ONLINE COOPERATIVE LEARNING: EFFECTS OF DESCRIPTIVE NORMS AND COOPERATIVE PARTNER MESSAGES ON ENGAGEMENT, MOTIVATION , AND ACHIEVEMENT By Hannah Klautke A DISSERTATION S ubmitted to Michigan State University in partial fulfillment of the requirements for the degree of Communication Doctor of Philosophy Educational Psychology and Educational Technology Doctor of Philosophy 2015 ABSTRACT ONLINE COOPERATIVE LEARNING: EFFECTS OF DESCRIPTIVE NORMS AND COOPERATIVE PARTNER MESSAGES O N ENGAGEMENT, MOTIVATION, AND ACHIEVEMENT By Hannah Klautke This study addresses effects of social influence on students engaged in a cooperative online learning activity on two levels: At the class level, the communication of completion percentage upd ates that indicate continued and timely participation in the activity as the norm for the course is hypothesized to improve engagement outcomes. At the dyadic level, the use of cooperative partner messages highlighting the interdependent nature of the task is hypothesized to increase cooperative - and decrease competitive task perceptions, along with subsequent desirable effects such as epistemic conflict regulation; positive affect toward the partner; and increased intrinsic motivation, engagement, and task performance. Specific hypotheses tested in this study are derived from the theory of normative social behavior, social in terdependence theory, and socio cognitive conflict theory, and related potential mechanisms and interactions with additional factors ar e discussed. To test the hypotheses and to explore related mechanisms, 193 undergraduate students were randomly assigned to four experimental conditions following a 2 x 2 experimental - control design. All participants engaged in a week - long asynchronous on line constructive controversy that required them to exchange one post per day regarding a controversial topic with a partner, before filling out an online survey about the dependent variables of interest. Using a confederate approach, participants in two o f these groups worked with a partner that used a set of cooperat ion - promoting partner messages such as statements highlighting the importance of consider ing argument joint integration essay . Participants in the other two g roups received the same substantive responses without these cooperative enhancements. The second experimental factor consisted of the presence/absence of daily messages within the activity site, indicating punctual participation by the majority of the grou p, R esults largely support social interdependence theory for the asynchronous mediated context examined here, in that cooperative partner messages increased perceived cooperation and reduced the pr eference to work on t his kind of task individually. Unexpectedly, c ooperative partner messages failed to signi ficantly reduce competitive percptions , but in line with expectations, they did produce increases in attraction, motivation, and engagement. Inter estingly, such messages also raised learning activity, an unexpected but potentially benefi cial side effect in the online learning context. The direct presentation of completion percentages di d affect perceived descriptive norms in the expected direction as well. However, it also led to significant reductions in motivational variables such as interest in the activity and perceived competence, eradicating some of the beneficial effects produced by cooperative partner messages. F indings support the detrimental effects of upward comparisons in educational contexts and indicate caution regarding the use of explicit normative messages within asynchronous online activities. I nterventions guiding onlin e learners to be conscious of constructive and destructive feedback loops in the context of asynchronous computer mediated intellectual conflict as found in many cooperative learning tasks are supported , to deliberately restore some of the cooperative perc eptions that previous work has shown to suffer under asynchronous mediated conditions. iv . v ACKNOWLEDGMENTS I would like to express my gratitude to all members of my dissertation committee: To Cary, for setting records in being thinly stretched and still managing to be generous with his time, effective, caring , and all - around wonderful. To Rand, for his support, sense of humor and wise counsel on all things, from the future of education to where to find oases in the desert (both figurat ively and literally ) . To Jim, whose pragmatism and encouragement made a world of difference. To Maria, whose expertise and constructive guidance sharpened my thinking. To Vernon, who m I could count on throughout the years for helpful input and discerning f eedback. Many ad ditional faculty members, staff members and doctoral students have contributed to this project in ways both big and small, from indispensable assistance with recruitment to thoughtful informal input throughout the pro cess. A special thank you to Ursula and Arndt for serving as an astute sounding board whenever needed; to Val eria , Andrii, and Aaron for their help with the pilot project and the website ; to Greg for filling in as the best doctoral cohort I could wish for ; r esearch team f or being the amazing colleagues they are. And to Sunny , whose interests lie far outside of cooperative online learning, but who made my life whole while working on this project . vi TABLE OF CONTENTS LIST OF TABLES ................................ ................................ ................................ ....................... viii LIST OF FIGURES ................................ ................................ ................................ ....................... ix K EY TO ABBREVIATIONS ................................ ................................ ................................ ......... x CHAPTER 1: INTRODUCTION ................................ ................................ ................................ ... 1 Statement of the Problem ................................ ................................ ................................ .......... 1 Constructive Controversy as an Exemplar for Cooperative (Online) Learning ....................... 3 Purpose of the Study and Theoretical Frameworks ................................ ................................ .. 3 Organization of Chapters ................................ ................................ ................................ .......... 5 CHAPTER 2: LITERATURE REVIE W ................................ ................................ ........................ 7 Growth and Challenges of Online Learning ................................ ................................ ............. 7 Online Learning via Asynchronous Course Discussions ................................ .......................... 8 Constructive Controversy, Face to Face and Online ................................ .............................. 10 Distal Social Influence: The Role of Course - wide Social Norms Messages .......................... 1 3 Using behavioral prevalence information to impact behavior: Descriptive norms interventions ................................ ................................ ................................ ..................... 1 3 The social norms approach ................................ ................................ ......................... 1 5 The theory of normative social behavior ................................ ................................ .... 1 7 Further possible moderators of norms effects ................................ ............................. 2 5 Descriptive norms can have unintended effects ................................ ................................ 2 8 Guarding against un intended norms effects ................................ ................................ ...... 3 0 An alternative mechanism for effects of descriptive norms in educational contexts: The big - fish - little - pond effect (BFLPE) ................................ ................................ ........... 3 6 Proximal Social Influence: The Role of Cooperative Partner Messages ................................ 39 Online argumentation as a situat ion at risk for individualistic and competitive goal construction ................................ ................................ ................................ ....................... 4 0 Scaffolding o nline course i nteractions ................................ ................................ .............. 4 1 Theoreti cal bases for the link between perceptions of c ooperation and productive a rgumentation ................................ ................................ ................................ ................... 4 5 Social interdependence t heory ................................ ................................ .................... 4 6 Fostering prom otive interactions online via cooperative partner messages ............... 4 9 ................................ ......................... 49 Socio - cognitive conflict t heory ................................ ................................ ................... 5 1 CHAPTER III: METHOD ................................ ................................ ................................ ............ 5 5 Overview of the Research Design ................................ ................................ ........................... 5 5 Study Context and Participants ................................ ................................ ............................... 5 5 Procedures ................................ ................................ ................................ ............................... 5 7 Me asured Variables ................................ ................................ ................................ ................ 60 Motivation and attraction ................................ ................................ ................................ .. 6 0 vii Perceptions of social interdependence ................................ ................................ .............. 6 2 Socio - cognitive conflict regulation ................................ ................................ ................... 63 Perceived descriptive norms ................................ ................................ ............................. 64 Potential moderators of norms - effects ................................ ................................ .............. 65 Additional measures ................................ ................................ ................................ .......... 6 7 Behavioral engagement and achiev ement ................................ ................................ ......... 6 7 Overview of Hypotheses, Data Sources, and Analyses ................................ .......................... 6 8 CHAPTER IV : RESULTS ................................ ................................ ................................ ............ 73 Participant Flow ................................ ................................ ................................ ...................... 73 Manipulation Check ................................ ................................ ................................ ................ 74 Factorial Analyses of Experimental Conditions ................................ ................................ ..... 75 Perceived descriptive norms ................................ ................................ ............................. 76 Social interdependence ................................ ................................ ................................ ..... 76 M otivation and r elate dness ................................ ................................ ............................... 77 Behavioral engagement and achievement ................................ ................................ ......... 80 Additional exploratory group comparisons ................................ ................................ ...... 81 Correlational Analyses ................................ ................................ ................................ ............ 81 Hierarchical Multiple Regression for Moderation of Norms Effects ................................ ..... 84 CHAPTER V: DISCUSSION ................................ ................................ ................................ ....... 87 Behavioral Prevalence Information, Social norms, and Engagement ................................ ..... 88 Factorial analyses ................................ ................................ ................................ .............. 88 Correlational and moderation analyses ................................ ................................ ............. 92 Additional effects of behavioral prevalence information ................................ ................. 93 Cooperation - promoting Messages and Social Interdependence ................................ ............. 95 The Role of Conflict Regulation ................................ ................................ ............................ 99 Limitations and Delimitations of the Study ................................ ................................ .......... 100 Implications ................................ ................................ ................................ ........................... 104 Implications for theory ................................ ................................ ................................ .... 105 The theory of normative social behavior ................................ ................................ .. 105 The big - fish - little pond effect and aversive perceptions of being controlled ........... 108 Social interdependence theory ................................ ................................ .................. 109 Sociocognitive co nflict theory ................................ ................................ .................. 114 The role of low - stakes conflict in SIT and SCCT ................................ ..................... 115 Implications for practice ................................ ................................ ................................ . 116 APPENDICES ................................ ................................ ................................ ............................ 122 Appendix A: Constructive Controversy site description with instructions .......................... 123 Appendix B: Constructive controversy materials ................................ ................................ . 126 Appendix C: F inal survey items ................................ ................................ ........................... 128 Appendix D: Participant Flow. ................................ ................................ ............................. 131 Appendix E: Actual completion percentages ................................ ................................ ........ 133 Appendix F: Message generation schedule ................................ ................................ ........... 134 REFERENCES ................................ ................................ ................................ ........................... 135 viii LIST OF TABLES Table 1: Sample characteristics . . 56 Table 2: Protot ypes of cooperative partner messages 59 Table 3: Levels of integration of arguments 68 Table 4: Research hypotheses, data sources, analyses 69 Table 5: Descriptive statistics . 74 Table 6: Intercorrelations of dependent variables ... 83 Table 7: Normative predictors of timeliness . . 85 Table 8: Normative predictors of step completion . 86 Table 9: Participant flow .. 131 Table 10: Actual step completions and completion percentages by day and time for sample week .. .... . 133 ix LIST OF FIGURES Figure 1: Theory of normative social behavior (e.g. Rimal et al., 2007) 18 Figure 2: Social interdependence Theory . 47 Figure 3: Socio - cognitive conflict Theory 52 Figure 4 a : Potential distal - proximal influence interactions: CPMs as moderator .. 53 Figure 4 b : Potential distal - proximal influence interactions: Norms messages as moderator ...... 54 Figure 5: Factorial design showing experimental and control groups . 55 Figure 6: Two - way ANOVA results for descriptive norms 76 Figure 7: Two - way ANOVA results for interest/enjoyment . 78 Figure 8: Two - way ANOVA results for competence .. 