importan t than the response rate. The ratio of observations per variable measured in the present s tudy was 12 to 1 for most variables. Data Collection Empirical data was collected using an online survey instrument. Participants were invited via email to complete an online survey admi nistered through SurveyMonkey.com. Access to the instrument was provided via email, wh ich included a unique link generated by the survey. The survey was housed in a secured and pri vate Survey Monkey web account. 38 Protecting Human Subjects. Every measure was taken to protect the identity o f willing participants to the maximum extent allowable by law . Participants agreed to an informed consent that outlined and described the purpose, po tential risks and benefits from participating in this study. The online survey instrument did not c ollect any personal identifying information and it did not track any IP addresses. The informed co nsent assured the respondents that their responses would be anonymous and would be kept in c onfidence. In addition, findings from the survey analysis are reported only in the aggregate. Participation in this study was voluntary and respondents had the choice to discontinue their par ticipation at any time without providing a reason. Participants were provided with contact in formation for the researcher, the director overseeing this study, and The Office of Human Rese arch Protection Program of Michigan State University. Instrumentation. The present study utilized the survey instrument t hat was developed by Cunningham, Fink, and Sagas (2005). In their st udy7, these authors adapted the 40 item scale that was developed by Mitchell and associates (2001 ) and designed and tested their global-item scale (a single-item indicator) for each embeddedne ss dimension. The instrument included one single item that measured the six dimensions of the embeddedness construct. Table 3.1 lists the single scale items and the associated construct. Table 3.1 Embeddedness: global-item scale Scale item Construct 1. I feel compatible with my organization Organization fit 2. I feel compatible with where I live Community fit 3. I feel a strong link to my organization Organization links 4. I feel a strong link to my community Community links 5. I would sacrifice a lot if I left this job Organization sacrifice 6. I would sacrifice a lot if I left this community Community sacrifice 7 Extension and further examination of the embeddedn ess construct. 39 In their comparative study, Cunningham and associat es (2005) measured the predictability of the embeddedness construct on int ent to stay on the job by analyzing data from embeddedness of two independent samples. These aut hors concluded that the figlobal-item scale predicted stay intentions–over [and beyond] Mitchel l et al.™s multi-item measurefl (p. 334). It appears that one of the motivations that Cunningham and associates had in designing and testing their six item embeddedness scale was to increase p articipant response rate. According to Hinkin (1995), fi[s]cales with too many items can cr eate problems with respondent fatigue or response biases (p. 972). Cunningham and associates (2005) concluded their st udy acknowledging that fifurther application and refinement of both scales are neede d in order to have full confidence in the validity of the measuresfl (p. 334). Further, in th e literature search for the present study, no other studies were found that utilized Cunningham et al.™ s single-item scale. Thus, the instrument™s validity and reliability was limited to the authors ™ claims that in their comparative study, their single-item scale predicted employee embeddedness o ver and above the multi-item scale developed by Mitchell and associates (2001). The p resent study makes important contributions to future use of the six item embeddedness scale. For the present research, scale items measuring par ticipants™ embeddedness were anchored on a 5-point Likert scale: Disagree-Strong ly, Disagree, Neutral, Agree, and Agree Strongly. The fiAgree Stronglyfl responses on the in strument compute a high level of embeddedness and fidisagree stronglyfl compute a low level of embeddedness. In addition to the Likert rating embeddedness scale items, the instrum ent included single-choice questions that collected data on nominal variables such as gender, rank, and length of tenure; as well as fiyesfl and finofl responses for questions pertaining to the binary variables (family status descriptors). 40 The family status variable included several questio ns: (1) fiare you married or live with a domestic partner?fl (2) if you are married or live w ith a domestic partner, does he or she work outside the home?fl (3) fido you have K-12 school-age children?fl and fido you own the house you currently live in?fl Scale items for family status were adapted from the Mitchell et al. (2001) study. These authors postulated that employees who are married, have children, and own their houses are more embedded in their community because they have many links; breaking these links would be appraised as high sacrifice. Validity. Validity is a multidimensional construct that requ ires due diligence towards ensuring content validity, criterion validity, and construct validity ; all of which are very important aspects in research design (Muijs, 2004). Content validity of the instrument was determined through the review of the literature and the theoretical framework that informed this study. The instrument was adapted from Cunningham and associates (2005) who examined and confirmed the ficonvergent, discriminant, and predic tive validityfl of the instrument (p. 319). Further, a review of precedent literature revealed that the Mitchell and associates (2001) instrument has been used in several empirical studi es (Biddle, 2010; Hardcastle, 2010; Gaylor, 2006; Lee at al., 2004) that measured the same dime nsions of the embeddedness constructs that are employed in this study. Content validity was a lso ensured through the inclusion of the fifull contentfl of scale items (Neuman, 2011, p. 212) meas uring the corresponding dimension of the embeddedness construct. Finally, content validity was also ensured through face validity from a pilot study that sought feedback from participants who fit the descriptors of the population of interest in this study. Similar to content validity, criterion validity was ensured through the review of the literature and theoretical framework informing this study (Muijs, 2004). Other studies (Berrios- 41 Ortiz, 2012; Hardcastle, 2010; Biddle, 2010; Holtom & Inderrieden, 2006; Ramesh, 2007; Gaylor, 2006; Cunningham et al., 2005; Lee et al., 2004; Mallol, 2002;), have demonstrated that the embeddedness construct predicts employee mobili ty (intent to leave or stay) with a high degree of confidence. The convergent and discrimin ant validity of the Cunningham and associates™ (2005) global-item embeddedness scale i s limited to their own study. A literature search did not yield any other empirical studies ha ving used the global-item scale. In their study, Cunningham and associates (2005) identified such a gap in the literature and recommended that future studies should use both ins truments to inform future theory and instrument development. Reliability. In their comparative study, Cunningham and associa tes (2005) found comparable reliability coefficients between their g lobal-item embeddedness scale and the original Mitchell and associates™ (2001) 40-item sc ale. They posited that fi–all reliability estimates (Cronbach™s ) [were] above .52 and most [were] above the tradit ional .70 cutofffl (p. 325). Several empirical studies (Berrios-Ortiz, 20 12; Hardcastle, 2010; Mallol, 2002; Biddle, 2010; Lee et al., 2004; Holtom & Inderrieden, 2006; Ramesh, 2007; Gaylor, 2006) have used the Mitchell and associates (2001) instrument and yield ed reliability coefficients above acceptable levels needed for quantitative research. For examp le, Gaylor (2004) used the Mitchell and associates (2001) instrument to examine business fa culty and the analysis yielded a 0.87 average reliability coefficient. Ramesh (2007) studied emb eddedness of customer service staff in the U.S. and India, and reported an average reliability coefficient of 0.68. Biddle (2010) studied faculty at Christian colleges and reported an avera ge reliability coefficient of 0.81. For the present study, the reliability estimate (Cronbach ) was .930 (n=116). 42 Scales and Measures As previously stated, this study adopted the instru ment that was developed by Cunningham and associates (2005) that measures one continuous dependent variable, embeddedness. According to the embeddedness theory, the embedde dness scale maps the instrument subscale items from the indicators of th e six dimensions (on-and-off the job links, fit, and sacrifices); and from the dimensions to the agg regate construct of embeddedness (Cunningham et al., 2005; Mitchell et al., 2001). Data on demographic attributes (personal and professional) were collected by asking participants about their personal attributes (gender, family status, immigration status, geographical area of or igin) and professional descriptors (rank, tenure, discipline). The demographic categories serve as t he independent variables. The yes/no responses were standardized by using dummy variable s (yes=1, no=2), which are numerical variables used in statistical analysis to represent subgroups within the study sample. Dummy variables allow for using statistical equations to represent multiple groups and look for statistically significant subgroup differences. Ta ble 3.2 below lists the variables measured in this study. Table 3.2 Variables of Interest Dependent Variables Independent Variables Organizational Embeddedness Personal Attributes Professional Attributes Links Œ organization Gender Rank Fit Œ organization Family status Org. Tenure Sacrifices Œ organization Immigration status Discipline Community Embeddedness Geographic Region Links Œ community Identify as International Fit Œ community Sacrifice Œ community 43 Research Questions, Hypotheses, and Statistical Tes ts The purpose of the proposed study was to investigat e the embeddedness of faculty members who are foreign-born and employed in STEM d isciplines at randomly selected universities in the United States. fiInstitutionfl i n this study simply serves as a means for identifying eligible participants. No data are col lected on the institutions. The present study sought to answer the following re search questions: 1. How embedded are foreign-born faculty members who a re employed in STEM disciplines at Master™s-level universities in the United States? 2. What personal and professional characteristics desc ribe the foreign-born faculty members who are employed in STEM disciplines at Master™s-le vel universities in the United States? 3. How are the demographic characteristics of these fo reign-born faculty members related to their embeddedness in their university and communit y? 3.1 How is the embeddedness of female foreign-born facu lty different than the embeddedness of male foreign-born faculty? 3.2 Does family status (single versus married/live with domestic partner, have/not have children enrolled in local K-12 school system, home ownership) matter in embeddedness of foreign-born faculty in the American professoria te? 3.3 Does immigration status (work visa, US citizenship/ permanent resident) matter in the embeddedness of foreign-born faculty? 3.4 Does geographic region of origin matter in the embe ddedness of foreign-born faculty? 4. How are the professional attributes related to embe ddedness of foreign-born faculty members in STEM disciplines in Master™s-level unive rsities? 4.1 Does length of employment matter in the embeddednes s of foreign-born faculty? 44 4.2 Does rank matter in the embeddedness of foreign-bor n faculty? 4.3 Does professional discipline matter in the embedded ness of foreign-born faculty? 5. What is the relationship between identifying as an international faculty member and their embeddedness at comprehensive universities in the U nited States? Hypotheses were developed in relation to the resear ch questions. The null hypothesis for the primary research question was: international f aculty members are not highly embedded. The alternative hypothesis was: international faculty m embers are highly embedded. I also expected that embeddedness would vary between and among subg roups. For example, I hypothesized that males, those who are married or live with domestic partners, those who have children enrolled in local K-12 school system, and those who are tenured and not on a temporary visa would display higher levels of embeddedness than their counterpar ts. This hypothesis was based on the general perception that faculty members who are single, emp loyed in non-tenure track positions, and are hired on temporary visas have not established stron g links in the community where they live and would not perceive relocating as costly 8 as their counterparts. I expected that foreign-bo rn faculty who are US citizens would display higher le vels of embeddedness than international faculty members who are employed on a temporary wor k visa. I also expected that international faculty who are tenured would be more embedded than their untenured counterparts. The assumption was that tenured faculty would perceive they have a lot to sacrifice if they were to leave their university. I also hypothesized that foreign-born faculty membe rs who are highly embedded in their community would also display high embeddedness in t heir university. I expected tenured faculty 8 According to Embeddedness Theory, people who are m arried, have a spouse or significant other who work s outside the home, and/or have children enrolled in local K-12 school system, perceive relocating as ve ry costly; doing so would require making a lot of sacrifices. 45 members and those in higher ranks would be highly e mbedded in their community because these individuals are perceived to be beyond the figetting tenuredfl time demands, they would have children, and as such be more involved and connecte d to their community. This hypothesis is based on the assumption that the tenured faculty me mbers who have established a family would also own their house rather than rent. My expectat ion was that faculty members who own their house would also have stronger connections to their community and would perceive that they would have a lot to sacrifice if they were to reloc ate. In addition, I hypothesized that foreign-born STEM faculty members who do not identify as international would demonstrate higher embeddedn ess than their international-identified counterparts. The assumption for this hypothesis w as that identifying as international may influence the foreign-born faculty members™ percept ion of fit at their university and community. I expected a positive relationship between identify ing as an international faculty member and the overall embeddedness. I hypothesized that identify ing as an international faculty member also predicts embeddedness of foreign-born faculty membe rs at their university. Methods of Analysis In this section I describe the steps, techniques, a nd the statistical tests utilized to analyze the data. I measured one continuous dependent vari able - overall embeddedness , which is a categorical variable comprised of six categories ( fit -organization, fit -community, links - organization, links -community, sacrifice - organization, and sacrifice -community), and seven independent variables (personal attributes: age, f amily status, immigration status, geographical region of origin; and professional characteristics: rank, tenure, and discipline). Demographic variables were converted into dummy variables, repr esenting and measuring non-metric responses for the associated scale items on the ins trument. Data was analyzed utilizing the 46 following statistical tests: descriptive statistic s, T-test for binary variables (gender, marital status, homeownership, having children, and spouse working outside the home, and identifying as international). Analysis of Variance (ANOVA) was used for nominal variables such as discipline, rank, geographic region of origin, immi gration status, and length of tenure. Correlations were conducted for identifying the rel ationship and strength of the scale items measuring the continuous dependent variable. In th e next section I provide a description and rationale for the selected statistical tests. Descriptive Statistics. I used descriptive statistics to generate graphica l representation of findings and visually present data on mutually incl usive and exclusive variables such as gender, family status, and immigration status. Findings on nominal variables such as geographical region of origin, rank, and professional discipline were also reported using descriptive statistics. Graphs, tables, and charts were utilized to visuall y represent findings. T-Tests. Independent sample t-tests were conducted to ident ify differences in embeddedness based on binary variables (gender and immigration status). The independent- samples t-test is commonly referred to as a between-groups design and assumes that each case must have scores on two variables that are mutually exclusive (i.e., male vs. female). The t-test evaluated whether the mean embeddedness value for o ne group differs significantly from the mean embeddedness value of the other group. The in dependent-samples t-test was used to measure the difference between the means of two ind ependent groups, those who are permanent residents or naturalized US citizens, and those who are on a work visa. In addition, the independent t-test measured differences between the means of the female and male faculty subgroups. That is, I evaluated whether the means for two independent groups were significantly different from each other. 47 Analyses of Variance (ANOVA). ANOVA was used to test for statistically significan t differences between subgroups, based on categorical variables such as personal attributes (family status, geographic region of origin) and profession al characteristics (rank and discipline). Categorical variables were converted into dummy var iables and analyzed using one-tailed and two-tailed ANOVA. According to Mitchell and assoc iates (2001), the embeddedness construct specifies a clear direction, depicted in the path d iagram (Figure1.2), thus, one-tailed and two- tailed tests with Analysis of Variance (ANOVA) were utilized. Embeddedness is the dependent and continuous variab le, and the rest of the variables are independent variables. Personal and professional d emographic characteristics are continuous categorical variables that are converted into dummy variables. The inclusion of dummy variables was the basis for analysis of variance (A NOVA). Each parameter b(X) becomes a mean comparison test for the categories of X. Interpretation of the results is straightforward for parameters with only two categories (0/1 coding). Parameters with more than two categories were also converted into dummy variables (2, 3, and so forth). Pearson Correlation Coefficient. I used the Pearson correlation coefficient to investigate relationships and strength (and directi on) of associations between independent variables (personal and professional characteristic s, and identifying as international) and the dependent variable embeddedness (comprised of the s ix dimensions of the embeddedness construct). According to Pallant (2011), the Pears on correlation is used to fi–explore the strength of the relationship between two continuous variablesfl (p. 103). The embeddedness construct and its dimensions are continuous variabl es, and I was interested in measuring both the direction (positive or negative) and the strength o f the relationship between independent and dependent variables. The Pearson correlation coeff icient indicates either a positive or negative 48 correlation between variables: a positive correlati on indicates that as the values of one variable increase, so do the values of the other; a negative correlation would indicate that as values of one variable increase, the other variable decreases (Pa llant, 2011). Procedures and Statistical Tests. To reiterate, the purpose of the proposed study wa s to investigate the embeddedness of faculty members who are foreign-born and employed in STEM disciplines at randomly selected masters-level univ ersities in the United States. fiInstitutionfl in this study simply serves as a means for identifying eligible participants. No data were collected on the institutions. The leading research question in this study was: ho w embedded are international faculty in STEM disciplines in Master™s-level universities? According to the embeddedness theory, embeddedness is an aggregate construct (Law, Wong, & Mobley, 1998) that is formed from six dimen sions: org-fit, org- links, org-sacrifices, com.-fit, com.-links, and com.-sacrifices (Mitchell et al., 2001). Thus, embeddedness is an aggregate composite variable. To measure embeddedn ess, I first computed the mean for each of the six dimensions of embeddedness constructs. By computing the mean of the six dimensions means, fithe composite embeddedness scores equally w eighs the influence of distinct dimensionsfl (Mitchell at al., 2001, p. 1111). I al so calculated the mean of the corresponding scale items measuring organizational embeddedness a nd community embeddedness respectively. Then, I generated the overall embeddedness scores b y calculating the mean of the organization- and-community means. As described above, scale ite ms measuring embeddedness were anchored on a 5-point Likert scale with responses r anging from fistrongly disagree=1fl to fiagree strongly=5fl. Participant responses ranged between 1 and 5 on each of the six embeddedness subscale items; thus, their composite embeddedness score ranged between 6 and 30, where six 49 represented the lowest level of embeddedness and th irty represents the highest level of embeddedness, respectively. The means and standard deviations are reported on the following table. Table 3.3 Dependent Variable and sample means N Mean Std. D. Embeddedness 120 3.71 .78 Fit-University 120 4.37 .81 Links-University 119 4.0 1.0 Sacrifices-University 119 3.63 1.16 Fit-Community 120 3.72 .95 Links-Community 119 3.49 1.05 Sacrifices-Community 119 3.08 1.20 In addition to examining the overall embeddedness, I was interested in understanding what describes the international faculty and how th eir demographic attributes are related to their embeddedness. The second research question explor ed what personal and professional characteristics describe international faculty in S TEM disciplines in master™s-level universities. Descriptive Statistics. I explored questions 2 and 3 in order to develop a n understanding about what characteristics describe the internation al faculty studied as a group. Data collected for question 2 were categorical in nature and were converted into dummy variables. Initial background profiles were created for international faculty as a group, summarizing data on all variables described above in the measurement sectio n on demographics. Simple descriptive statistics generated percentages of subgroups. I u sed graphical representation to visually present data on mutually inclusive and exclusive variables such as gender, family status, immigration status, and the geographical region they identify w ith. In addition, the instrument collected data on subjects™ rank, tenure (years at the university and community) and professional discipline. Graphs, tables, and charts were utilized to visuall y represent findings from survey items informing question 3. 