78 Figure 9: Two - way ANOVA results for attraction . 79 Figure 10: Factorial results for length of final statement d raft .. 80 Figure 11: Screen shot of the controversy website, Day 1 . 123 Figure 12: Message generation schedule balancing substantive content across conditions . . 134 x KEY TO ABBREVIATIONS BFLPE Big - fish - little - pond effect CC Constructive c ontroversy CMC Computer - mediated c ommunication CMS Course management s ystem (e.g. Blackboard, Desire2Learn) CPMs Cooperation - promoting partner messages FtF Face - to - face IMI Intrinsic motivation inventory TNSB Theory of normative social behavior SCCT Socio - cognitive conflict theory SIS Social interdependence scale SIT Social interdependence theory SN Social norms SNA Social norms approach 1 CHAPTER 1 : INTRODUCTION Statement of the Problem One of the main concerns regarding online education generally and the quality of online student discussions specifically is the problem of loss of engagement and student attrition over time. At the course level, students drop out at significantly higher rates than they do in traditional fa ce to face courses (e.g. Carr, 2000; Patterson & McFadden, 2009). Similarly, investigations at the level of individual course discussion activities have indicated low completion rates for online versions of their face - to - face equivalents (Roseth, Saltarell i, & Glass, 2011) and dying carrying out course discussions online have been attributed to a lack of social presence (i.e, a awareness of the o ther person/persons in the course; e.g., Stodel, Thompson, & MacDonald, 2006; Gunawardena, 1995); lost momentum due to the asynchronous nature of the interactions ( e.g., Guzdial & Turns, 2000; Hara, Bonk, & A ngeli, 2000; Hewitt, 2005 ); and uncertainty reg arding instructor expectations (e.g., Rourke, Anderson, Garrison, & Archer, 2001; Sammons, 2007), among other factors. At the root of most of the proposed explanations lies the reduced availability of feedback, which is relatively immediately available fr om both peers and instructors in face - to - face contexts classmates and instrumental task - related information, an hour spent in a face - to - face setting can go much f arther than the same amount of time spent interacting virtually (Walther & Parks, 2002), where direct questioning becomes necessary to gather both social (Tidwell & Walther, 2006) and task related information. In contrast, for ongoing (informal) task - relat ed feedback, 2 student groups working on cooperative learning tasks in traditional face - to - face contexts are able to glean expression of the instructor who is walking by their group and overhea ring suggestions made, or similar informative cues. They can easily observe other group s engagement with the task and even overhear how nearby other groups are advancing. This information provides informal benchmarks, or social norms, to which their own be havior can be compared. In a face - to - face classroom, a ctivities are therefore regulated in part by the environment, including peers and the instructor, increasing the odds that students stay involved with the task. In the absence of such regulating cues, d emands on - regulation skills are particularly high in asynchronous online learning contexts where the place and time to engage in coursework has to be set by the individual student (Joo, Bong, & Choi, 2000; Williams & Hellman, 2004; Young, 19 96 ). Additionally, face - to - face contexts may be somewhat more conducive to fostering in group members a sense of accountability to their peers e.g. for meeting required course sub - goals, along with a sense of cooperation as the best way of achieving thes e goals which can provide mechanisms for increased and sustained engagement. By definition, cooperative learning activities that require students to discuss course content are characterized by a mutual goal that requires input from multiple individuals a nd individual accountability (Johnson & Johnson, 1994). However, largely due to the reduced availability of feedback on these different levels, the instructions and scaffolding of such interdependency - fostering cooperative activities that are successful in face - to - face contexts may not be transferable to online contexts without careful consideration of how underlying mechanisms are affected by features of the context and of possible adaptations that address such effects (e.g. Bures, Abrami, & Schmid, 2010; Saba & Shearer, 2004). 3 Constructive Controversy as an Exemplar for Cooperative (Online) Learning The cooperative learning technique of constructive c ontroversy (Johnson & Johnson, elational processes students argue for opposing sides of an issue before switching perspectives and eventually generating an joint statement integrating the best e vidence from both sides, has a long track - record of beneficial outcomes on student motivation and achievement ( Johnson & Johnson, 2007, 2009 ), but it has been shown to lose some of its beneficial effects in asynchronous online contexts. Examination of such a well - established procedure further allows for an exploration of some of the possible accommodations needed to replicate its desirable learning effects in such contexts, specifically, of how to reconstruct productive kinds of social influence at various levels (school/course/small group/dyad) that play important roles in traditional educational