50 Inferential Statistics. The third research question, and its subquestions, investigated for variance in embeddedness based on personal and prof essional attributes. I hypothesized that female international faculty would display lower le vels of embeddedness than male international faculty. Mean scores were calculated for female and male binary variables. An independent sample t-test was used to explore subgroup difference. The hypothesis underlying this question was based on the assumption that female faculty enc ounter greater challenges in STEM disciplines than male faculty. Other research (Ols en, Maple, & Stage, 2014; Xu, 2008; Cropsey et al., 2008; Gardner, 2012) has documented a multi tude of factors leading to female faculty attrition, such as (a) lack of support, (b) work-li fe balance issues, (c) low job satisfaction, and (d ) negative workplace climate. In addition, I was interested in learning whether f amily status (single versus married/living with a domestic partner, having/not having children enrolled in local K-12 school system, and homeownership) matters in the embeddedn ess of foreign-born faculty in the American professoriate. Hypothesis 3.2 stated that foreign-born faculty members who are married or live with a domestic partner will displa y higher embeddedness compared to their single counterparts in the American professoriate. Hypothesis 4 stated that foreign-born faculty whose spouse or domestic partner works outside the house are likely to be more embedded as compared with faculty members whose spouse or domes tic partner is a homemaker. Additionally, I hypothesized that embeddedness woul d vary significantly between foreign-born faculty members who have children enrolled in the l ocal K-12 education system as compared with the foreign-born who do not have children enro lled in the local K-12 school system. For homeownership, I hypothesized that foreign-born fac ulty members who own their house are likely more embedded as compared with the faculty w ho rent. 51 For hypotheses testing for family status variable c ategories, I utilized independent sample t-tests. First, I calculated the mean scores for th e following categorical variables: marital status (single vs. married/living with a partner), have or do not have children enrolled in local K-12 education system, and homeownership. Research question 3.3 sought to examine embeddednes s of foreign-born faculty members based on their visa or immigration status. Visa or immigration status is a categorical variable. The scale item asked if participants were naturaliz ed U.S. citizens, permanent residents (green card holder), or on a temporary work visa. Hypothe sis 7 stated that foreign-born faculty members who are naturalized US citizens are likely to have the highest embeddedness scores when compared with faculty members who are permanen t residents and those who are on temporary visas. I also proposed that foreign-born faculty members who are permanent residents in the U.S. are likely to have higher embeddedness scores when compared with faculty members who are on a temporary visa. The hypothesis under lying research question 3.3 was based on the assumption that faculty who are employed on tempora ry work visas do not feel strongly linked to their university and their community; thus, they ar e less embedded. I expected that faculty members on a temporary work visa are more likely to be renting instead of owning a house. Their counterparts, on the other hand, who are perm anent residents or U.S. citizens, are more likely to be linked to their community and universi ty. Before testing the aforementioned hypotheses, I calculated the mean scores for the im migration status categorical variables: (a) work visa, and (b) permanent resident/U.S. citizen. Analysis of variance (ANOVA) test was utilized to test for difference in embeddedness amo ng subgroups. Research question 3.4 explored the difference in e mbeddedness of foreign-born faculty members based on their geographical region of origi n. Prior literature has documented the 52 challenges foreign-born professors encounter as the y learn to integrate into the American professoriate. For example, in an opinion piece, N gwainbi (2006) described the challenges African and African-American professors encounter i n interacting with their colleagues. However, it is unclear from the existing literature how similar or different the experiences of foreign-born faculty are within the group. Althoug h foreign-born faculty members are not homogenous, I did not expect their embeddedness to vary significantly based on their geographical region of origin. Hypothesis 8 stated that no significant difference will be observed for embeddedness of foreign-born faculty members ba sed on their geographic region of origin. Origin is a categorical variable that included 5 g eographical regions of origin: Africa, Asia, Europe, South and Central America, and North America. Before testing the hypothesis, I first calculated the mean scores for the geographic al region of origin subcategories. Next, ANOVA was used to test for subgroup differences. The fourth research question explored how the profe ssional attributes of foreign-born faculty are related to their embeddedness in the Am erican professoriate. Specifically, research questions 4.1, 4.2, and 4.3 sought to examine diffe rence in embeddedness based on length of tenure, rank, and discipline, respectively. Accord ing to the embeddedness theory, the longer an employee has worked at the organization the more em bedded he or she becomes (Mitchell et al., 2001). Thus, the assumption for this research ques tion was that faculty members with more years at the university likely feel they have more to sacrifice if they were to leave. Hypothesis 9 stated that foreign-born faculty members with the l ongest tenure at the university are likely to be more embedded as compared with faculty with few yea rs of employment. Prior to testing for subgroup difference based on length of tenure at th e university, I first calculated the mean scores 53 for the tenure variable. ANOVA was conducted to te st for difference in embeddedness of foreign-born faculty members based on their length of tenure. For the rank variable, I was first interested in i nvestigating if tenured versus not tenured status was significant in the embeddedness of the f oreign-born faculty members who participated in the present research. Next, I was interested in learning whether embeddedness varied significantly between respondents™ tenured ranks (a ssistant professor tenure-track, associate professor, and full professor). The assumption for this question was that tenured faculty members perceive they have a lot to sacrifice if th ey were to leave their university. Additionally, untenured international faculty members employed on tenure-track positions also perceive they are more linked, and thus more embedded, than their counterparts who are employed in non- tenure-track positions. Thus, the following hypoth eses were tested for the rank variable: foreign- born faculty members who are employed in non-tenure track positions are significantly less embedded in the American professoriate as compared with tenure-track and tenured faculty members. And, foreign-born faculty members who are at the full professor rank will likely display significantly higher embeddedness as compar ed with faculty members who are at the associate and assistant professor ranks. The mean scores were calculated for the rank variable and its subcategories. ANOVA was utilized to test for subgroup variance in embeddedness. In addition, I was interested in learning whether e mbeddedness of foreign-born faculty varies based on their discipline. Prior research i s unclear about the effect of discipline on one™s embeddedness. Thus, the researcher hypothesized th at no significant subgroup difference will be observed based on discipline in the embeddedness of foreign-born faculty members. The mean scores were calculated for the fidisciplinefl categor ical variable and ANOVA was used to test for discipline effect on subgroups for embeddedness. 54 Finally, I was interested in examining the directio n and strength of the relationship between embeddedness of foreign-born STEM faculty m embers and their identification as international faculty. Correlation analysis was us ed to investigate for the relationship between these two variables. In addition, I was interested in testing whether the fiidentify as internationalfl variable was a significant factor in the embeddedness of foreign-born STEM faculty members. Single sample t-test analysis was utilized to test the null hypothesis, which stated, there is no significant difference in the e mbeddedness of foreign-born STEM participants who identify as international faculty and those who do not. Limitations The present study has several limitations, which in clude: (a) participant selection procedures, (b) sample size and response rate, and (c) instrument design. Identifying potentially eligible participants for the current study proved to be very challenging for two reasons. First, universities do not keep lists of foreign-born facu lty. Second, this study was not designed to compare the embeddedness of foreign-born faculty me mbers with the embeddedness of U.S.- born faculty. Thus, it was not appropriate to send an invitation for participation in this study to all STEM faculty members and then filter out respon dents who were not foreign-born. A comparative study would facilitate the possibility of reaching a larger population sample. Participant Selection. For the present study, to identify potential partic ipants I used the following procedures. First, I created a list of r andomly selected comprehensive universities that had clear listings of STEM departments. Next, I re viewed each faculty member™s bio or curriculum vitae posted on university and departmen t public web pages, looking for foreign-born identifiers such was born or completed undergraduat e or graduate degrees in a country other than USA. Finally, I created a master list containing t he email addresses of potentially eligible 55 participants. To make the study feasible within ti me constraints for degree completion, it was necessary to keep the number of sites small. After many hours of reviewing several thousand webp ages, the population sample for the current study included 630 potentially eligible par ticipants. Of these, only 598 received the email invitation through the SurveyMonkey platform; 32 invitations did not get delivered due to inactive email address and potential participants h aving opted out from SurveyMonkey. A total of 149 participants (n=149) responded to the invita tion for participation, yielding an overall response rate of 25 percent. Of the 149 respondent s, 29 self-identified as born in the United States, and were not eligible to participate in the current study. After filtering for missing data, only 120 observations were included in the data ana lysis. Sample Size and Response Rate. Sample size for the current study was deemed adequate for the statistical analysis relevant to t he research questions. However, caution must be exercised when making inferences about findings fro m the current study. The current study™s sample was not representative of foreign-born STEM faculty at universities across the United States. For example, the sample for this study did not include respondents from Research One universities, Ivy League or private institutions of higher education. In addition, the overall response rate for the current study was 25 percent. A higher response rate would strengthen the confidence level regarding findings from the statis tical analysis. Future studies may find better ways to ensure a larger sample size and response ra te. Instrument Design. The online survey for current study did not utiliz e forced choice application for data collection. This allowed for participants to skip responding to some scale items, which created the issue of missing data. Al so, with regards to exploring the effect of having K-12 school-age children on participants™ em beddedness, the survey instrument did not 56 collect data on whether or not respondents had chil dren in other age groups. Future studies should consider collecting data on the age ranges o f participants™ children in order to improve interpretation of findings. Whether participants h ave children and controlling for the children™ age would inform interpretation of findings regardi ng the effect of this variable on community embeddedness. In addition, the instrument utilized in this study did not include items measuring respondents™ intent to leave their university and/o r their community. This study focused on investigating participants™ embeddedness and how th eir embeddedness differed based on personal and professional characteristics. The present study provides a snapshot of embeddedne ss of international faculty members who are foreign-born and employed in STEM disciplin es at comprehensive randomly selected public universities across the United States. This snapshot of their embeddedness was based on analysis of respondents™ perceptions of their own o n-and-off the job links, fit, and sacrifices . The constructs examined in this study may have been con text-laden and may change over time. In addition, this study did not collect institutional and context-specific data. Although the institutions were randomly selected, the institutio n served only as a means for selecting a sample of the population of interest that is clustered at these institutions. A longitudinal and national study that measures similar constructs with a large r sample over time would provide a better understanding of the embeddedness of international faculty in the American professoriate. 57 CHAPTER 4: DATA ANALYSIS AND FINDINGS In this chapter I present research findings from an alyzing empirical data collected through an online survey instrument. In the first s ection, following listing of research questions and summary of data, I report findings from descrip tive analysis to provide a profile of foreign- born faculty who are employed in the STEM disciplin es at randomly selected comprehensive universities in the United States. Biographical da ta are reviewed on personal characteristics such as gender, family status, immigration status, geogr aphic region of origin, as well as professional attributes such as rank, discipline, and years empl oyed at the current university. Next, using statistical analysis of survey results, findings ar e reported on aggregate embeddedness scores for the study sample, on-and-off the job embeddedness s cores, as well as observations from comparing subgroups. The chapter concludes with a summary of findings relevant to the research questions. The purpose of this study was to examine the embedd edness of foreign-born faculty who are employed in the STEM disciplines at comprehensi ve universities in the United States. I sought to answer the following research questions: 1. How embedded are foreign-born faculty members who a re employed in STEM disciplines at Master™s-level universities in the United States? 2. What personal and professional characteristics desc ribe the foreign-born faculty members who are employed in STEM disciplines at Master™s-le vel universities in the United States? 3. How are the demographic characteristics of these fo reign-born faculty members related to their embeddedness in their university and communit y? 3.1 How is the embeddedness of female foreign-born facu lty different than the embeddedness of male foreign-born faculty? 58 3.2 Does family status (single versus married/live with domestic partner, have/not have children enrolled in local K-12 school system, home ownership) matter in embeddedness of foreign-born faculty in the America n Professoriate? 3.3 Does immigration status (work visa, US citizenship/ permanent resident) matter in the embeddedness of foreign-born faculty? 3.4 Does geographic region of origin matter in the embe ddedness of foreign-born faculty? 4. How are the professional attributes related to embe ddedness of foreign-born faculty members in STEM disciplines in Master™s-level unive rsities? 4.1 Does length of employment matter in the embeddednes s of foreign-born faculty? 4.2 Does rank matter in the embeddedness of foreign-bor n faculty? 4.3 Does professional discipline matter in the embedded ness of foreign-born faculty? 5. What is the relationship between identifying as an international faculty member and their embeddedness at comprehensive universities in the U nited States? Data Overview Procedures for data collection are described in Cha pter 3. A total of 630 potentially eligible respondents were invited to participate vo luntarily in the current study by completing an online survey. Of the 630 invited participants, 14 9 responded; 481 did not respond. Among those who did not respond, 21 opted out and 11 emai l invitations for participation fibouncedfl and were undeliverable. Follow-up reminders were sent, initially bi-weekly and then weekly, to the sample population. To secure a large enough sample , the population sample was increased throughout the data collection period by searching for and adding potential participants to the list of invited participants. Data collection efforts c oncluded with securing over 100 complete surveys. During the design of the current study, t he goal was to obtain a sample size of at least 59 100, resulting in ten observations (completed surve ys) for each measured variable. The current study analyzes findings from 120 observations; the other 29 respondents identified as fiborn in USA,fl thus, were disqualified from the current stud y. The instrument collected data on demographic varia bles such as gender, length of tenure, rank, discipline, marital status, immigration or vi sa status, world region of origin, having K-12 school-age children, spouse/partner working outside the home, and home ownership. The instrument included a 6-item scale measuring one de pendent variable, embeddedness . As described in the previous chapters, embeddedness is a composite variable, a 2-by-3 dimensional construct that includes on-and-off the job embeddedness (in this case, UE=university embeddedness, and CE=community embedd edness). UE is measured by three scale items: (1) fit-university, (2) links-university, and (3) s acrifices-university . CE is also measured by three scale items: (1) fit-community, (2) links-community, and (3) sac rifices- community. The following section provides results from the da ta analysis on the variables of interest. Results In this study, embeddedness is the dependent variab le and was measured by a six-item scale that was developed by Cunningham, Fink and Sa gas (2005). Table 4.1 provides the six scale items and the corresponding construct dimensi on that each item measures. Participants recorded their responses to scale items on a Likert scale (1=Disagree Strongly, 2=Disagree, 3=Neutral, 4=Agree, 5=Agree Strongly). Table 4.1 Scale items measuring embeddedness const ruct Items Construct dimensions 1 I feel I am compatible ( a good match ) with the university where I work Fit-university 2 I feel I am compatible ( a good match ) with the community where I live Fit-community 60 Table 4.1 (cont™d) 3 I feel a strong link ( attachment ) to the university where I work Links-university 4 I feel a strong link ( attachment ) to the community where I live Links-community 5 I would sacrifice a lot if I left this job Sacrif ice-university 6 I would sacrifice a lot if I left the community w here I live Sacrifice-community Correlation analysis was conducted for instrument s cale items measuring the continuous dependent variable and its sub-dimensions. Results are reported in Table 4.2. Analysis for instrument validity revealed that scale items are p ositively and highly correlated to embeddedness (aggregate measure). Cronbach™s Alpha was 0.892 and 0.902 on the standardized scale items measuring aggregate embeddedness; corre lation was significant at the 0.05 level. The high correlation between scale items and embedd edness demonstrate that the scale items indeed measure the intended variable. Table 4.2 Correlations Variables 1 2 3 4 5 6 1. Embeddedness 2. Fit Œ University .519 ** 3. Links Œ University .658 ** .593 ** 4. Sacrifices Œ University .659 ** .246 ** .387 ** 5. Fit Œ Community .811 ** .310 ** .387 ** .368 ** 6. Links Œ Community .815 ** .272 ** .418 ** .350 ** .783 ** 7. Sacrifices Œ Community .831 ** .208 * .322 ** .550 ** .674 ** .722 ** **. Correlation is significant at the 0.01 level (1 -tailed). *. Correlation is significant at the 0.05 level (1- tailed). Participant responses to scale items measuring the aggregate embeddedness and its sub- dimensions were converted into numerical values tha t ranged from 1 to 5. An embeddedness score was calculated for each respondent by calcula ting the mean of means for each of the scale items. Table 4.3 exhibits sample aggregate means ( x=sample mean) and standard deviations for aggregate embeddedness and the sub-dimensions scale items (n=120). 61 Table 4.3 Means and Standard Deviation Scale Items Mean SD Embeddedness (aggregate) 3.714 0.783 University Embeddedness 3.997 0.803 Fit Œ University 4.375 0.810 Links - University 4.000 1.000 Sacrifices - University 3.630 1.163 Community Embeddedness 3.430 0.961 Fit - Community 3.471 0.954 Links - Community 3.487 1.048 Sacrifices - Community 3.084 1.183 As exhibited in Table 4.3, overall on a scale of 1- 5, respondents demonstrated high aggregate embeddedness means. It is also interesti ng that university embeddedness was greater than community embeddedness across all categories. Descriptive Analysis. One of the goals of the current study was to const ruct a profile of foreign-born faculty members who were employed in S TEM disciplines at comprehensive universities across the United States. This sectio n reports findings from descriptive statistical analysis on personal and professional background va riables such as gender, marital status, having children enrolled in local K-12 school system, spou se/partner working outside the house, visa or immigration status, world region of origin, discipl ine, rank, and length of tenure. In this section I also report findings on what described participants who identify as foreign-born faculty members. Identify as International. As indicated in the sample composition section of C hapter 3, only respondents who self- identified as fiborn outside United Statesfl were eli gible to participate in the current study. 77.3 percent of respondents identified as international faculty, and 22.7 percent did not perceive themselves as international faculty; one respondent chose not to answer this item. Findings are reported in Figure 4.1 (n=120). 62 Figure 4: 1 Identify as International Faculty It is rather interesting that nearly 23 percent of participants in the current study did not identify as fiinternationalfl faculty members, although they s elf-identified as foreign-born. It is plausible that some of the respondents who did not identify a s international might be born to American expatriate parents employed outside the United Stat es. One other possible explanation is that after living in the United States for a long period of time and having become US citizens, foreign-born faculty might identify more as America n than international. This finding is important for other faculty members and university administrators to take into consideration when referring to all foreign-born faculty members as fiinternationalfl because such practice may have implications for faculty inclusion. Gender Representation. Gender representation in the sample (n=120) includ ed 31.4 percent female (n=37) and 68.6 percent males (n=81) ; two respondents chose not to report their gender. Note that the number of males was more tha n double that of female faculty. This finding is consistent with existing literature rega rding gender distribution in STEM disciplines. With regards to gender distribution by rank, the ma jority (40.5%) of female respondents were represented at the associate professor rank, follow ed by tenure-track assistant professors (27%), and full professor (18.9%). Similar to females, th e majority of male respondents reported they 63 were at the associate professor rank (39.5%). In c omparison to females, male full-professors were 29.6 percent. Twenty-one percent of males sel f-identified as assistant tenure-track professors. Within the gender category, the percen tage of female instructors not on the tenure- track was more than double that of males (5.4% vers us 2.5% respectively). With regards to gender gap, females made up 37 percent of the assis tant tenure-track professors, 31.9 percent of the associate tenured professors, and 22.6 percent of professors in the sample for this study. For gender distribution by length of tenure (years employed at current university), 50 percent of male respondents reported having been em ployed ten years or more at their current university. For the female category, 38.9 percent of respondents reported having been employed ten years or more at their current university. A l arger percentage of females (30.6%) reported employment in the 6-8 years range as compared with the male respondents (21.8%). A similar pattern was observed at the 0-2 years of employment category. Nineteen point two percent of male respondents reported employment at the 3-5 yea rs, compared to 16.7 percent of female respondents. Overall, a majority of respondents (5 3%) had been employed for ten or more years at their current university, followed by 6-8 years (28%), 3-5 years (21%), and 0-2 years (12%) respectively. Findings are reported in the followi ng table (n=114). Table 4.4 Gender representation by length of emplo yment Length of Employment Total 0-2 3-5 6-8 > 10 Gender Female 5 6 11 14 36 Male 7 15 17 39 78 Total 12 21 28 53 114 With regards to gender representation by discipline , the biggest gap was observed in the technology field where there were five times as man y males compared to female participants 64 (though sample size within the category was very sm all). Findings are reported in Table 4.5 (n=118). Table 4.5 Gender representation by discipline Discipline Total Sciences Tech. Engineering Math Gender Female 20 1 10 6 37 Male 42 5 22 12 81 Total 62 6 32 18 118 Family Profiles. Family status is a categorical variable that inclu des the following sub- categories: (a) marital status (married/living with domestic partner), (b) having a spouse/partner who works outside the home, and (c) having children enrolled in local K-12 school system. Regarding marital status, 88.7 percent of the respo ndents were married or lived with a domestic partner; and, 73.3 percent of them reported that th eir spouse/partner worked outside the home. A majority of the participants (54%) reported they di d not have children enrolled in the local K-12 school system. For the homeownership variable, 82. 6 percent of the respondents reported they owned their dwelling place. For the gender variable and family status categorie s, an overwhelming majority of male respondents (nearly 91%) reported being married or lived with a domestic partner; compared to only 83.3 percent of the female respondents. Surpr isingly, the percentage of female respondents who reported having school-age children was the sam e as those who reported they did not have school-age children. In contrast, there were fewer (44.0%) male respondents who reported having school-age children as compared with those ( 56.0%) who had school-age children. The data analysis also revealed that 93.5 percent o f the female respondents reported their spouse/partner worked outside the home. In contrast , only 65.3 percent of the male respondents indicated their spouse/partner worked outside the h ouse. This finding suggests that, in 65 comparison to female participants in this study, a significant percentage of male respondents seem to enjoy greater support at home since their s pouse/partner does not work outside the home. Immigration Status Profile. Figure 4.4 (n=120) provides a graphical representa tion of the data according to immigration or visa status. More than half (51.7%) of the respondents reported being naturalized U.S. citizens, followed by permanent residents (35.8%), and those holding a work visa (12.5%) of one type or another. This finding is interesting with regards to the percentage of faculty members on work visa. At the time this study was conducted, the U.S. Government increased the work visa allotment for hi ghly skilled individuals in the STEM disciplines. The researcher expected a larger perc entage of work visa participants in this study sample. Figure 4.2 Representation by immigration status Geographical region of origin. Research question 3.4 investigated for difference in embeddedness based on respondents™ geographical reg ion of origin. Origin was a categorical variable that included five world regions: Africa, Asia, Europe, Central and South America, North America, and Oceania. Respondents were asked to select one of the five geographical regions of origin they identify with. A majority o f respondents self-identified with Asia (40.3%), 66 followed by Europe (32.8%), Africa (10.9%), North A merica (8.4%), and Central and South America (7.6%). There were no respondents from Oce ania. Figure 4.3 exhibits data distribution by respondents™ geographical region of origin (n=11 9). Figure 4.3 Representation by geographical region o f origin Profiles by Rank. Figure 4.4 provides a graphical distribution of da ta by rank (n=118). A quick inspection of the data shows that a majorit y of respondents self-identified as associate professors (39.2%), followed by full professors (27 .5%), and assistant professors employed in tenure-track positions (22.5%). 7.5 % were assista nt professors employed in non-tenure track positions, and 3.3% were non-tenure track instructo rs. 67 Figure 4.4 Rank distribution For distribution of rank status by gender, the perc entage of male respondents (77.4%) in the tenured professor rank was more than three time greater than females (22.6%). At the tenured associate professor rank and non-tenure-tra ck assistant professor rank, there were twice as many males (68.1% and 66.7%, respectively) as th ere were female respondents (22.6% and 33.3% respectively). The gender gap was smaller at the assistant professor tenure-track rank: 63% males as compared with 37% females. Discipline Profiles. Figure 4.5 (n=120) shows that a majority of foreig n-born faculty members were clustered in the sciences (52.5%), fol lowed by engineering (26.7%), mathematics (15%), and technology disciplines (5.8%). The tech nology discipline had the fewest complete observations in the dataset. 68 Figure 4.5 Discipline representation Figure 4.6 reports gender distribution across disc iplines (n=118). An inspection of figure 4.6 shows that in almost all disciplines there were twice as many male as there were female respondents. In the technology discipline, the per centage of male respondents was slightly greater than double the percentage of female partic ipants. Figure 4.6 Gender across discipline Length of Employment. In higher education length of employment is often synonymous with tenure; the longer the faculty member is emplo yed at the university the more likely it is that 69 he or she progresses through the ranks (tenure syst em). Descriptive analysis revealed that a majority of the respondents (47.4%) reported employ ment in the 10 plus years category; followed by 24.1% in the 6-8 years category, 18.1% were in the 3-5 years category, and 10.3% of cases were in the 0-2 tenure category. The rese archer was interested to examine the length of employment by rank; expecting not to find any respo ndents at the full professor rank to be employed in the 0-2, 3-5, and 6-7 years of employme nt categories. Figure 4.7 exhibits years of employment at the current university by rank (n=118 ). One interesting observation is that there were tenu red associate professors and tenured professors in the sample who were employed at their current university for as few as 0-2 and 3-5 years, respectively. Taking into consideration the number of years it takes to progress through the tenure system in the American professoriate, th e following graph demonstrates the mobility of full professors and associate professors from a different university to the current university. This finding suggests that tenured status is not sy nonymous with life-time employment at the same university. Although mobility past achieving tenure may not be as common, this finding illustrates the willingness of foreign-born faculty members to change institutions and their ability to negotiate rank as part of their employment offer . 70 Figure 4.7 Years of employment by rank Figure 4.8 exhibits distribution of years of employ ment by gender. A quick inspection of the graph reveals that the gender gap is very large at the ten and above years of employment and smaller at the other three categories for lengths o f tenure. The gender gap for length of tenure prompted the researcher to dig deeper into the data and look for any variation for distribution of years of employment based on having school-age chil dren. Prior literature has demonstrated that, due to publish or perish pressures associated with the early-career faculty stage, female faculty members consider childrearing to be detrime ntal to their career progress. 71 Figure 4.8 Years of employment by gender. Table 4.6 reports distribution of years of employme nt data for respondents with and without school-age children (n=113). Within the ye ars of employment category, a majority of respondents (64.2%) who were employed for 10 plus y ears at the current university reported they do not have school-age children. 66.7 percent of r espondents in the 6-8 years of tenure reported having school-age children. ANOVA shows that there is not a significant difference across years of employment categories (p value = .507). This in dicates that whether or not a faculty member had school-age children was not statistically signi ficant for their length of employment at the current university. Table 4.6 Length of employment by having school-ag e children Years Total Number 0-2 3-5 6-8 > 10 Have school-age children Yes 4 11 18 19 52 No 8 10 9 34 61 Total 12 21 27 58 113 Table 4.7 exhibits distribution of years of living in the community data for respondents who had and did not have school-age children enroll ed at a local K-12 school (n=112). Within 72 the length of community tenure category, a majority of respondents (58.6%) who lived in the same community for 10 plus years reported they did not have school-age children at the time the current study was conducted. 63.6 percent of respo ndents in the 6-8 years of community tenure reported having school-age children. ANOVA showed that there is not a significant difference across categories of years of living in the communi ty (p value = .350). This indicates that whether or not the foreign-born faculty members had school-age children was not statistically significant for their length of tenure in the commu nity. Table 4.7 Years of living in the community by havi ng school-age children Years Total Number 0-2 3-5 6-8 > 10 Have school-age children Yes 5 9 14 24 52 No 7 11 8 34 60 Total 12 20 22 58 112 Inferential Analysis The following section reports results as they relat e to the specific research questions. This section is organized by the research questions , hypotheses, and propositions that were investigated in the present study. The principal r esearch question investigated the aggregate embeddedness of the foreign-born faculty members. I was interested in understanding how embedded the foreign-born faculty members are who are employed in STEM disciplines at Master™s-level universities in the United States. In addition, I was interested in examining variance in embeddedness between subgroups based on their personal and professional characteristics. Further, exploring variance in a ggregate embeddedness based on community versus university embeddedness was also of interest in the current study. The statistical analysis relevant to each of the research questions is repor ted in the following pages. 73 Embeddedness of Foreign-Born Faculty. Data analysis showed that as a group, faculty members who are foreign-born and employed in STEM d isciplines at comprehensive universities included in this study are highly embedded (x =3.71). Figure 4.8 shows distribution of aggregate means of foreign-born faculty members™ embeddedness across the sample (n=120). Embeddedness values on the interquartile of Figure 4.8 represent the very few observations of very low embeddedness (n=1) and low embeddedness (n =3). On the opposite end of the spectrum, 5 percent of cases demonstrated very high embeddedness. The statistical analysis of distribution of means shown in Figure 4.9 also reve aled that the overall range between very low embeddedness scores and very high embeddedness scor es was very small (.613). A majority of respondents demonstrated high embeddedness; the sam ple mean was 3.71, and a substantial percentage of observations fall above the sample me an. Figure 4. 9 Distribution of embeddedness data Findings did not support the null hypothesis, which stated that foreign-born faculty members are not highly embedded in the American pro fessoriate. The researcher accepted the 74 alternative hypothesis: foreign-born faculty member s are highly embedded. Findings about university and community embeddedness are reported in the next section. University and Community Embeddedness. The following figures exhibit results measuring university and community embeddedness. D ata analysis showed that respondents had a higher mean for university embeddedness (x =4.00) than community embeddedness (x =3.43). Figure 4. 10 Distribution of university embeddedne ss data Figure 4. 11 Distribution of community embeddednes s data 75 Subgroup Differences in Embeddedness. Independent sample t-test analyses were utilized to determine if any significant embeddedne ss differences exist among various subgroups of foreign-born faculty based on binary variables s uch as gender, marital status, having school- age children enrolled in local schools, whether or not one™s spouse or domestic partner works outside the home, and homeownership. I conducted an analysis of variance (ANOVA) to exam ine subgroup differences in embeddedness based on categorical variables such as discipline, rank, length of tenure, immigration status, and geographical region of orig in. Findings are reported in the following section. Embeddedness by Gender. Research question 3.1 investigated for difference in embeddedness based on gender. The researcher hypot hesized that female participants are less embedded as compared with the male participants. A n independent-sample t-test was conducted to compare differences by gender in aggregate embed dedness, university embeddedness, and community embeddedness. For aggregate and universi ty embeddedness, no significant difference was observed based on gender (p>.05); bo th male and female respondents demonstrate similar means (x =3.76 and x =3.56, respectively). Figure 4.12 shows that no statistical significant difference was found in the distribution of gender scores (n=119). Significance was set at p < .05. We can see this c learly because the resulting confidence intervals overlap, demonstrating that the mean valu e for each group can be the same. Although males demonstrated slightly higher aggregate embedd edness than females, the means for male and female overlap significantly, which means that no significant statistical difference was observed based on gender for aggregate embeddedness . These findings (see Table 4.8) do not 76 support the stated hypothesis that female responden ts would demonstrate lower levels of aggregate embeddedness than male respondents. Figure 4.12 Embeddedness across gender Table 4.8 T-Tests results comparing male and femal e participants on embeddedness variables Male x Female x t df p Aggregate Embeddedness 3.76 3.56 -1.28 116 .203 University Embeddedness 3.98 3.98 -.076 116 .939 Community Embeddedness 3.53 3.15 -2.06 116 .041* *. Significant at the 0.05 level (2-tailed) The table above also reports findings from follow-u p analysis that was conducted to determine the differences between males and females on their community embeddedness (CE) and university embeddedness (UE) scores. Significance was set at p < .05. Gender was not a Mean Mean±0.95 Conf. Interval MaleFemale Gender 3.20 3.30 3.40 3.50 3.60 3.70 3.80 3.90 4.00 Embeddedness 77 statistically significant factor in the university embeddedness dimension (p=.939). However, the t-test results for community embeddedness showed th at gender was statistically significant (p=.041); males had a statistically significantly h igher community embeddedness than females (x =3.537 and x =3.148, respectively). Follow-up analysis revealed that gender difference was significant at the p<.05 level for the community fi t dimension (p=.021) and community sacrifice (p=.035) of the aggregate construct. These finding s support the stated hypothesis that, at the community embeddedness level, gender was a signific ant factor; females demonstrated a lower embeddedness than males. Embeddedness by Family Status. Research question 3.2 examined difference in embeddedness based on family status. The researche r hypothesized that embeddedness would vary significantly based on family status. Family status is a categorical variable that consists of the following sub variables: (a) married or lived w ith domestic partner, (b) spouse/domestic partner worked outside the home, (c) respondent had K-12 school-age children, and (d) respondent homeownership status. Responses were bi nary (yes / no). Findings for hypothesis testing for each family variable subcategory are re ported in the next section. Embeddedness by Marital Status. Marital status is computed from survey responses regarding whether or not respondents were married o r lived with a domestic partner. Follow-up analysis for marital status was conducted to determ ine intergroup differences for university embeddedness and community embeddedness (significan ce set at p<0.05). The analysis revealed no statistically significant difference for univers ity embeddedness (p=0.536) based on marital status. Similarly, no statistically significant su bgroup difference was observed on community embeddedness based on marital status (p=0.865). Fi ndings did not support the stated hypothesis. 78 As reported in following table, the t-test for mari tal status revealed no statistically significant difference on aggregate embeddedness scores (p=0.81 7) for significance set at p<0.05. Table 4.9 T-Tests results comparing married and si ngle participants on embeddedness variables (n=115) Index Married x (n=102) Single x (n=13) t df p Aggregate Embeddedness 3.74 3.79 -.232 113 .817 University Embeddedness 4.02 4.15 -.620 113 .536 Community Embeddedness 3.47 3.42 .170 113 .865 *. Significant at the 0.05 level (2-tailed) Embeddedness by School-age Children. Research question 3.2 also investigated for statistical significant difference in embeddedness based on having school-age children enrolled in a local K-12 school system. The researcher hypo thesized that respondents who had school- age children would demonstrate higher embeddedness means compared to respondents who did not indicate having school-age children. Table 4.1 0 exhibits t-test analysis of data by school-age children enrolled in a local K-12 school system (n= 113). The Significance level was set at p<0.05. The t-test results showed no statistically significant difference between groups on aggregate embeddedness based on the school-age chil dren variable (p=0.517). Findings did not support the stated hypothesis. 