settings. Purpose of the Study and Theoretical Frameworks This study examines performance in a cooperative learning activity: (1) an explicit awareness of other progress, conceptualized here as the behavioral prevalence of timely completion of sub - tasks at the class level, and (2) confederate - generated cooperation - promoting partner messages as based on overlapping frameworks of cooperative skills described in the next chapter. Various dependent variables hypothesized to be either directly or indirectly affected are assessed. Engagement and achievement are conceptualized here as differences i n completion rates, proportion of daily task steps completed in a timely fashion, length of the final group statement, and level of integration of the joint statement, and factual content knowledge. Theory - based mechanisms producing different outcomes are also explored, including (a) perceived descriptive 4 course norms, (b) the perception of the task as competitive, individualistic, or cooperative in nature, and (c) approaches to conflict regulation (such as a focus on issues a versus focus on out - arguing th e other person ). Possible moderators of such effects, such as identification with the reference group and ego - involvement in the activity are assessed as well. The proposed hypotheses are derived from a set of complementary theories including the theory of normative social behavior (Rimal & Real, 2005; Lapinski & Rimal, 2005), social interdependence theory (Deutsch, 1949; 1962; Johnson & Johnson, 1989; 2005) and socio - cognitive conflict theory (Doise & Mugny, 1984), described in the next chapter. Specifica lly, this study will address whether exposing online learners to messages indicating that regular and continued participation in an online constructive controversy activity is the norm for the course leads to more persistence with the cooperative activity; and whether exposure to messages that invite and acknowledge input and reiterate the interdependent nature of the task lead to increased cooperative perceptions, along with the subsequent desirable effects. Such effects include a focus on learning (rather than a focus on mechanically rebutting partner messages), positive affect toward the partner, and increased motivation expressed as enjoyment, perceived utility of the task, reduced perceived cost of engaging in the task , and reduced work avoidance relate d to the task. Beyond relational affect and motivation constructs, ultimate desirable effects include increased engagement (evidenced e.g. by continued reliable and punctual participation) and achievement (including the quality of the integration statement crafted and factual content knowledge) . The study aims to contribute to our understanding of the extent to which some fundamental propositions of social interdependence theory hold within the asynchronous computer - mediated communication context which ha s previously been shown to involve effects that may delineate boundary conditions for SIT after the deliberate introduction 5 of strategies to counteract these effects via cooperative partner messages. Given the complexity of effects of behavioral, normati ve information on behavior, the study pursues more exploratory aims with regards to this factor, particularly with respect to motivational effects. Organization of Chapters To deepen the cursory overview of the problem, frameworks employed, and goals of the study provided in this first chapter, chapter 2 will provide a review of the pertinent literature that addresses the two different levels of influence examined in this study the more distal, course - wide information on daily completion percentages, f ollowed by the more proximal, social - interdependence - theory based cooperative partner messages. After taking a look at asynchronous online course discussions and findings regarding their potential for and limitations in fostering motiva tional and achieveme nt - related student outcomes, an overview of the literature pertaining to the first aspect is provided, examining the mixed results of using behavioral prevalence information to impact behavior, and exploring how findings relate to student engagement in onl ine constructive controversy. Concluding that first part of the chapter, alternative mechanisms unique to educational and achievement contexts are considered. The second part of the chapter explores why students may be more likely to approach an asynchrono us, online constructive controversy assignment with more individualistic and competitive and less cooperative perceptions than a face - to - face version of the same task, and describes the theoretical linkages between cooperative goal perceptions and positive student outcomes. Hypotheses and research questions are presented throughout chapter 2 and summarized in the subsequent chapter 3, which provides a detailed description of the participants, procedures, and materials, measures, and observations used in the study. Chapter 4 provides details on the flow of participants through the different steps and conditions along with the quantitative results and details on the analyses 6 employed. Finally, chapter 5 discusses the results in light of the hypotheses and rese arch potential relevance considered in the literature review . It also examines implications for theory, practice, and further research. 