79 Table 4.10 T-Tests results comparing participants™ embeddedness by having/not having school-age children Index Yes Children x (n=52) No Children x (n=61) t df p Aggregate Embeddedness 3.70 3.79 -.649 111 .517 University Embeddedness 3.95 4.11 -1.11 111 .255 Community Embeddedness 3.45 3.47 -.116 111 .907 *. Significant at the 0.05 level (2-tailed) Follow-up analysis was conducted to determine inter group differences based on having children and respondents™ university embeddedness a nd community embeddedness. No statistically significant difference (significance set at p<0.05) was observed on either the university embeddedness (p=0.255) or community embe ddedness (p=0.907) based on having school-age children. Findings did not support the s tated hypothesis in either the aggregate embeddedness or its subcategories. Embeddedness by Working Spouse/Partner . Research question 3.2 also sought to examine intergroup differences in embeddedness base d on spouse/partner working outside the home. The researcher hypothesized that respondents whose spouse or domestic partner worked outside the home would exhibit higher embeddedness than their counterparts. T-test was conducted to determine intergroup difference (n=105 ). Figure 4.13 shows that the distribution of subgroup means overlaps significantly, which means that no statistically significant difference (significance set at p<0.05) was observed on aggreg ate embeddedness based on spouse/partner working outside the home variable (p=0.932). Thus, findings did not support the stated hypothesis. 80 Figure 4.13 Effect of working spouse/partner on em beddedness Follow-up analysis was conducted to determine inter group differences for university embeddedness and community embeddedness based on a spouse/domestic partner working outside the home. Findings are reported in Table 4 .11 (n=105). No statistical difference was observed on either the university embeddedness (p=. 719) or community embeddedness variables (p=.651) based on having a working spouse/domestic partner. Findings did not support the stated hypothesis. Table 4.11 T-Tests results comparing participants™ embeddedness by spouse/domestic partner working outside the home Index Yes x (n=77) No x (n=28) t df p Aggregate Embeddedness 3.75 3.77 -.086 103 .932 University Embeddedness 4.05 3.99 .361 103 .719 Community Embeddedness 3.46 3.55 -.453 103 .651 *. Significant at the 0.05 level (2-tailed) Mean Mean±0.95 Conf. Interval noyes Working Spouse 3.40 3.50 3.60 3.70 3.80 3.90 4.00 4.10 Embeddedness 81 Embeddedness by Homeownership. Respondents were asked if they own the house they live in. Responses were binary variables (1=Y es, 2=No). A majority of respondents (82.6%) reported they owned the house they live in. T-test was conducted to explore for a statistically significant difference in aggregate e mbeddedness scores based on home ownership; a majority of the respondents (82.6%) indicated they owned the house they live in. Statistical analysis for homeownership showed a statistically s ignificant difference was observed at the p<0.05 level for aggregate embeddedness score. How ever, the difference was very small. Findings are reported in Table 4.12 (n=115). Table 4.12 Results of T-Tests comparing participan ts™ embeddedness by homeownership Index Yes x (n=95) No x (n=20) t df p Aggregate Embeddedness 3.81 3.46 1.94 113 .0.55* University Embeddedness 4.07 3.83 1.30 113 .195 Community Embeddedness 3.54 3.10 1.99 113 .049* *. Significant at the 0.05 level (2-tailed) A follow-up analysis was also conducted for impact of homeownership on community embeddedness and university embeddedness (significa nce was set at p<0.05). Results reported in Table 4.12 also show that no statistical signifi cance was observed for homeownership factor on university embeddedness (p=0.195). T-test for i mpact of homeownership on community embeddedness showed statistically significant diffe rence (p=0.049) between groups. Embeddedness by Immigration Status. Research question 3.3 examined the relationship of immigration status on embeddedness. Immigration status was a categorical variable that included the following subcategories: (a) naturalized US citizen, (b) permanent 82 resident or green card holder, and (c) employed on a work visa. The researcher hypothesized that foreign-born faculty members who are naturaliz ed citizens of the United States would have the highest embeddedness scores as compared with fa culty members who are permanent residents and those who are on temporary visas, res pectively. ANOVA was conducted to determine subgroup differences based on immigration status. Figure 4.14 exhibits distribution of aggregate embeddedness means by immigration status (n=120). The data analysis revealed that the sample means overlaps significantly, which sign ifies that no statistically significant difference was observed (p=0.203) among subgroups. Figure 4. 14 Effect of immigration status on embed dedness ANOVA was conducted to determine intergroup differe nce based on immigration status for university embeddedness (UE) as well as for com munity embeddedness (CE) dimensions; results are reported in Table 4.13 (n=120). Signif icance was set at p<0.05. An inspection of mean scores on immigration status revealed no signi ficant statistical difference for university embeddedness (p=0.979) and community embeddedness ( p=0.771). The findings do not support 83 the stated hypothesis. Immigration status was not statistically significant factor in any dimension of the embeddedness variable. Table 4.13 Means and significance of difference fo r immigration status on embeddedness Variable U.S. Citizen (n=62) Permanent Resident (n=43) Work Visa (n=15) F p Aggregate Embeddedness 3.69 3.71 3.79 .098 .907 University Embeddedness 4.01 3.97 4.01 .021 .979 Community Embeddedness 3.38 3.45 3.58 .261 .771 *. Significant at the 0.05 level Embeddedness by Geographical Region of Origin Research question 3.4 explored difference in embedd edness based on geographical region of origin. The hypothesis stated that no significa nt difference would be observed for embeddedness based on geographical region of origin . ANOVA was conducted for aggregate embeddedness based on geographical region of origin . Results are reported in table 4.14 (n=119). Table 4.14 Means and significance of difference fo r origin on embeddedness Variable Africa (n=13) Asia (n=48) Europe (n=39) Americas (n=9) North America (n=10) F p E 3.90 3.55 3.83 4.24 3.28 2.86 .026* U.E. 4.23 3.77 4.14 4.63 3.57 4.04 .004* C.E. 3.56 3.33 3.51 3.85 2.98 1.22 .305 *. Significant at the 0.05 level Origin was found statistically significant at p<0.0 5 level for aggregate embeddedness (p=.026). Follow-up one way ANOVA with Tukey™s HSD (honest significant difference) post hoc test was conducted to determine which groups in the sample differ. Tukey™s HSD test is used with ANOVA tests to compare all possible subgr oup pairs and identify sample means that 84 are statistically significantly different from each other (Cohen & Cohen, 1983). Results also showed that geographical region of origin was stati stically significant at the p<.05 level for the university embeddedness dimension of the construct (p=.004); but not significant for the community embeddedness dimension (p=.305). Analysi s showed that respondents from Central and South America (Americas) demonstrated the highe st mean scores for aggregate embeddedness constructs and its two dimension Œ uni versity and community embeddedness. Embeddedness by Professional Status and Rank. Research question 4 investigated difference in embeddedness based on professional at tributes. Question 2.2.1 explored the impact of rank on aggregate embeddedness (E). Rank is a c ategorical variable that included the following categories: (a) tenure-track assistant pr ofessor; (b) non tenure-track assistant professor; (c) associate professor; (d) professor; and instruc tor. The researcher hypothesized that tenured respondents would demonstrate higher embeddedness t han non-tenured participants. ANOVA tests that were conducted to measure the influence of rank and subgroup differences on aggregate embeddedness (E), university embeddedness (UE), and community embeddedness (CE) (statistical significance was set at p=0.05). Find ings are reported in Table 4.15 (n=120) Table 4.15 Means and significance of difference fo r rank status on embeddedness Variable Assist. Prof. Tenure- track (n=27) Assoc. Prof Tenure- track (n=47) Professor Tenured (n=33) Assist. Prof. no tenure (n=9) Instructor no tenure (n=4) F p E 3.59 3.65 3.92 3.65 3.70 .884 .476 U.E. 4.03 3.89 4.19 3.74 4.00 .932 .448 C.E. 3.16 3.40 3.67 3.55 3.75 1.086 .367 *. Significant at the 0.05 level Findings showed no statistically significant differ ences at the p<.05 level for aggregate embeddedness sample means based on the rank variabl e (p=.476). Similar findings also resulted 85 from follow-up ANOVA for community embeddedness (p= .367) and university embeddedness (p=0.448). Findings did not support the stated hy pothesis. Although no statistically significant difference in embeddedness was observed based on rank, analysis of data showed that the foreign-born faculty members who were ranked as full professors demonstrated slightly higher mean for ag gregate embeddedness in all dimensions of the embeddedness constructs. It was interesting th at non-tenure track respondents at the assistant professor rank demonstrated slightly higher means a s compared with participants who were tenure-track and assistant professor rank. Discipline Differences in Embeddedness. Research question 4.3 explored intergroup difference in aggregate embeddedness scores based o n professional discipline. The researcher hypothesized that no significant difference would b e observed in embeddedness based on professional discipline (significance was set at p < .05). ANOVA was conducted to explore subgroup differences; findings are reported in Tabl e 4. 16 (n=120). An inspection of distribution of means suggests that, based on discipline, there was no statistically significant difference among subgroup embeddedness at the p<.05 level of s ignificance. An inspection of the mean scores reported in Table 4.18 indicates that techno logy had a higher embeddedness score (x =4.09); but, this discipline area also had the fewe st observations in the sample distribution. Table 4.16 Means and significance of difference fo r discipline on embeddedness Variable Sciences (n=63) Technology (n=7) Engineering (n=32) Mathematics (n=9) F p E 3.70 4.09 3.56 3.89 .884 .476 U.E. 3.95 4.20 3.82 4.39 .932 .448 C.E. 3.44 4.00 3.30 3.40 1.08 .367 *. Significant at the 0.05 level 86 Table 4.16 also reports finding from follow-up ANOV A with Tukey™s HSD post hoc tests, which were conducted to determine group difference for un iversity embeddedness (UE) and community embeddedness (CE) based on the respondents™ discipl ine. No statistically significant difference was observed between groups based on discipline eit her on the university embeddedness (p=0.448) or the community embeddedness (p=0.367) d imensions. Findings support the stated hypothesis. Embeddedness by Length of Employment. Majority (47.4%) of the foreign-born faculty in the current study had been employed at t heir current university for ten or more years, followed by 24 percent who were employed for 6-8 ye ars, 18.1 percent employed for 3-5 years, and 10.3 percent that reported 0-2 years of employm ent. ANOVA was conducted to examine intergroup differences in university embeddedness b ased on the number of years foreign born faculty were employed at their current university; findings are reported in Table 4.17 (n=120). Results showed no statistically intergroup differen ce in aggregate embeddedness scores (p=0.733), community embeddedness dimension (p=0.59 4), or university embeddedness dimension (p=0.467). The following table provides ANOVA findings between groups using Tukey™s HSD post hoc test measuring embeddedness an d its two aggregate dimensions (community and university embeddedness) across tenu re. Table 4.17 Means and significance of difference fo r length of employment on embeddedness Variable 0-2 (n=12) 3-5 (n=21) 6-8 (n=28) >10 (n=55) F p E 3.58 3.67 3.75 3.81 .428 .733 U.E. 3.71 4.11 4.03 4.06 .855 .467 C.E. 3.44 4.24 3.47 3.56 .634 .594 *. Significant at the 0.05 level 87 An inspection of results from the table above shows that no statistically significant difference was observed at the p<.05 level for effect of lengt h of tenure on embeddedness. The following figure illustrates these findings; the overlap of d istribution of embeddedness means indicates that there is no statistically significant difference in aggregate embeddedness based on the length of tenure factor (n=120). Findings from the data anal ysis did not support the stated hypothesis for the effect of length of tenure on embeddedness. Figure 4. 15 Effect of length of employment on emb eddedness Embeddedness by Identify as an International Facult y Member Respondents were asked if they identify as an inter national faculty member. Responses were binary variables (1=Yes, 2=No). A majority of respondents (77.3% versus 22.7%) reported they identify as an international faculty member. A T-test was conducted to explore for a statistically significant difference in aggregate e mbeddedness scores based on identifying as an international faculty member; equal variance betwee n the means was assumed. Findings are reported in Table 4.19 (n=119). Mean Mean±0.95 Conf. Interval 0-2> 103-56-8 Tenure 2.80 3.00 3.20 3.40 3.60 3.80 4.00 4.20 Embeddedness 88 Table 4.18 T-Test results comparing embeddedness b y identify as international faculty member Yes x No x t df p Aggregate Embeddedness 3.69 3.80 -.658 117 .512 University Embeddedness 4.02 3.91 .611 117 .543 Community Embeddedness 3.36 3.69 -1.571 117 .119 *. Significant at the 0.05 level (2-tailed) A negative relationship is observed for aggregate e mbeddedness and identifying as an international faculty member; however, the relation ship is not statistically significant at the p<.05 level. A similar finding is observed for the commu nity embeddedness dimension. Although a positive relationship is observed for university em beddedness and identifying as a foreign-born faculty, the relationship is not statistically sign ificant at the p<.05 level. Summary This study sought to examine the embeddedness of pa rticipants who were foreign-born and employed in STEM disciplines at randomly selected c omprehensive universities in the United States. Table 4.20 provides a summary of findings on statistical significance of personal and professional factors on aggregate embeddedness (E), university embeddedness (UE), and community embeddedness (CE). Table 4.19 Summary of factors significance on embe ddedness Factors E (Aggregate) UE CE Personal Factors Gender ns ns * Immigration status ns ns ns School-age children ns ns ns Homeownership * ns * Origin * * ns 89 Table 4.19 (cont™d) Marital status ns ns ns Identify as international ns ns ns Professional Factors Rank ns ns ns Length of tenure ns ns ns Discipline ns ns ns * p < .05; ** p < .01; *** p < .001; ns=no signifi cance observed Findings showed that participants in the current st udy were highly embedded. This finding supported rejecting the null hypothesis and accepti ng the alternative hypothesis. T-tests and ANOVA tests were conducted to test hypotheses for i ntergroup differences in embeddedness based on the following variables: (a) personal char acteristics such as gender, family status, immigration status, geographic region of origin; an d (b) professional attributes such as rank, discipline, and length of tenure at the current uni versity. The impact of homeownership factor on embeddedness variable was also tested. Findings from the current study showed that gender was not a statistically significant factor for aggregate embeddedness; therefore, the r esearcher rejected the stated hypothesis. However, results showed that gender was significant for the community embeddedness dimension of the embeddedness construct; female par ticipants were less embedded in their community than male respondents. For community emb eddedness, findings supported the stated hypothesis. Further, statistical analysis did not show support for the significance of family status, immigration status, rank, and length of ten ure factors on the aggregate embeddedness. However, findings from more in-depth data analysis showed statistically significant difference for geographical region of origin factor on aggrega te embeddedness and university embeddedness. No statistically significant differe nce was observed for geographical region of 90 origin factor on community embeddedness dimension. Table 4.21 provides a summary of hypothesis testes and findings. Finally, identifyi ng as a foreign-born faculty member served as a control variable. A negative relationship was obse rved between aggregate embeddedness and identifying as a foreign-born faculty member, howev er, the relationship was not statistically significant. The variance was not statistically si gnificant for embeddedness means of participants who do or do not identify as internati onal faculty members. Table 4.20 Summary of hypotheses and findings Hypothesis Supported / Rejected 1. Foreign-born faculty members are likely to demonstr ate high embeddedness in the American professoriate. Supported 2. Female foreign-born faculty members are less embedd ed compared with male faculty members. Partially rejected; except for CE 3. Single foreign-born faculty members are less embedd ed compared to faculty members who are married or live with a domestic partner. Rejected 4. Foreign-born faculty whose spouse or domestic partn er works outside the house are likely more embedded as compared with faculty members whose spouse or domestic partner is a homemaker. Rejected 5. Foreign-born faculty members who own their house ar e likely more embedded as compared with the faculty w ho rent. Supported 6. Foreign-born faculty members who have K-12 school- age children are significantly more embedded compar ed with faculty members who do not have school-age children. Rejected 7. Foreign-born faculty members who are naturalized US citizens likely to have the highest embeddedness co res when compared with faculty members who are permanent residents and those who are on temporary visas. Rejected Table 4.21 (cont™d) 7a. Foreign-born faculty members who are permanent residents in the US are likely to have highest embeddedness cores when compared with faculty members who are on a temporary visa. Rejected 8. No significant difference will be observed for embeddedness foreign-born faculty members based on their geographic region of origin. Partially rejected; except for CE 91 9. Foreign-born faculty members with the longest tenur e at the university are likely to be more embedded as compared with faculty with few years of employment. Partially supported, but not statistically significant difference 10. Foreign-born faculty members who are at the full professor rank will likely display significantly hi gher embeddedness as compared with faculty members who are at the associate and assistant professor ranks. Rejected; though full professors did display higher embeddedness mean 11. Foreign-born faculty members who are employed in non-tenure track positions are significantly less embedded in American professoriate as compared with tenure-track and tenured faculty members. Rejected 12. No significant subgroup difference will be observed based on discipline in the embeddedness of foreign- born faculty members. Supported 13. No significant subgroup difference will be observed in aggregate embeddedness based on whether or not participants identify as international faculty memb er. Supported 92 CHAPTER 5: SUMMARY, CONCLUSIONS, AND DISCUSSION Two overarching objectives guided the present study . First, I was interested in understanding what describes the foreign-born facul ty who are employed in the Science, Technology, Engineering, and Mathematics (STEM) dis ciplines at comprehensive universities in the United States; and what factors influence their embeddedness in the American professoriate. This chapter begins with a summary of the impetus f or conducting the current study and then provides an overview of the Embeddedness Theoretica l Framework, followed by a summary of results from the data analysis. The chapter progre sses with a discussion of findings through the lenses of the Embeddedness Theoretical Framework an d literature related to embeddedness. I conclude with a section on implications for researc h and practice. Impetus for Current Study The impetus for the current study was literature th at has demonstrated four important trends in the American professoriate. For example, Schuster and Finkelstein (2006) and Thelin (2004) reported changing demographics in the profes soriate such as the graying of the faculty in American higher education. Other studies (Doyle, 2 008) observed that the number of faculty, ages 55-64 and 71 and above, grew by 2 percent from 1998 to 2004. Further, Sabharwal (2011) observed that fifull-time foreign-born faculty membe rs employed in science and engineering disciplines rose 7 percentage points from 21% in 19 92 to 28% in 2003, with a few disciplines like mathematics, physical sciences, engineering, a nd computer sciences experiencing a 47% growth rate (p. 143-144). This trend is concurrent with large numbers of faculty members in the United States who have been approaching retirement and a limited number of STEM degrees 93 awarded to American-born graduate students in recen t years. (Kuenzi, Matthews, and Magnan, 2006). At the time when this study was conceptualized, mor e than 30 percent of all new faculty hires in the United States were foreign-born (Kim, Wolf-Wendell, & Towmbly, 2011c; Skachkova, 2007). This trend suggests that STEM di sciplines depend significantly on international scholars (Nelson & Rogers, 2005; NSB, 2003). One study (Wei, 2007) at a Research One university reported that foreign-born faculty members made up 35% - 50% of some STEM departments and in some units were fiexce eding 50 percentfl (p. 90). Finally, literature also indicates two other important trend s: (a) increased global competition for international talent (Douglas & Edelstein, 2009; Ha wthorne, 2005; Hercog, 2008; The Economist, 2006; Batalova, 2007; Iredale, 2000); an d (b) an increase in the departure of international scholars from the United States in re cent years (Finn, 2010; Kim et al., 2011a). Such trends are particularly problematic in the STE M disciplines where foreign-born scholars make up a significant percentage of the faculty in the American professoriate. This trend may become a major problem unless foreign-born faculty members are highly embedded in their university and community. In the following paragra phs I provide a brief summary of the Embeddedness Theory and its assumptions. Summary of Embeddedness Theoretical Framework The embeddedness theoretical framework was develope d by Mitchell, Holton, and Lee (2001) as a construct that focuses on why people st ay in their jobs. It accounts for the individual™s appraisal of on-the-job and-off-the-jo b perceived