7 CHAPTER 2 : LITERATURE R EVIEW This chapter starts with a brief review of research on online learning and asynchronous course discussions as the context in which cooperative activities while very successful in traditional settings seem to lose momentum. Additional details on c onstructive controversy are also provided. In front of this background the chapter lays out, in turn, some of the pertinent theory and research on social - norms - based behavior modification on one hand and on cooperation - relevant partner behavior in online d iscussions on the other. Theories guiding the expectations of effects of social norms and partner behaviors include the theory of normative social behavior (Rimal and Real, 2003; 2005), the big - fish - little - pond effect (Marsh & Parker, 1984; Marsh, 1987); s ocial interdependence theory (Deutsch, 1949; Johnson & Johnson, 2005) and socio - cognitive conflict theory (Doise & Mugny, 1984 ). Growth and Challenges of Online Learning Even if the formerly exponential growth in enrollment is leveling off, online courses are widespread and here to stay: more than a third of students in U.S. higher education have taken at least one course o nline; online enrollment growth far exceeded that of overall higher education ; and over seventy percent of academic leaders consider onl ine learning as critical to their institutions long - term strategy (Allen & Seaman, 2015). Accordingly, research into the unique strengths and limitations of online course delivery as well as design research geared at improving learning outcomes online abou nds. As mentioned in the introduction, a perceived lack of community and diminished social presence (i.e., degree of awareness of the other person/persons in the course) and their detrimental effects on learning (e.g., Stodel, Thompson, & MacDonald, 2006; Gunawardena, 1995) are common concerns in the literature. As a result of diminished 8 - to - face course, ilities and personalities, and their own fit with the course, which can make them defensive and unwilling to take risks (Rourke, Anderson, Garrison, & Archer, 2001; Sammons, 2007) and inhibit collaborative knowledge building ( Kreijns, Kirschner, Jochems, & van Buuren, 2004; Slagter & Bishop, 2012) . Such uncertainty, coupled with the often criticized lack of community and support felt in online courses, contributes to the up to over 20 percent higher dropout rates observed in online learning (Carr, 2000; Pat terson & McFadden, 2009) as well. Online Learning via Asynchronous Course Discussions Guided text - based, asynchronous discussions, be they course - wide or small - group based, are frequently a key element of online courses (Chen & Chiu, 2008; Guzdial & Turns , 2000; Kim social presence via elaborate technological solutions involving audio, video and real - time interaction (e.g. Aragon, 2010; McLellagn, 1999), asynchronous, text - based course discussions are unlikely to become obsolete given their unique advantages such as high flexibility and low technological demands that make them suitable for learners across different time zones, conflicting schedules, and varying degree s of technological savvy and equipment (De Wever, Schellens, Valcke, & Van Keer, 2005; Jaggars, 2014; Tallent - Runnels et al., 2006). Additionally, their asynchronous nature provides advantages for those course discussions that benefit from reflection and in formation processing (Harasim, 1993; Pena - Shaff & Nicholls, 2004). If sustained sufficiently long among mutually supportive learners, they hold the potential to promote critical thinking and social construction of knowledge (Garrison, Anderson, Archer, 200 0; McLoughlin & Luca, 2002; Ruberg, Moore, & Taylor, 1996), tapping into a long tradition 9 cognition as internalized dialogue. Especially the deliberate and constru ctive consideration of multiple viewpoints regarding an issue is particularly conducive to both individual and group learning (Johnson, Johnson, & Tjosvold, 2000). For these reasons, asynchronous course discussions are likely to continue playing a large ro le in online teaching and learning (e.g., Jeong & Joung, 2007). However, maybe not surprisingly, online course discussions often fall short of high pedagogical aspirations. Instructors commonly observe a rather mindless adherence to the required posting mi post, making instructors wonder if that is all they have actually read of the discussion. According to Dav is and Rouzie (2002), reasoned deliberation is often sorely missing instead, one finds short threads - yet - respectful discourse instructors would like to stimulat e. In their case study on using Toulminian virtually non - Numerous computer - supported collaborative learning (CSCL) and argumentation - specific computer - supported collaborative argumentation (CSCA) systems have been developed to support learning in virtual groups with technical solutions, such as visual guides, note starters, and other scaffolds, as described further below. Yet many questions remain regarding the sociocognitive mechanisms that can undermine and foster productive online cooperative learning. 