fit, links, and sacrifices (Mitchell et al., 2001). These authors theorized that fit ref ers to the perceived congruency between the individual™s needs and expectations and the provisi ons and accessible amenities provided by the 94 employing organization and their community. In rel ation to the current study, the university™s mission would be an important dimension to a foreig n-born faculty member™s perception of congruency. For example, participants in the curre nt study were from comprehensive universities where greater emphasis is placed on th e teaching mission of the institution as compared with the research focus at the Research On e universities. Participants who are more teaching-oriented than research-oriented would like ly provide a high rating for the fit- university dimension of the embeddedness construct if their cu rrent appointment emphasized teaching responsibilities over research. The links to the organization and the community dimension of the embeddedness construct describes the extent to which an individu al is attached to other people and activities as well as the number of people who depend on her with in the organization and community. Links can be both formal and informal ties. Owning a hom e, having children in the local school system or family members living in the same area, s erve as examples of links to the community. The sacrifice dimension of the embeddedness construct describes t he individual™s perception of what she must give up if she were to leave the organization, and perhaps also the community in which she lives. In other words, if t he employee were to voluntarily leave the organization, she would give up associated benefits such as status and potential promotion, as well as certain perks associated with job seniority . Examples of community sacrifice could mean removing children from their schools, one™s spouse having to resign a professional position and, leaving a support network and an area that feels fa miliar and safe. Mitchell et al. (2001) theorized that the more an employee perceives she w ould have to give up when leaving, the more enmeshed she would feel in the organization and the community. In summary, the degree of 95 embeddedness depends on the faculty member™s apprai sal of on-and-off the job perceived fit, links and sacrifices. For the purpose of this study, embeddedness is unde rstood as a positive construct that explains the degree foreign-born faculty members pe rceive they are enmeshed in their university and community. I collected and analyzed data on re spondents™ perceptions about their fit with their university and community, links to their univ ersity and community, and the level of sacrifices they believed they would make if they we re to leave their university and community. I also gathered personal and professional demographic data. The survey instrument did not collect any institution-related data. Embeddedness was the dependent variable and the demographic data served as independent variables. The demograp hic variables also informed understanding of this sample of foreign-born faculty in STEM disc iplines in the American professoriate. In the next section I provide a summary of the rese arch results, followed by a discussion of findings from descriptive analysis and the stati stically significant factors that were influential on the embeddedness of participants in the present study. Summary of Findings Descriptive analysis of the data showed that an ove rwhelming majority (77.3%) of the participants (n=120) identify as international. As ian and European origins made up the largest percentage of the sample, and only 12.5% were on a work visa while the others were naturalized U.S. citizens (51.7%) and permanent residents (35.8 %), respectively. Females made up only 31.4% and were largely represented in the lower ran ks of the tenure system. Most of the participants (88.7) were married or lived with a do mestic partner; 54% reported having children enrolled in the local K-12 school system; and, 77.3 % reported their spouse/partner worked outside the house. 82.6 percent reported owning th eir place of living. With regards to discipline, 96 52.5% of the respondents were in the sciences, foll owed by 26.7% in engineering, 15% in mathematics, and only 5.8% in technology. A majori ty of the respondents were employed for ten or more years (47.4%) and 6-8 years (24.1%) at the current university. Regarding embeddedness, data analysis showed that a s a group, faculty members who are foreign-born and employed in the STEM disciplines a t comprehensive universities included in this study are highly embedded. The sample mean fo r aggregate embeddedness was 3.714 on the 1 Œ 5 interval (1=very low embeddedness and 5=very high embeddedness). The sample mean for university embeddedness dimension was higher th an the sample mean for community embeddedness (x =3.997 and x =3.430, respectively). The university fit and links sub-dimensions were the highest rated scale items (x =4.375 and x =4.000, respectively). The sacrifices sub- dimension for both university and community dimensi on were the lowest rated scale items (x =3.630 and x =3.084, respectively) in the embeddedness construct . Analysis of the data demonstrated that homeownershi p was a statistically significant factor at the p<.05 level for the aggregate embedde dness and the community embeddedness dimension of the embeddedness construct. In additi on, gender was a statistically significant factor at the p<.05 level only for the community em beddedness dimension of the embeddedness construct. And, geographic region of origin was fo und to be an influential factor at the p<.05 level only for the university embeddedness dimensio n of the embeddedness construct. Data analysis showed no statistically significant differ ence at the p<.05 level for the following variables: rank, length of tenure, and discipline ( professional characteristics), marital status, visa or immigration status, having K-12 school-age child ren, and spouse or partner working outside the home (demographic characteristics). In the nex t section I discuss findings from the 97 descriptive analysis, followed by discussion of res ults from inferential statistical analysis, and then conclude with a section on implications and re commendations for future research. Profile of Foreign-Born Faculty Research question 2 sought to explore what describe s the foreign-born faculty members who are employed in STEM disciplines at comprehensi ve universities. One of the survey questions asked whether or not participants identif y as international faculty members. Data analysis showed that 22.7 percent of the respondent s in the present study indicated they did not identify as international, and 77.3 percent identif ied as international. This finding was somewhat surprising. One plausible explanation for this fin ding is that foreign-born faculty members in the current study have already committed to long-term s ervice to their institution; a majority of them (47.4%) had been employed at their current universi ty for ten or more years, followed by 24.1 percent who were employed for 6-8 years. A caution ary interpretation of this finding is that some of the participants in this study may have bee n born outside the US, but might have been born to expatriate US citizens abroad. In addition , it is very likely that these faculty members identify strongly with the U.S. context as a countr y of immigrants; 51.7 percent of the respondents were naturalized U.S. citizens and 35.8 percent were permanent residents (green card holders). Another important consideration regarding this find ing is the ambiguous connotation of identifying as international . This term seems to resonate differently with the foreign-born faculty, and the survey did not provide a definitio n regarding how international faculty is understood in the current study. Although this stu dy did not explore dimensions of international- related engagement of participants, it is likely th at the faculty members in this study might have considered international to refer to being engaged in international scholarship and service. One 98 important conclusion is that this finding brings to question what it means to be international. According to the 22.7 percent of the participants i n the present study, being international was not perceived as synonymous with being foreign-born. T his finding from the present study also confirms prior research indicating that foreign-bor n faculty members at universities in the United States do not perceive themselves as different from other faculty members (Wei, 2007). This finding draws attention to the preference of foreig n-born faculty members to be noticed for their academic achievements rather than for being foreign -born. The varied ways in which foreign-born faculty self- identify also speaks to the present reality that this subgroup in the professoriate is not homogeneous. Consistent with other studies (Skachkova, 2007), foreign-born faculty in the pres ent study were also diverse; 40.3 percent were of Asian origin, followed by 32.8 percent Euro pean, 10.9 percent African, 8.4 percent North American (Canada and Mexico), and 7.6 percent identified with Central and South America. As noted, a majority of respondents ident ified as Asian. This finding is consistent with Schuster and Finkelstein™s findings (2006) who observed a shift in the source of foreign- born talent coming into the American professoriate from European countries to Asian countries. Included in this demographic change is also the inc rease in the percentage of female faculty among the number of new entrants into the American professoriate. With regards to gender representation, a majority o f the respondents in the present study were males (68.6%) and 31.4 percent were females. Although this finding is consistent with existing literature (National Science Board, 2008) about the gender gap in STEM disciplines, in this study the gender gap was significantly smaller compared to the general STEM faculty gender gap in the American professoriate. Accordin g to a report by the National Science Foundation (2015), the percentage of women made up only 24.3 percent of the full-time faculty 99 in STEM disciplines at all 4-year colleges and univ ersities. In the sciences, engineering and math disciplines female respondents made up 47.6 pe rcent, 45.4 percent, and 50 percent of the participants respectively. For the gender gap in t he technology discipline in the present study (there were 5 times as many male participants compa re to female participants), the results were skewed by the very small sample size in this subgro up. Further, the data analysis also revealed that femal e foreign-born faculty members were better represented at the lower faculty ranks (37% assistant tenure-track professor, 31.9% associate tenured professor, 22.6% professor). Thi s finding is consistent with prior research dating back three decades (Menges and Exum, 1983) a nd more recent literature (NSF, 2015; Hill and associates, 2010; Cropsey, 2008) about represen tation of women faculty in STEM disciplines at the lower faculty ranks. In a recen t study published by the American Association of University Women (AAUW), female faculty members made up less than– fi–one-quarter of the faculty in computer and inform ation sciences (22 percent), math (19 percent), the physical sciences (18 percent), a nd engineering (12 percent). In the life sciences, an area in which many people assume that women have achieved parity, women made up only one-third (34 percent) of the faculty. [And], in all cases women were better represented in lower faculty ranks than in higher r anks– (Hill, Corbett, Rose, 2010, p. 15). However, this finding from the current study also p oints to the growing number of female faculty members who are entering the STEM professoriate. It would be very important that future studies track and report the percentage of foreign- born female STEM faculty members who advance into the higher ranks of the tenure system. For the gender gap in non-tenure-track faculty app ointments, findings from the current study showed that there were twice as many male non -tenure-track assistant professors as females. No gender difference was observed for non -tenure-track instructor rank in the present 100 study. One important observation from the analysis of data in the present study is the percentage of tenured female foreign-born professors compared to national data sources (Hill et al., 2010; Corpsey, 2008). Findings from the current study show a different tr end for foreign-born female faculty members in the STEM disciplines. The percentage of female full professors in the current study was 3.6 times larger than the national percentage r eported by Burrelli (2008). A similar difference is the percentage of female assistant pr ofessors; 37% in this study versus 42% in the Burrelli study. At the associate professor rank, f indings from this study showed that female foreign-born STEM faculty members were 2 percent be hind the national average. It is important to note here that the sample for the current study is not representative of STEM faculty in general. Participants in the current study were fr om comprehensive universities, and these institutions may have a somewhat distinct faculty p rofile, especially in STEM disciplines, as compared with faculty at research universities. With regards to marital status, descriptive analysi s of data from the present study showed that 85 percent of respondents were married or live d with a domestic partner, 10.8 percent were single, and 4.2 percent skipped this question. A r elatively similar percentage of males (90.9%) and females (83.3%) were married or lived with a do mestic partner. However, findings from the present study showed that the percentage of single female foreign-born faculty members was nearly twice as large as the percentage of single m ale participants. This finding might be explained by inequities in the workplace and pressu res women are reported to experience in STEM departments such as discrimination due to gend er and work-family imbalances, and mentoring inaccessibility (Blackwell, Snyder, Mavri plis, 2009). 101 The percentage of female respondents who had K-12 s chool-age children was the same as the percentage of participants who did not report h aving school-age children. It is important to observe that the number female respondents who repo rted not having school-age children also correlates with these respondents being early-caree r faculty members; a large percentage of female respondents were at the assistant professor rank in this study™s dataset. Other research studies ( Armenti, 2000 ) have documented the high pressures female faculty members feel to balance academic productivity with childrearing (Wa rd and Wolfe-Wendel, 2004; Finkel and Olswang, 1996). A study by Mason and Goulde (2004) reported that having children and the fitiming of babiesfl was an important factor in the g ap between men and women faculty achieving tenure. Some women faculty members either delay ha ving babies or opt not to have children altogether. An earlier study by Davis and Astin (1 990) reported that, compared to male faculty fionly women identify family commitments as obstacle s to publication and productivityfl (as cited in Armenti, 2010, p.66). Similar to US-born female faculty, foreign-born female faculty members are probably not immune to pressures for re search and publication in order to become tenured. For male respondents, 56 percent reported they did not have school-age children. It is important to observe that nearly 30 percent of male respondents were also full professors, and 50 percent of males in the subgroup had been employed at the university for more than ten years. This could mean that the 56 percent of male respond ents in this study who said they did not have school-age children might have had children, but, a t the time these respondents participated in the survey their children might have been older tha n the K-12 age group. With regards to visa or immigration status, a major ity of the respondents were naturalized US citizens (52.2%), followed by permanent resident s (34.8%); only 13 percent were on a work 102 visa. It is not surprising that a majority (87%) o f foreign-born faculty members in the present study reported having a permanent immigration statu s (naturalized citizens and permanent residents combined). This finding is consistent wi th findings from the Kim et al. (2001c) study that reported 64 percent of foreign-born faculty me mbers were naturalized US citizens. This finding is important for two reasons. First, the l ow percentage of participants holding a work visa might support Finn™s (2010) findings that fewe r foreign-born scholars are opting to stay in the US after completion of their doctoral education at American universities. The researcher expected a higher number of respondents on work vis as, considering recent immigration legislation reform that has increased the number of visas issued to highly skilled workers such as scholars in STEM disciplines (Stratford, 2012). Further, the combined 87 percent of the sample who identified as either US citizens or on permanent resident status suggests that these parti cipants have strong links, and according to the embeddedness theory, they are less likely to leave the American professoriate for figreener pasturesfl in other countries. Said differently, a foreign-born professor holding permanent resident status (also known as a green card) would sacrifice a lot if opting to live outside the US for more than 6 months during a calendar year; that would be considered having abandoned permanent residency in the United States. Permanen t resident status also suggests that the individual has established other strong links to th e university and the community, which leads to higher embeddedness. As it has been the profession al experience of the researcher, many foreign-born professors obtain their permanent resi dent status in the United States through the university sponsoring the change from non-immigrant visa status to permanent resident status. For the visa or immigration status variable, the re searcher expected the embeddedness scores would vary significantly based on respondent s™ immigration status. Other research (Kim, 103 Wolf-Wendell, & Towmbly, 2011c), has indicated that foreign-born faculty mobility is constrained by their immigration or visa regulation s that tie the immigrant to the sponsoring institution, discouraging or inhibiting voluntary t urnover. The current study assumed a positive and significant relationship among immigration stat us and embeddedness of international faculty. I expected that respondents who were on a work visa status would demonstrate a lower embeddedness score. Findings from the analysis of data did not support the stated hypothesis. No statistical significant difference was observed at the p<.05 level for embeddedness of participants who were naturalized US citizens as co mpared with foreign-born faculty members on a temporary work visa. However, it is also impo rtant to acknowledge that foreign-born faculty members in the current study who were on a work visa would have a lot to sacrifice if they left their university because their authorizat ion to work in the United States is tied to the employer who sponsored their work visa. A majority of the respondents in this category (73%), most of whom were males (62.5%) indicated they woul d sacrifice a lot if they left their job. In the next section I discuss findings from inferen tial statistical analysis examining the relationship of personal and professional character istics (independent variables) to the embeddedness (dependent variable) of foreign-born f aculty members in this study. I interpret findings through the Embeddedness Theoretical Frame work and compare results from the present study with prior research that has measured the embeddedness of faculty in the American professoriate. The discussion of findings is organ ized into two sections: (a) embeddedness of foreign-born faculty, and (b) effect of demographic variables on embeddedness. Embeddedness of Foreign-Born Faculty 104 The leading research question sought to examine how embedded are the foreign-born faculty members who are employed in STEM discipline s at randomly selected comprehensive 9 public universities in the United States. And the subset research questions investigated the effect of personal and professional characteristics on emb eddedness of respondents in the current study. The null hypothesis for the current study stated th at foreign-born faculty members are not highly embedded. The alternative hypothesis stated that f oreign-born faculty members are highly embedded. Chapter 4 reported results from the stat istical analysis of responses from 120 respondents who participated in the online survey. Participants™ embeddedness was measured using a six-item scale that was developed by Cunnin gham, Fink, and Sagas (2005); responses were recorded on a 5-point Likert scale. A mean sc ore was calculated for each respondent to generate an aggregate embeddedness score, and the m ean of means was calculated to generate the embeddedness score for the sample in this study . Results are presented in Table 4.1. The composite embeddedness score was expected to fall o n an interval between 1 and 5, where 1 represents very low embeddedness; and 5 represents very high embeddedness. Analysis of the data showed that, as a group, the f oreign-born faculty members in this study are highly embedded; the overall embeddedness mean score was 3.71 and standard deviation was .783. Based on findings from the cur rent study, the researcher rejected the null hypothesis and accepted the alternative hypothesis Œ the foreign-born STEM faculty members who participated in this study were highly embedded . Later in this chapter I discuss factors that were found to influence the embeddedness variable. At the time the current study was undertaken there were few other embeddedness studies that examined embeddedness of faculty at higher edu cation institutions. Biddle (2010) examined 9 Comprehensive universities refer to Carnegie Maste r™s classification of universities. 105 embeddedness of faculty at Christian higher educati on institutions using a 5-point Likert scale and reported low embeddedness scores (2.686 for agg regate embeddedness; 2.591 university embeddedness; and 2.780 for community embeddedness) . Hardcastle (2010) also examined embeddedness of university faculty at Christian ins titutions using a 7-point Likert scale and reported a fi–mild degree of [aggregate] embeddednes sfl (p. 89). Gaylor (2006) used an 8-point Likert scale to compare the embeddedness of busines s faculty at secular and Christian-affiliated higher education institutions and colleges and repo rted a mean score of 5.12 for job embeddedness. Wei (2007) examined the attachment o f foreign-born faculty members at a large Research One university and found that participants in that study fidemonstrated low and passive attachment to their employing institutionfl (p. 99). Contrary to findings from these other studies, the analysis of data from the current study showed that foreign-born faculty members demonstrated high embeddedness ( x=3.714 on a 5-point Likert scale). However, some v ariance was observed between university and community embed dedness dimensions of the construct. In the next section I discuss findings relevant to the university and community embeddedness of foreign-born faculty members in the current study. University and Community Embeddedness According to the data analysis, a higher embeddedne ss mean score was observed for the university dimension ( x=3.997) compared to the community dimension ( x=3.430) of the embeddedness construct. This finding was not surpr ising because, in general, faculty tend to identify strongly with their academic community. I n a recent study about foreign born faculty members™ institutional attachment (Wei, 2007), part icipants fi–spent much of their time in labs and libraries. They either show no interest or have no access to resources in the local community lives and other cultural valuesfl (p. 69). 106 One might argue that lower community embeddedness, as compared with university embeddedness, can be generally expected in light of these faculty members™ recent arrivals in their new community. However, analysis of data for length of employment variable in the current study showed that 45.8 percent of responden ts reported ten or more years of employment at their university. According to the embeddedness theory, lower community embeddedness is disconcerting because, when controlling for length of employment, weak community embeddedness implies the likelihood of voluntary tu rnover (Cunningham et al., 2005; Mitchell et al., 2001). Further, female foreign-born faculty m embers were found to be less embedded in their community than male foreign-born faculty. Th is finding should concern university administrators regarding retention of female foreig n-born faculty members. The effect of the gender variable is further discussed later in this chapter. Interestingly, the highest rated embeddedness dimen sion in the current study was perceived congruency (fit) with the university ( x=4.375). This finding is both surprising and encouraging, particularly in light of findings from the Wei (2007) study where foreign-born faculty members perceived they did not belong in Am erican academe. Wei reported that the cultural differences foreign-born faculty members e ncountered made them feel alienated. Other studies (Chatman, 1989) have suggested that organiz ational fit mediates the employee™s socialization experiences, productivity, and job sa tisfaction. University administrators would find it encouraging that results from the current s tudy showed that foreign-born faculty members appraised highly their organizational fit with thei r university. Olsen and associates (1995) cautioned, however, that organizational fit is more complex than an employee™s perception; it fimay be confounded by misperceptions of the individ ual™s professional values and strengthsfl (p. 272). 107 In the current study, the lowest rated embeddedness dimension was community sacrifices (x=3.084). This finding, although not seriously low on a 1-5 scale (1=very low embeddedness and 5=very high embeddedness), should alert univers ity administrators and policy makers because a low perception of community sacrifice is a mobility indicator; the less an individual perceives she has to lose by relocating, the less e mbedded the person is (Mitchell et al., 2001). According to the Embeddedness Theory, the community sacrifice dimension of the embeddedness construct measures what the individual would lose by leaving the community and relocating elsewhere. For example, if participants in the current study who identified as permanent residents were to relocate abroad, they w ould risk losing their green card and authorization to work in the United States. Other examples of community sacrifice could mean removing children from their schools, one™s spouse having to also resign from a job, leaving a support network and an area that feels familiar and safe. Despite their higher perceptions about university fit and links ( x=4.375 and x=4.000 respectively), it was surprising that the sample me an score for university sacrifice ( x=3.630) was lower than the other two sub-dimensions (fit and li nks). Cunningham and associates (2005) postulated that fi–organizational sacrifice might be the most important facet [dimension] of embeddedness [construct] – and that other facets mi ght, in combination, lead to overall embeddednessfl (p. 231). According to the Embeddedn ess theory, a low mean score for organization sacrifices suggests respondents percei ve that they could acquire similar benefits by relocation to another job. Contrary to findings fr om the Cunningham and associates™ (2005) study, in the current study sacrifices (community s acrifices, in particular) were a less important facet in the embeddedness construct as compared wit h fit and link dimensions. 108 At the outset of the present study, the researcher expected higher community sacrifice scores when considering factors such as homeownersh ip, having children enrolled in the local K- 12 school system, and also having a spouse/domestic partner working outside the home. Analysis of data in the current study revealed that a majority of the respondents (82.6%) owned the house in which they live, 85 percent were marri ed or live with a domestic partner Œ 76% of whom reportedly worked outside the home, and only 4 5.9 percent of the respondents had children enrolled in the local K-12 school system. According to the embeddedness theory, these finding suggest that foreign-born STEM faculty memb ers who participated in the current study are highly embedded and likely to stay. Results fr om an independent sample t-test indicated that homeownership was a moderately significant factor ( p=.055) in the aggregate embeddedness of foreign-born faculty members and a statistically si gnificant factor at the p<.05 level for community embeddedness (p=.049). This finding is co nsistent with other research where owning and selling a house could be factor in opting to st ay in the community and keep the job at the university (Hardcastle, 2010). Mitchell and associates (2001) theorized that havin g school-age children is an important factor for the community embeddedness dimension and aggregate embeddedness construct. Other studies (Giosan, 2004) reported that the numb er of school-age children was a significant factor in overall embeddedness of respondents. In the current study, having school-age children was not a statistically significant factor for the embeddedness of participants. A lower community sacrifice score (x =3.084) also suggests that foreign-born faculty members did not expect it would be very costly to t hem (high sacrifice) if they left the community in which they live. This finding is cons istent with results from the Hardcastle (2010) study of the embeddedness of faculty members at Eva ngelical colleges and universities. 109 Hardcastle observed that scale items measuring comm unity sacrifice might not be clear enough for respondents to appraise associated costs with l eaving the community. According to the embeddedness theory, a low mean for community sacri fice suggests a bigger challenge for university administrators because there is little t hat a university can do to influence the congruency perceptions that foreign-born faculty me mbers have regarding the community in which they live. However, there are practical stra tegies university administrators may employ in increasing the links of foreign-born faculty to the community. These are discussed in the section on implications for policy and practice. In summary, findings from the current study showed that foreign-born faculty members who participated in the current study are highly em bedded (x =3.714). Based on the data analysis, respondents demonstrated high university embeddedness (x =3.997) and moderate community embeddedness (x =3.430). Participants appraised very highly their university-fit (x =4.375) and university-links (x =4.000). Although the university and community sac rifice dimensions were rated moderately high and average ( x=3.630 and x =3.084) respectively, according to the embeddedness theory, participants in the current study would be expected to stay at their universities and in their communities . In the next section I discuss findings based on the relationship of key demographic factors to the respondents™ embeddedness. Relationship of Demographic Factors to Embeddedness Research questions 3 and 4 (and subquestions) explo red differences in embeddedness based on personal and professional characteristics, respectively. Personal characteristics included variables such as gender, marital status, having school-age children, immigration status, and geographical region of origin. Professional va riables included discipline, length of tenure, and rank. At the time the current study was undert aken, existing embeddedness studies 110 associated personal attributes (i.e., gender, famil y status) and professional characteristics (i.e., length of employment, rank) with on-the-job and-off -the-job embeddedness. For example, Hardcasle (2010) studied institutional commitment o f faculty members at evangelical institutions and found that spouse working outside the home, hom eownership, and tenure were positively and significantly correlated with aggregate embedde dness and institutional commitment. Similarly, Gaylor (2004) studied organizational com mitment and embeddedness of faculty members at public and private (Christian-based) col leges and universities, and found that employees remain with an organization because they are linked with the community in which they live or believe they are a fit within the orga nization. Other studies (Smart 1990; Olsen, 1993; Johnsrud and Heck, 1994) have demonstrated th at demographic variables and perceptual variables about work-life in the institution were s ignificant in differentiating between those faculty members who stayed and those who left the o rganization. The current study examined the relationship between the background characteristics (personal and professional) of foreign-born STEM fa culty members and their embeddedness in their universities and communities. Based on the d ata analysis from this study, three personal characteristics were statistically significant at p <.05 level: (a) gender, (b) region of origin, and (c) homeownership factors for embeddedness of forei gn-born faculty members in their university and community. Other personal and professional cha racteristics were not found to be statistically significant at the p<.05 level. In t he next section, I discuss the significance of gend er in intergroup embeddedness, followed by a discussio n on homeownership and geographical region of origin. Relationship of Gender on Embeddedness. Research question 3.1 investigated for differences in embeddedness based on gender. Other studies have suggested that fiwomen are 111 twice as likely as men are to quit their jobfl (Schw artz, 1989 as cited in Holtom and Inderrieden, 2006, p. 443). The present study hypothesized that female foreign-born faculty members are less embedded as compared with male foreign-born faculty . Findings from the current study did not support hypothesis 2 at the aggregate embeddedness level; female foreign-born faculty demonstrated slightly lower aggregate embeddedness scores as compared with male faculty, but the difference was not statistically significant (p =.203) at the p<.05 level. This finding was consistent with other research studies (Sanders, 20 15; Biddle, 2010) where male and female faculty members demonstrated almost identical aggre gate embeddedness scores. However, in the current study, there was a statistically signif icant difference at the p<.05 level for community embeddedness (p=.041), and particularly for the fit and sacrifices dimensions with p values reported at .021 and .035, respectively. Although findings from the present study indicate t hat female foreign-born faculty members are embedded in their university, a statist ically significant lower community embeddedness finding raises the question whether or not university embeddedness alone is strong enough for foreign-born female faculty membe rs to stay. Prior research has also demonstrated that employees™ decision to stay or le ave is affected by non-work influences such as having a spouse and children (Lee and Maurer 199 9) and prioritizing family needs (Maertz & Campion, 1998; Steers and Mowday, 1981; Mobley, 197 9). In the current study, a majority of the female faculty members reported they were marri ed or live with a domestic partner, have children, and their spouse worked outside the home; which suggest they are likely to stay. Interestingly, no statistically significant differe nce was observed for the gender factor on any of the three dimensions of university embeddedn ess. This finding is consistent with findings from other studies (Biddle, 2010; Holtom and O™Neil , 2004). Other research (Xu, 2008) has 112 demonstrated that fiwomen and men are equally commit ted to their academic careers in STEMfl (p. 607). The same study went on to document that workplace related factors were a greater influence on their turnover intentions rather than family responsibilities. The low means for community embeddedness of female foreign-born faculty members in the current study also correlate with the finding t hat a majority of the female participants were represented at the lower academic ranks (assistant and associate professor) and the length of tenure. Prior literature (Armenti, 2010; Mason and Goulde, 2004; Armenti, 2003; Austin, 2002; Fox, 2001; Johnsrud and Heck, 1994) has documented work-life balance issues for early-career female faculty. For example, Armenti (2004) report ed that female faculty members fibelieve that having children before obtaining tenure is detrimen tal to their career prospectsfl (p. 212); some of them wait until they obtain tenure before having ch ildren. According to the embeddedness theory, the female foreign-born faculty members in the present study may have not developed the type of community links that Mitchell and assoc iates (2001) would consider very costly to break. Homeownership Effect on Embeddedness. Research question 3.2 investigated for difference in embeddedness based on family status. Family status was a categorical variable that included binary variables such as single versus mar ried/live with domestic partner, have/not have children enrolled in local K-12 school system, and homeownership. For the family status variable, the analysis of data showed that only hom eownership was moderately significant at the p<.05 level for the aggregate embeddedness (p=.055) of foreign-born faculty members in this study, and statistically significant for the commun ity embeddedness dimension of the construct (p=.049). The researcher hypothesized that foreign -born faculty members who own their house are likely more embedded as compared with the facul ty who do not. Results from the data 113 analysis moderately supported the stated hypothesis that embeddedness would vary significantly based on homeownership status. One important question among employers is whether f eeling embedded causes one to purchase their home or vice versa. According to th e Embeddedness Theory, homeownership causes one to feel more embedded by way of creating stronger links in the community and having more to sacrifice if contemplating leaving ( accounting for housing market conditions, actual and sentimental costs associated with sellin g a house); and, embeddedness does not cause one to buy a house (Mitchell et al. 2001). Whether or not this assumption applies the same way in a higher education setting (where there is a ten ure system) as compared to other settings, is an important consideration and beyond the scope of the current study. It is plausible to expect that faculty members who are tenured are more likely to also be homeowners; therefore, a positive and significant correlation would be expected betwe en the homeownership variable and other variables such as faculty rank, years of employment at the university, and length of time living in the community. On the contrary, however, correlati on analysis of the data from the current study showed a negative and statistically significant rel ationship between homeownership and faculty rank, as well as homeownership and length of employ ment at the university. Prior research (Hardcastle, 2010) that measured emb eddedness of higher education faculty observed that homeownership was significant in the embeddedness of the participants in that study. Hardcastle concluded that faculty memb ers perceived selling their house to be costly Œ a factor that influences their community embedded ness. Findings from the current study showed that homeownership was only a moderately sta tistically significant predictor variable for embeddedness of foreign-born STEM faculty members. 114 According to the embeddedness theory, this finding is very significant because homeownership increases community links and is high ly correlated with the community sacrifices dimensions of the embeddedness construct (Cunningham et al., 2005; Mitchell et al., 2001). As Hardcastle (2010) observed, selling a ho use and relocating can be an arduous endeavor and financially very costly for faculty me mbers. 82.6 percent of the respondents in the current study indicated they owned the house in whi ch they live. This finding may suggest positive news to university administrators and poli cymakers because it implies that many foreign-born faculty members may have made a long-t erm commitment to their university and the community. In addition, it is also important to observe that a majority of the other 17.4 percent of the respondents who did not own their home reported the y were on a work visa. The 2-tailed correlation analysis test of the data from the curr ent study showed a positive, and statistically significant relationship (p=.002) between homeowner ship and immigration status. This finding was expected when considering the temporary nature of work visas, difficulties that non- immigrants encounter in obtaining a mortgage loan w ithout an established credit history, and the long-term financial commitment in homeownership. L ending practices for non-immigrants vary from one financial institution to another. At a ti me when the housing market remains volatile in the United States, and there are increasingly stric ter regulations for mortgage lending, homeownership for foreign-born faculty members on t emporary (non-immigrant) work visas is expected to remain an important factor for the comm unity embeddedness of this subgroup in the American professoriate. Origin. Research question 3.4 investigated for difference in respondents™ embeddedness based on their geographical region of origin. Orig in was a categorical variable that included five 115 regions: Africa, Asia, Europe, Central and South Am erica, and North America. The researcher hypothesized that no significant difference would b e observed on embeddedness of foreign-born faculty members based on their geographical region of origin. However, the data analysis showed that origin was statistically significant at the p<.05 level for aggregate embeddedness (p=.026), and for university embeddedness (p=.004). The origin factor was statistically significant for each of the three factors/dimension s (fit, links, sacrifices) that measured university embeddedness (p=.013, p=.038, and p=.026 respective ly). No statistically significant difference was observed for community embeddedness based on or igin (p=.305). These findings did not support the stated hypothesis, and the researcher a ccepted the alternative hypothesis. Region of origin does matter in how foreign-born faculty memb ers in the present study appraise their perceived fit and links in their university. Data analysis also revealed lower means for commu nity embeddedness as compared with university embeddedness means for each subgrou p. Respondents who identified with fiNorth Americafl demonstrated the lowest university and community embeddedness means in the sample (x =2.98, and x =3.57, respectively). This subgroup did not percei ve they would sacrifice a lot if they left their community (x =2.50) and their university (x =3.10). One plausible explanation for this finding is that foreign-born f aculty members who identified as North American may find it easier to relocate than other participants in this study. This finding should alert university administrators and policy makers b ecause according to the embeddedness theory, low means for community sacrifice, coupled with low means for organizational sacrifices, serve as indicators for potential voluntary turnover (Mit chell et al., 2001). Voluntary turnover is a utility of at least one of the following three poss ible considerations: (a) unmet expectations (Porter and Steers, 1973); (b) life changing events (i.e., childbirth, disability, illness, retirement , 116 etc.), which Mitchell and associates (2001) conside red as shock events; and (c) perceived better rewarding employment and quality of life elsewhere (March and Simon, 1958). Some of the findings in the current study seem coun terintuitive and contrary to theoretical assumptions of Embeddedness Theory. Fo r example, one would expect that faculty members who are tenured would be less likely to lea ve, partially because, in the American higher education system, tenured status is often understoo d as lifelong employment (satisfactory performance is assumed) at the same university. Ac cording to the Embeddedness Theory, tenured status would be associated with higher embe ddedness, and thus a tenured professor would less likely to leave the organization. One p lausible explanation for the lack of statistical significance of the tenured (rank) variable in embe ddedness of foreign-born STEM faculty members is that these faculty members might perceiv e they can obtain similar or better rewarding employment and quality of life elsewhere. Descriptive statistics from the current study showed that some of the participants were ten ured full-professors and employed for as few as 0-2 and 3-5 years at their current university. This finding should alert university administrators to the mobility of foreign-born STEM faculty. Implications for Theory and Research The present research is one of very few studies tha t have utilized the embeddedness theoretical framework for research in the higher ed ucation context. This is the only known study that has investigated the embeddedness of for eign-born faculty members in the American professoriate. In addition, this was also the firs t study to utilize Cunningham and associates™ (20015) single-item scale to measure the six dimens ions of the embeddedness construct for this sample. The embeddedness theory focuses on why ind ividuals stay in their jobs and accounts for factors that go beyond job satisfaction. Mitchell and associates (2001) theorized that individuals 117 stay because they perceive the sacrifices they woul d make by leaving their university and their community outweigh the benefits of relocating elsew here. Cunningham et al. (2005) argued that sacrifices is the important dimension of the embeddedness con struct. In contrast, findings from the current study demonstrated that, for this sampl e, the sacrifices dimension in the embeddedness construct, especially community sacrif ices, was the least significant dimension. Future studies should further investigate the signi ficance of the sacrifices dimension of embeddedness to higher education faculty. Research studies utilizing regression analysis would provide a better understanding about what factors c ontribute the most to the embeddedness of foreign-born STEM faculty members at US universitie s. Future studies should also take a comparative appro ach in collecting and analyzing data on the embeddedness of US-born and foreign-born fac ulty members who are employed in STEM disciplines at universities across the United State s. Such an approach would ensure a broader understanding of faculty embeddedness in STEM disci plines. A comparative study would also inform the reader whether being foreign-born or US- born is a significant factor in embeddedness of faculty members in the American professoriate. It is plausible that embeddedness may mean something different for foreign-born faculty member s compared to domestic faculty members. The present study was more a cross-sectional explor atory analysis of the embeddedness of foreign-born faculty; further study is needed on the effects of other institutional and community factors on participants™ embeddedness. F or example, findings from the current study demonstrated that female faculty members were signi ficantly less embedded in the community as compared with the male respondents. Future studies might include follow-up questions that investigate reasons why participants provided a cer tain rating for the single scale items that measure community and university embeddedness. 118 It would also be a valuable contribution to the lit erature if future studies collected and analyzed more in-depth data on the quality of respo ndent™s experiences at their university and in their community beyond what was measured in this st udy by the single scale items. For example, collecting and analyzing institutional dat a such as financial pressures, student enrollment and retention, workplace climate, and fa culty workload issues could shed light into whether a foreign-born faculty member might perceiv e having a future at their current university. Community related data, such rural versus urban, co uld inform findings about embeddedness of foreign-born faculty members when considering their own socio-economic background and lifestyle preferences. Factors such as job satisfa ction and workplace climate have been demonstrated by prior literature (Cunningham et al. , 2005; Mallol, 2007; Mitchell et al., 2001) as antecedents of job search behavior (updating vitae, conducting job searches, and participating in job interviews). The present study did not inquire about job satisfaction and workplace climate. One interesting finding from the present study was the number of participants who reported being foreign-born but did not identify as international faculty members. Future research will need to investigate what it means to be an international faculty member. The present research did not inquire about participants ™ engagement in international activities. Future studies might investigate the relationship between being foreign-born, identifying as international faculty member, and engagement abroad as well as in internationalization activities at the employing university. Another recommendation for future research is the n eed to better understand the embeddedness of foreign-born faculty and their role s in the university and community. The present study did not inquire about the many roles that describe faculty work such as administrative functions (i.e., department chair, p rogram director, dean of the school), researcher, 119 teaching, and service (i.e., committee work, etc.). Faculty roles might influence faculty embeddedness at the university. For example, the u niversity-links dimension of the embeddedness construct might be influenced by the m any interactions that the faculty member might have as a program director, department chair, or school dean. A final recommendation for future research is to ex amine embeddedness as an independent variable. Data collection and analysis should test for strength and direction of relationship as well as measure the effect of the e mbeddedness factor on outcome variables such as retention, productivity, satisfaction, and organ izational commitment of foreign-born faculty at comprehensive universities in the United States. A comparative approach to such as study is recommended. Implications for Policy and Practice The present study examined the embeddedness of fore ign-born faculty members who were foreign-born and employed in STEM disciples at comprehensive universities in the United States. Findings from the present study have sever al implications for policy and practice as they inform recruitment and retention of highly skilled scholars. Duderstadt (2000) posited that the fiquality and achievements of [faculty]–determine th e quality of the institutionfl (p. 54). For several decades, American universities have success fully attracted growing numbers of foreign- born faculty members, especially in the STEM discip lines, and the question has been whether these faculty members will stay. Data analysis demonstrated that, as a group, foreig n-born faculty members who participated in the present study are highly embedd ed. Findings also revealed that these faculty members are highly embedded in their universities a nd less embedded in their communities. On one hand, this is good news, because high universit y links increase organizational attachment 120 (Mossholder, Setton and Henagan 2005; Friedman and Holtom 2002). On the other hand, low community embeddedness may be cause for concern. S pecifically, the low means for community sacrifice and prior literature on faculty voluntary departure caution interpretation of findings for university embeddedness in the current study. Brown (1967) observed that faculty mobility is sometimes to be expected in higher educ ation fi–because loyalty to discipline transcends loyalty to school and because teaching-r esearch skills are readily transferable among schoolsfl (p. 25). One conclusion from the findings in this study is that the high university embeddedness sample mean is likely to demonstrate m ore accurately the embeddedness of foreign-born faculty members in their discipline an d the university might be perceived as a proxy (to carryout scholarly activities) rather than an a nchor (hindering voluntary turnover). Future studies might seek to investigate this possibility. The lack of statistical significance of tenured sta tus, as measured by the rank variable, on the aggregate embeddedness and university embeddedn ess should alert university administrators and policymakers that tenured status may not mean l ife-long employment at the same university for foreign-born STEM faculty members. For several decades many foreign-born scholars have pursued careers in the American professoriate for d ifferent reasons, which may include limited opportunities for faculty jobs in their own countri es, political unrest and economic reasons. However, the reasons why foreign-born STEM faculty members chose to pursue careers in the American professoriate, and also stay at their univ ersity, may be different than those assumed by administrators, policymakers, and the tenure system in American higher education. Future studies should seek to investigate such reasons amo ng foreign-born STEM faculty and test some of the assumptions associated with the tenure syste m in American higher education. 121 University administrators and policy makers should pay close attention to the lower embeddedness in the community. Gender and homeowne rship were significant factors for embeddedness of foreign-born faculty in the present study. With regards to gender™s relationship to community embeddedness, female participants in t he current study indicated they did not perceive the community in which they live was congr uent (a good fit). Taking into consideration that a majority of the female faculty members in th is study were represented in the lower faculty ranks, it is likely that these faculty members are new to the area where they live, may feel isolated, and lack a sense of community outside the ir workplace. Promoting and facilitating community socials and engagement opportunities has the potential to help them feel connected and belonging. Universities could also increase co mmunity embeddedness and by providing their faculty with information and support for comm unity and social engagement (Holtom and Inderrieden, 2006; Mitchell et al., 2001). One other recommendation for practice includes lear ning more about the community needs and interests of female foreign-born faculty members and finding ways to meet those needs and interests. For example, data from the pr esent study suggests that these faculty members are mostly represented in their early and m id-career stages and might need adequate support balancing their faculty roles (teaching, re search, and service) and childrearing. Encouraging professional engagement in community-ba sed professional organizations (Holtom and O™Neil, 2004) and providing support with childc are has the potential to facilitate a sense of community building and appreciation. Findings from the current study also demonstrated t hat many foreign-born faculty members did not perceive they would have a lot to s acrifice if they were to leave their university or community. This finding has implications for fa culty retention, especially at a time when 122 institutions and governments across the world have promoted greater internationalization and the global competition for international talent has int ensified (Douglas & Edelstein, 2009; Hawthorne, 2005; Hercog, 2008; The Economist, 2006; Batalova, 2007; Iredale, 2000). A very low sense of community sacrifice is a function of w eak community links. Data analysis also demonstrated that homeownership was a significant f actor in embeddedness of foreign-born faculty in the present study. Another recommendation for practice to strengthen c ommunity links and increase the appraisal of sacrifice-community is to promote home -buying (Holtom and O™Neil, 2004). Policymakers need also to consider that foreign-bor n faculty members, especially those employed on a work-visa, may not have established t he level of credit history mortgage lenders often look for in order to approve applications for loans. At a time when the housing market in the United States remains volatile and there are in creasingly stricter regulations for mortgage lending, lenders are reluctant to approve mortgage loans for customers who are on a temporary non-immigration work visa. Universities are in a position to mitigate this bar rier for foreign-born faculty members in several ways. For example, some universities that own houses may sell houses to faculty members and guarantee buying them back when employe es want to sell them or move elsewhere; such is the case at Lynchburg College in Virginia. Other universities offer financial assistance for home-buying. To support faculty and staff retention through encouraging homeownership, several universities in the United S tates have instituted their own financial assistance programs to encourage homeownership. Fo r example, according to the Human 123 Resources webpage 10 of Drexel University, employment benefits includes the Home Purchase Assistance Program that was designed to provide eli gible staff members a forgivable loan towards the purchase of a home. Other universities (e.g., Penn State University, University of Washington) offer their employees special financing and interest rates negotiated with third- party mortgage lenders. Finally, based upon findings of the present study, we know that being foreign-born is not synonymous with seeing oneself as international. O ver twenty percent of the respondents in this study indicated that they do not identify as an int ernational faculty member. This finding has implications regarding references in university int ernationalization literature and comments at social settings that emphasize the fiothernessfl of f aculty members. Wei (2007) reported that foreign-born faculty members in her study preferred to be known for their academic and scholarly accomplishments instead of being singled out because they were born in a country different than the United States. This may be a co ntrary position to some who feel that recognizing the global diversity represented at uni versity campuses is an institutional accomplishment. However, as university communities strive towards cultivating a sense of belonging and inclusion, findings from this study e ncourage administrators and academics to be sensitive and learn first about how their colleague s prefer to identify before making certain assumptions. Foreign-born faculty members are very diverse. They are not homogeneous. Conclusion In the present quantitative study, I utilized the embeddedness theoretical framework to explore the embeddedness of a cluster sample of for eign-born faculty members who were employed in STEM disciplines at comprehensive unive rsities across the United States. Findings 10 Drexel University Human Resources Webpage http://www.drexel.edu/hr/benefits/voluntary/homepur chase/homepurchase/ 124 indicate that foreign-born faculty members are high ly embedded in the American professoriate. However, data analysis also demonstrated significan t difference for university and community embeddedness. Low community embeddedness was prima rily a function of low appraisal of participants™ perceptions of sacrifices they would make if they were to leave their community. This finding should keep policy makers and universi ty administrators on alert and caution them not to become overconfident in the high aggregate e mbeddedness of their foreign-born faculty by overlooking or dismissing the low community embedde dness dimension. Findings demonstrated that factors such as gender, homeownership, and geographical region of origin were influential on the embeddedne ss of the sample included in the present study. No significant difference was found for emb eddedness based on other demographic and professional characteristics. The current study s upports findings from prior literature about gender differences in STEM higher education and hig hlighted the need for further embeddedness research from a comparative perspective to include perspectives of U.S.-born faculty in the analysis. In addition, there is a need to investigate further the significance of the sacrifices dimension of embeddedness of foreign-born faculty m embers in American higher education. The current study focused only on perspectives of p articipants who are currently employed; it did not include any data from respondents who had made sacrifices by leaving their university or community. As STEM disciplines at universities acr oss the United States rely significantly upon the foreign-born faculty members™ teaching, researc h, and service, understanding their embeddedness in the university and community will f urther inform recruitment and retention policies and practices. University administrators and policymakers who intentionally consider the utility of embeddedness of foreign-born faculty members in STEM disciplines would be 125 well-positioned to foster higher appraisal among th ese faculty members regarding perceived fit, links, and sacrifices on and off the job. 126 APPENDIX 127 APPENDIX A SURVEY 11 Qualifying question: 1. Were you born in USA? 1.1 YES (response disqualifies participation) 1.2 NO Demographic questions: 2. Do you identify as an international faculty member? 2.1 YES 2.2 NO 3. Which of the following best describes your rank at your current university? (single choice items) 3.1 Assistant Professor (tenure-track) 3.2 Associate Professor (tenured) 3.3 Assistant Professor (NOT tenure-track) 3.4 Instructor (NOT tenure-track) 3.5 Professor (tenured) 3.6 Other (please specify) 4. Which of the following categories best describes yo ur academic discipline? (single choice items) 4.1 Sciences 4.2 Technology 4.3 Engineering 4.4 Mathematics 5. What is your gender? 5.1 Female 5.2 Male 6. Which of the following best describes your status i n the United States? (single choice items) 6.1 U.S. Citizen 6.2 Permanent Resident (green card holder) 6.3 Work Visa (i.e., J-1, H-1, E-3b, TN, etc.) 7. Which of the following describes your geographical region of origin? (single choice items) 7.1 Africa 7.2 Asia 7.3 Europe 7.4 Central and South America 7.5 North America 7.6 Oceania 11 This is a condensed version of the survey instrume nt; reflecting the data that was analyzed and repor ted in the present study. The complete survey instrument incl uded the 40-item scale developed by Mitchell and as sociates (2001). 128 8. Please check the box that best describes your circu mstances (YES or NO responses) 8.1 Do you own the home you live in? 8.2 Are you currently married or live with a partner? 8.3 If you are married or live with a partner, does s/ he work outside the home? 8.4 Do you have any children enrolled in the local ele mentary/middle/high school? 9. How many years have you worked at the current unive rsity: (single choice item) 9.1 0-2 9.2 3-5 9.3 6-8 9.4 >10 Embeddedness scale items (anchored on a 5-point Lik ert scale): 10. Please indicate your level of agreement or disagre ement with the following statements. 9.1 I feel I am compatible (a good match) with the uni versity where I work. 9.2 I feel a strong link (attachment) to the universit y where I work. 9.3 I would sacrifice a lot if I left this job. 9.4 I feel I am compatible (a good match) with the com munity where I live. 9.5 I feel a strong link (attachment) to the community where I live. 9.6 I would sacrifice a lot if I left the community wh ere I live. 129 REFERENCES Alberts, H. C. (2008). The challenges and opportun ities of foreign-born instructors in the classroom. Journal of Geography in Higher Education, 32 (2), 189Œ203. Allen, D. (2006). Do organizational socialization tactics influence newcomer embeddedness and turnover? Journal of Management , 32(April), 237-256. Armenti, C. (2000). Women academics blending private and public lives (Order No. NQ53860). Available From ProQuest Dissertations & Theses Glob al. (304657952). Retrieved from http://ezproxy.msu.edu/login?url=http://search.proq uest.com/docview/304657952?accoun tid=12598 Armenti, C. (2010). Women faculty seeking tenure a nd parenthood: Lessons from previous generations. Cambridge Journal of Education , 34(1), 65-83. Armenti, C. (2003). May babies and posttenure babi es: Maternal decision of women professors. Review of Higher Education, 27(2), 211-231. Austin, A. (2002). Preparing the next generation o f faculty: Graduate school as socialization to the academic career. Journal of Higher Education , 73(1), 94-122. Arney, W. R. (1990). Understanding Statistics in the Social Sciences. New York: W. H. Freeman and Company. Baily, E. A. (1997). Hertzberg™s job satisfaction Œ job dissatisfaction theory revisited: A national study of its application to chief housing officers in higher education . Doctoral Dissertation, Mississippi State University. Barnes, L. B., Agago, M. O., & Combs, W. T. (1998). Effects of job-related stress on faculty intention to leave academia. Research in Higher Education , 39(4), 457-469. Batalova, J. (2007). The fibrain gainfl race begins with foreign students. Immigration Policy in Focus . Migration Policy Institute. Washington D.C. Bergiel, E. B., Nguyen, V. Q., Clenney, B. F., & Ta ylor, G. S. (2009). Human resource practices, job embeddedness and intention to quit. Management Research News , 32(2), 205-219. Berrios-Ortiz, E. (2012). Employee job embeddedness: A quantitative study of information technology personnel in the workplace. Capella University. Besich, J. (2005). Job embeddedness verses traditional models of volun tary turnover: A test of voluntary turnover prediction . Doctoral Dissertation. The University of North T exas. 130 Biddle, W. A. (2010). Intent to stay: An examination of the impact of job embeddedness, job satisfaction and job search on faculty retention at Christina colleges . Doctoral Dissertation. Anderson University. Blank, W. R. (1993). Factors associated with job satisfaction and dissat isfaction among college student affairs professional staff . Doctoral Dissertation, University of Northern Colorado. Blau, G. (1993). Further exploring the relationshi p between job search and voluntary individual turnover. Personnel Psychology , 46(2), 313-330. Blau, G. (1994). Testing a two-dimensional measure of job search behavior. Organizational Behavior and Human Decision Processes , 59(2), 288-312. Bluerdorn, A. C. (1982). A unified model of turnov er from organizations. Human Relations , 35(2), 135-153. Boswell, W. R., Boudreau, J. W., & Tichy, J. (2005) . The relationship between employee job change and job satisfaction: The honeymoon-hangover effect. Journal of Applied Psychology , 90(5), 882-892. Bowen, H. R., & Schuster, J. H. (1986). American Professors: A National Resource Imperiled . New York: Oxford University Press. Bretz, R. D. & Judge, T. A. (1994). Person-organiz ation fit and the theory of work adjustment: Implications for satisfaction, tenure, and career s uccess. Journal of Vocational Behavior, 44 , 32-54. Brown, D. G. (1967). The mobile professors. Washington D.C.: American Council on Education. Burt, R. S. (2001). Attachment, decay and social n etworks. Journal of Organizational Behavior, 22, 619-643. Cable, D. M., & Judge, T. A. 1996. Person-organiza tion fit, job choice decisions, and organizational entry. Organizational Behavior and Human Decision Processes , 67, 294Œ 311. Carrell, M. R., Kuzmits, F. E., & Elbert, N. F. (19 92). Personnel / Human Resource Management (4 th ed.). New York, NY: Macmillan. Chatman, J. A. (1989). Improving interactional org anizational research: A model of person- organization fit. Academy of Management Review, 14(3), 333-349. Cohen, J. (1977). Statistical Power Analysis for the Behavioral Scien ces . New York: Academic Press. 131 Cohen, A. M. (1974). Community college faculty job satisfaction. Research in Higher Education , 2(4), 369-376. Connolly, M. D. (2010). The market for skilled imm igrants: The role of student stay rates. Paper presented at Asia-pacific Economic Associatio n Meetings, Hong Kong, July 8-9, 2010. Available online from http://apeaweb.org/con fer/hk10/papers/connolly_md.pdf Corley, E., & Sabharwal, M. (2007). Foreign-born a cademic scientists and engineers: Producing more and getting less than their U S.-born peers? Research in Higher Education , 48 (8), 909Œ940. Creswell, J. W. (2003). Research Design: Qualitative, Quantitative, and Mix ed Methods Approaches (2 nd Ed.). Thousand Oaks: Sage. Crossley, C. D., Bennett, R. J., Jex, S. M., & Burn field, J. L. (2007). Development of a global measure of job embeddedness and integration into a traditional model of voluntary turnover. Journal of Applied Psychology , 92(4), 1031-1042. Cropsey, K. L., Masho, S. W., Shiang, R., Sikka, V. , Kornsetin, S. G., Hampton, C. L. (2008). Why do women faculty leave? Reasons for attrition of women and minority faculty from a medical school: Four-year results. Journal of Women™s Health , 17(7), 111-1118. Cunningham, G. B., Fink, J. S., and Sagas, M. (2005 ). Extensions and further examination of the job embeddedness construct. Journal of Sport management , 19 (July), 319-335. DeVaus (2002). Surveys in social research (5 ed.). Australia: Routledge. Diaspora. (2013). In Merriam-Webster.com . Douglas, J. A., & Edelstein, R. (2009). The global competition for talent: The rapidly chan ging market for international students and the need for a strategic approach in the U.S . Center for Studies in Higher Education. University of California Œ Berkeley. Doyle, W. R. (2008). The babyboomers as faculty: W hat will they leave behind? Change, 40(6), 56-59 Drexel University Human Resources Webpage http://www.drexel.edu/hr/benefits/voluntary/homepur chase/homepurchase/ Drysdale, D. S. (2005). Faculty job satisfaction: Retaining faculty in the new millennium . Doctoral Dissertation. Montana State University. Duderstadt, J. J. (2000). A University for the 21 st Century . Ann Arbor: University of Michigan Press. Ehrenberg, R. G., Rizzo, M. J., & Condie, S. S. (20 03). Start-Up Costs in American Research Universities. Working Papers. Paper 38. Available o nline from http://digitalcommons.ilr.cornell.edu/workingpapers /38 132 Ehrenberg, R. G., Kasper, H., & Rees, D. (1991). F aculty turnover at American colleges and universities: Analysis of AAUP data. Economics of Education Review , (10), 99-110. Emmerik, H. & Sanders, K. (2004). Social embeddedn ess and job performance of tenured and non-tenured professionals. Human Resource Management Journal , 14(1), 40-54. Feldman, D. C & Bolino, M. C. (1998). Moving on ou t: When are employees willing to follow their organization during corporate relocation? Journal of Organizational Behavior, 19 , 275-288. Felps, W., Mitchell, T., Hekman, D., Lee, T., Holto m, B., Harman, W. (2009). Turnover contagion: How coworkers™ job embeddedness and job search behaviors influence quitting. Academy of Management Journal , 52(3), 545-561. Finn, M. G. (2012). Stay Rates of Foreign Doctorat e Recipients from the U.S. Universities 2009. Oak Ridge, TN: Oak Ridge Institute for Science and Education. Available online from http://orise.orau.gov/files/sep/stay-rates-foreign- doctorate-recipients-2009.pdf Finn, M. G. (2010). Stay Rates of Foreign Doctorat e Recipients from the U.S. Universities 2007. Oak Ridge, TN: Oak Ridge Institute for Science and Education. Available online from http://orise.orau.gov/files/sep/stay-rates-foreign- doctorate-recipients-2007.pdf Finn, M. G. (2007). Stay Rates of Foreign Doctorat e Recipients from the U.S. Universities 2005. Oak Ridge, TN: Oak Ridge Institute for Science and Education. Available online from http://orise.orau.gov/files/sep/stay-rates-foreign- doctorate-recipients-2005.pdf Finn, M. G. (2005). Stay Rates of Foreign Doctorate Recipients from the U.S. Universities 2003. Oak Ridge, TN: Oak Ridge Institute for Science and Education. Available online from http://orise.orau.gov/files/sep/stay-rates-foreign- doctorate-recipients-2003.pdf Finn, M. G. (2003). Stay Rates of Foreign Doctorat e Recipients from the U.S. Universities 2001. Oak Ridge, TN: Oak Ridge Institute for Science and Education. Available online from http://orise.orau.gov/files/sep/stay-rates-foreign- doctorate-recipients-2001.pdf Finkel, S. K., & Olswang, S. G. (1996). Child Rear ing as a Career Impediment to Women Assistant Professors. Review of Higher Education , 19(2) 123-139. Fox, M. (2001). Women, science, and academia: Grad uate education and careers. Gender & Society , 15(5), 654-666. Friedman, R. A. & Holtom, B. (2002). The effects of network groups on minority employee turnover intentions. Human Resources Management, 41 , 405-421. Garner, S. K. (2012). I couldn™t wait to leave the toxic environment: A mixed methods study of women faculty satisfaction and departure from one r esearch institution. NASPA Journal About Women in Higher Education, 5(1), 71-95. 133 Gaylor, Kristena P. (2004). fi Organizational Commitment in Higher Education: A Multidimensional Perspective .fl Doctoral Dissertation, Jackson State University . Giosan, C. (2004). Antecedents of Job Embeddedness . Doctoral Dissertation. New School University. Granovetter, N. (1995). Getting a Job: A Study of Contacts and Careers. Chicago: University of Chicago Press. Granovetter, M. (1985). Economic action and social structure: The problem of embeddedness. Journal of Sociology, 9(3), 481-510. Griffeth, R. W., & Hom, P. W. (2001). Retaining valued employees . Thousand Oaks: Sage. Griffeth, R. W., Hom, P. W., & Gaertner, S. (2000). Meta-analysis of antecedents and correlates of employee turnover: Update, moderator test, and r esearch implications for the next millennium. Journal of Management , 3, 463-475. Hagedorn, L. S. (2000). Conceptualizing faculty jo b satisfaction: Components, theories, and outcomes. In What contributes to job satisfaction among faculty and staff, ed. L. S. Hagedorn 5-20. San Francisco, CA: Jossey-Bass. Hamilton, M. B. (2003). Online survey response rat es and times: Background and guidance for Industry. White paper. Available online from http://www.supersurvey.com/papers/supersurvey_white _paper_response_rates.htm Hardcastle, I. B. (2010). Institutional commitment of faculty in evangelical colleges and universities. Doctoral dissertation. The Southern Baptist Theol ogical Seminary. Hawthorne, L. (2008). The growing global demand for students as skilled m igrants . Washington, D.C.: Migration Policy Institute. Avai lable online from http://www.migrationpolicy.org/transatlantic/intlst udents.pdf Hercog, M. (2008). The role of the state in attrac ting highly-skilled migrants: The case of the Netherlands. EIPAScope, (3), 19-24. Available online from http://aei.pitt.edu/11032/1/20090130170607_SCOPE200 8-3_3_MetkaHercog.pdf Herzenberg, F. I. (1959). The Motivation to Work . New York, NY: John Wiley and Sons. Hocking, R. R. (1996). Methods and Applications of Linear Models: Regressi on and Analysis of Variance. New York: Wiley. Holland, C. B. (1992). The influence and organizational commitment on inte ntion to leave of nurse educators . Doctoral dissertation, Louisiana State Universit y. Holtom, B. & O™Neill, B. (2004). Job embeddedness: A theoretical foundation for developing a comprehensive nurse retention plan. The Journal of Nursing Administration , 34(5), 216- 227. 134 Holtom, B. C., & Inderrieden, E. J. (2006). Inte grating the unfolding model and job embeddedness model to better understand voluntary t urnover. Journal of Managerial Issues , 18(4), 435-452. Hom, P. W. & Griffeth, R. W. (1995). Employee Turnover. Cincinnati, OH: South-Western College Publishing. Hom, P. W., & Kinicki, A. J. (2001). Towards a gre ater understanding of how dissatisfaction drives employee turnover. Academy of Management Journal , 44(5), 975-987. Hutcheson, G. D. (2011). Ordinary Least-Squares Re gression. In L. Moutinho and G. D. Hutcheson, The SAGE Dictionary of Quantitative Mana gement Research (224-228). Los Angeles: SAGE Publications. Igbaria, M., & Siegel, S. R. (1991). The reasons f or turnover of information systems personnel. Information and Management , 23, 321-330. International. 2013. In Merriam-Webster.com . Iredale, R. (2000). Migration policies for the hig hly skilled in the Asia-Pacific region. International Migration Review , 34(3), 882-906. Johnsrud, L. K. & Rosser, V. J. (2002). Faculty me mbers™ morale and their intention to leave. The Journal of Higher Education, 73(4), 518-542, Keppel, G. (1991). Design and Analysis: A Researcher™s Handbook. New Jersey: Prentice Hall. Kim, D., Wolf-Wendel, L., & Twombly, S. (2011). In ternational faculty: Experiences of academic life and productivity in U.S. universities . The Journal of Higher Education, 82 (6), 720-747. Kim, D., Wolf-Wendel, L., & Twombly, S. (2011). Similarities and differences in mobility for U.S. and Non-U.S. citizen faculty: Exploring the un derlying motivations for why they leave . Paper presented at the Annual Conference for the Study of HIgher Education, Charlotte, North Carolina. Kim, D., Wolf-Wendel, L., & Twombly, S. (2011). Foreign-born faculty mobility: Longitudinal study of foreing-born faculty mobility . Paper presented at the Annual Conference of American Education Research Association (AERA), New Orleans, Louisiana. Kittleson, M. J. (1997). Determining effective fol low-up of email surveys. American Journal of Health Behavior , 21(2), 193-196. Kuenzi, A. J., Matthews, C. M., & Magnan, B. F. (20 06). Science, Technology, Engineering, and Mathematics (STEM) Education Issues and Legisla tive Options. Congressional Research Service Reports. Paper 34. The Library of Congress. 135 Kuenzi, Jeffrey J., (2008). Science, Technology, E ngineering, and Mathematics (STEM) Education: Background, Federal Policy, and Legislat ive Action. Congressional Research Service Reports. Paper 35. Available online from http://digitalcommons.unl.edu/crsdocs/35 Labatmediene, L., Endriulaitien, A., & Gustainiene, L. (2007). Individual correlates of organizational commitment and intention to leave th e organization. Baltic Journal of Management, 2, 196-212. Lauver, K. J. & Kristof- Brown (2001). Distinguish ing between employees™ perception of person-job fit and person-organization fit. Journal of Vocational Behavior, 59, 454- 470. Law, K. S., Wong, C., & Mobley, W. H. (1998). Towa rd a taxonomy of multidimensional constructs. Academy of Management Review , 23(4), 741-755. Lee, T. W. & Maurer, S. (1999). The effects of fam ily structure on organizational commitment, intention to leave and voluntary turnover. Journal of Management Issues, 11, 493-513. Lee, T. W., Mitchell, T. R., Holtom, B. C., McDanie l, L., & Hill, J. (1999). The unfolding model of voluntary turnover: A replication and exte nsion. Academy of Management Journal, 42, 450-462. Lee, T. W., Mitchell, T. R., Sablynksi.C. J., Burto n, J. P., Holtom, B. C. & Brooks, C. (2004). The effects of job embeddedness on organizational c itizenship, job performance, volitional absences, and voluntary turnover. Academy of Management Journal , 47 , 711- 722 Lee, T. W., & Mowday, R. T. (1987). Voluntary leav ing an organization: An empirical investigation of Steers and Mowday™s model of turno ver. Academy of Management Journal, 30(4), 721-743. Maertz, C. P., & Campion, M. A. (1998). 25 years o f voluntary turnover research: A review and critique. International Review of Industrial and Organization al Psychology , 13, 49-81. Maertz, C. P., & Campion, M. A. (2004). Profiles i n quitting: Integrating process and content turnover theory. Academy of Management Journal , 47 (4), 566-582. Maertz, C. P. & Griffeth, R. W. (2004). Eight motiv ational forces and voluntary turnover: A theoretical synthesis with implications for researc h. Journal of Management , 30 (5), 667- 683. Available online from http://jom.sagepub.com/c ontent/30/5/667.full.pdf+html Mallol, C. M., Hotom, B. C., & Lee, T. W. (2007). Job embeddedness in a culturally diverse environment. Journal of Business Psychology. 22, 35-44. Mamisieshvili, K. & Rosser, V. J. (2011). Examini ng the relationship between faculty productivity and job satisfaction. The Journal of the Professoriate , 5(2), 100-132. Available online from http://jotp.icbche.org/2011/5 -2_Mamiseishvili_p.100.pdf 136 Manrique, C. G., & Manrique, G. G. (1999). Third world immigrant women in American higher education. In G. A. Kelson & D. L. DeLaet (Eds), Gender and Immigration . New York: New York University Press. Master's Colleges and Universities. (2011). Master 's colleges and universities. (Classification Description) [Web page]. The Carnegie Foundation f or the Advancement of Teaching. Retrieved November 15, 2012, from http://classifications.carnegiefoundation.org/descr iptions/basic.php Mason, M. A., Goulden, M. (2004). Do babies matter (Part II)? Closing the baby gap. Academe , 88(6), 21-27. Menges, R., & Exum, W. (1983). Barriers to the pro gress of women and minority faculty. Journal of Higher Education , 54(2), 123-144. Miller, H. E., Karterberg, R. & Hulin, C. L. (1979) . Evaluation of the Mobley, Horner and Hollingworth model of employee turnover. Journal of Applied Psychology, 64(5), 509- 517. Mitchell, T., Holton, B., and Lee, T. (2001). Why people stay: Using job embeddedness to predict voluntary turnover. Academy of Management Journal . 44(6), 1102-1121. Mobley, W. H. (1977). Intermediate linkages in the relationship between job satisfaction and employee turnover. Journal of Applied Psychology , 62(2), 237-240. Mossholder, K. W., Settoon, R. P. & Henagan, S. C. (2005). Relational perspective on turnover: Examining structural, attitudinal, and behavioral p redictors. Academy of Management Journal , 48 , 607-618. Mowday, R. T., Porter, L. W., & Steers, R. M. (1982 ). Employee-Organizational linkages: The psychology of commitment, absenteeism, and turnover . New York: Academic Press. Mueller, C. W & Price, J. L. (1990). Economic, psyc hological and sociological determinants of turnover. Journal of Behavioral Economics, 19, 321-335. Muijs, D. (2004). Doing Quantitative Research in Education with SPSS . Los Angeles: SAGE Publications. Nagowski, M. P. (2004). Associate professor turnover at America™s public an d private institutions of higher education (CHERI Working Paper #31). Available online from http://digitalcommons.ilr.cornell.edu/student/4/ National Science Board. 2008. Science and Engineering Indicators 2008 . Arlington, VA: National Science Foundation. National Science Foundation (20015). Women, Minorities, and Persons with Disabilities in Science and Engineering . Arlington, VA (NSF 15-311). Source: http://nsf.gov/statistics/2015/nsf15311/digest/them e7.cfm#women 137 Nelson, D. J., & Rogers, D. C. (2005). A national analysis of diversity in science and engineering faculties at research universities . Neuman, W. L. (2011). Social research methods: Qualitative and quantitati ve approaches (6, illustrated ed.). Boston: Allyn and Bacon. Neuman, Y. (1978). Predicting faculty job satisfac tion in university departments. Research in Higher Education , 9(3), 261-275. Ng, T. W, & Feldman, D. C. (2007). Organizational embeddedness and occupational embeddedness across career stages. Journal of Vocational Behavior , 70 (2), 336-351. NSB. (2003). National Science Board: " The Science and Engineering Workforce: Realizing America's Potential" . Washington, DC: National Science Foundation. Nunnally, J. (1978). Psychometric theory . New York: McGraw-Hill. Olanrewaju, L. A. (2002). Job satisfaction of the business faculty in Virgini a Community College system: An examination of Hertzberg™s motiv ation-hygiene theory . Doctoral Dissertation, George Mason University. Olsen, D. (1993). Work satisfaction and stress in the first and third year of academic appointment. Journal of Higher Education , 64(4), 453-471. Olsen, D., Maple, S.A., & Stage, F.K. (1995). Wome n and minority faculty job satisfaction: Professional role interests, professional satisfact ions, and institutional fit. The Journal of Higher Education , 66(3), 267-293. Ott, R. L. & Longnecker, M. (2008). An introduction to statistical methods and data ana lysis. Duxbury Press. Pallant, J. (2007). SPSS survival manual: A step-by-step guide to data analysis using SPSS for windows version 15. Australia: Allen & Unwin. Pfeffer J. (1994). Competitive Advantage Through People . Boston: Harvard Business School Press . Porter L, & Steers R. M. (1963). Organizational, wo rk, and personal factors in employee turnover and absenteeism. Psychological Bulletin, 80 , 151-176. Price, J. L. (1977). The study of turnover. Ames: Iowa State University Press. Ramesh, A., & Gelfand, M. J. (2010). Will they sta y or will they go? The role of job embeddedness in predicting turnover in individualis tic and collectivistic cultures. Journal of Applied Psychology . 95(5), 807-823. Ramesh, A. (2007). Replicating and Extending Job Embeddedness Across C ultures: Employee Turnover in India and The United States . Doctoral Dissertation. University of Maryland. 138 Rea, L. M., & Parker, R. A. (1997). Designing and Conducting Survey Research: A Comprehensive Guide . San Francisco: Jossey-Bass. Rothblum, E. (1988). Leaving the ivory tower: Fact ors contributing to women™s voluntary resignation from academia. Frontiers , 10(2), 14-17. Sabharwal, M. (2011). High-skilled immigrants: How satisfied are foreign-born scientists and engineers employed at American universities?fl Review of Public Personnel Administration , 31(2), 143-170. Sabharwal, M., & Corley, E. A. (2009). Faculty job satisfaction across gender and discipline. The Social Science Journal , 46, 539-556. Sanders, J. W. (2015). The relationships among leader-member exchange, off -the-job embeddedness, and turnover intention . Doctoral Dissertation. Sanderson A., V. Phua, and D. Herda. 2000. The Ame rican Faculty Poll. Chicago: National Opinion Research Center. Available online from http://www.norc.org/PDFs/publications/SandersonA_Am er_Faculty_Poll_2000.pdf Sapnas, K. G. and Zeller, R. A. (2002). Minimizing sample size when using exploratory factor analysis for measurement. Journal of Nursing Measurement . 10(2), pp. 135-154. Schuster, J. H., & Finkelstein, M. J. (2006). The American Faculty . Baltimore: Johns Hopkins University Press. Scotland, J. (2012). Exploring philosophical underp innings of research: Relating ontology and epistemology to the methodology and methods of the scientific, interpretive, and critical research paradigms. English Language Teaching , 5(9), 9-16. Shaffer, M. A., & Harrison, D. A. (1998). Expatria tes™ psychological withdrawal from international assignments: Work, nonwork, and famil y influences. Personnel Psychology , (51), 87-118. Shaw J. D., Delery, J.E., Jenkins, G. D. & Gupta, N .(1998). An organizational level analysis of voluntary and involuntary turnover. Academy of Management Journal, 41 , 511-525. Skachkova, P. (2007). Academic careers of immigran t women professors in the U.S. Higher Education , (53), 697-738. Smart, J. C. (1990). A causal model of faculty tur nover intentions. Research in Higher Education , 31(5), 405Œ424. Stratford, M. (November 30, 2012). House passes vi sa bill for foreign graduates, over democrats™ objections. The Chronicle of Higher Education . Tabachnick, B. G., & Fidell, L. S. (2007). U sing Multivariate Statistics . Boston: Pearson. 139 Tack, M. W. & Patitu, C. L. (1992). Faculty job satisfaction: Women and minorities in p eril. ASHE-ERIC higher education report no. 4. Washington , DC: Association for the study of higher education. Tanova, C., & Holton, B. C. (2008). Using job embe ddedness to explain voluntary turnover in four European countries. The International Journal of Human Resource Managem ent , 19(9), 1553-1568. The Economist (October 5, 2006). The Battle for br ainpower. Available online from http://www.economist.com/node/7961894/print Thelin, J. R. (2004). A History of American Higher Education . Baltimore, MD: The Johns Hopkins University Press. Theobald, R. (2007). Foreign-Born Early-Career Faculty in American Highe r Education: The Case of the Discipline of Geography . Doctoral Dissertation. University of Colorado a t Boulder. Thurstone, L. L. (1947). Multiple-factor analysis . Chicago: The University of Chicago Press. U.S. Department of Homeland Security, Immigration a nd Customs Enforcement, STEM Designated Degree Programs (2012). Retrieved Febru ary 03, 2013, from http://www.ice.gov/sevis/stemlist.htm Umbach, P. D. (2004). Web surveys: Best practices. New Directions for Institutional Research, (121), 23-38. Ward, K., & Lisa Wolf-Wendel, L. (2004). Academic Motherhood. Review of Higher Education , 27(2) pp. 233-257. Wasem, R. E. (2012). Immigration of Foreign Nationals with Science, Tech nology, Engineering, and Mathematics (STEM) Degrees . Congressional Research Service Report R42530. Available online from http://www.fas.org/sgp/crs/mi sc/R42530.pdf Wei, N. (2007). Do we belong here? An exploration of foreign-born f aculty's organizational attachment at a U.S. research university. Dissertat ion Abstracts International Section A: Humanities and Social Sciences , 1633-1633. Weiler, W. C. (1985). Why do faculty members leave a university? Research in Higher Education , 23(2), 270Œ278. Wells, R. (2007). International faculty in U.S. com munity colleges. New Directions for Community Colleges. (133), 77-82. Available online from http://www.accbd.org/articles/index.php/attachments /single/188 Xu, Y. J. (2008). Gender disparity in STEM discipl ines: A study of faculty attrition and turnover intentions. Research in Higher Education , 49, 607-624. 140 Zhou, Y. & Volkwein, J. F. (2004). Examining the i nfluences of faculty departure intentions: A comparison of tenured versus nontenured faculty at research universities using NSOPF- 99. Research in Higher Education, 45(2), 139-176