10 Constructive Controversy, Face to Face and Online As a p articularly well - established cooperative learning technique, constructive controversy (Johnson & Johnson, 2009) lends itself as an exemplar to be studied in the online environment. Its thoroughly validated beneficial learning outcomes include greater achie vement, higher level reasoning, more accurate perspective taking, greater achievement motivation, more positive attitudes toward the task, and more positive attitudes toward individuals holding conflicting positions, surpassing individual learning approach es as well as similar cooperative and debate - based techniques (Johnson & Johnson, 2009). The five - step scaffolded procedure involves deliberate discussion of two sides of a controversial issue related to class content, such as the question whether civil d isobedience is constructive or destructive . Students a re assigned either the Pro - or Con position to research and defend while interacting with another student tasked to argue for the opposing view (Johnson, Johnson, & Tjosvold, 2000). After presenting the ir best opening arguments and engaging in open discussion (steps 1 - 3), advocacy roles are reversed in step 4, followed by the task of drafting a mutually agreeable consensus statement that represents the most important points from both perspectives in step 5. The desirable learning outcomes produced by this learning procedure have been explained with the constructive nature of cooperatively framed intellectual conflict: Rather than sticking to an initial conclusion and supporting set of evidence, confrontat ion with evidence for the opposing viewpoint creates ambiguity in participants which in turn motivates a search for more information. Thanks to the switching of sides and the ongoing shared goal of achieving a well - thought - out consensus on the issue, the c ontroversy is a cooperative effort distinct from competitive debate formats , where one side wins , and from individual deliberations that do not demand accommodation of suggestions. Skills such as perspective taking, 11 creative problem solving, and a focus critiquing the quality of arguments, rather than the people making those arguments, are fostered (Johnson et al., 2000). And yet, taking this tried - and - true educational tool and placing it into an asynchronous online discussion environment und ermines its beneficial effects drastically (Roseth, Saltarelli, & Glass, 2011). Holding the instructional treatment constant via scripted constructive controversies for students to engage in, Roseth and colleagues (2011) isolated effects of (a)synchroni cit y and various delivery modes an on student outcomes and found that while synchronous computer - mediated communication via video - or audio chats seemed to support the motivational and asynch ronous CMC (via reciprocal voicemails, videologs, or most relevant for frequent online educational practice text - based discussions) was associated with reduced cooperative perceptions, motivational losses, and lower achievement. Various explanations for these losses are plausible: The logistical burden placed upon students who need to wait for a partner to respond is larger than for synchronous controversy asynchronous online course discussions losing momentum when feedback is not received instantaneously or never at all (e.g. Weaver & Albion, 2005) is a prominent concern in the online learning literature. It is also plausible that a controver sy that is per ceived as mostly constructive in the classroom ends up being perceived as uncomfortably antagonistic given that the exclusive reliance on written communication alone can lead to anxiety, defensiveness, and misinterpretations due to the absence of what psyc hiatrist Edward Hallowell (1999) has referred describes similarly disruptive effects, and the CMC literature is rich in additional mechanisms 12 that may contribute to a cooperative learning activity becoming less engaging. Some of the r elevant findings are explored in the second part of the chapter. (2014) findings of a reduced sense of cooperation and an increased sense of competition in the as ynchronous CMC conditions compared to the face - to - face condition are in line with such disruptive effects. Their experimental study of perceived belongingness, revisited further below, directly addressed some aspects of this underlying uncertainty as a mec hanism: An experimental induction of focus on the issue at hand, and increas ed motivation compared t o the control conditions. This finding supports - determination theory, which postulates that feelings of belonging and social connection, along with feelings of competence and autonomy, are prerequisites to self - regulation, intrinsic motivation, and well - being. Heeding the previously cited call to examine how constructive mechanisms involved in cooperative learning activities such as constructive controversy need to be addresse d differently online than in the original face - to - face version, the next section will consider relevant research regarding the role of other online course members beyond the dyad on individual course research are considered: First, relevant interventions and theory explicating the many contingencies involved the effects of group - level norms on individual behavior are explored . Second, and focusing back on the dyadic level, specific cooperative partner messages that are likely to affect the sociocognitive mechanisms of interest as mediators of motivation, relational quality, task engagement and achievement outcomes are discussed . 13 Distal Social Influence: The Role of Course - wide Social N orms M essages social persuasion reveals that one of the factors people use in making behavioral decisions pertains to their assessment as to whether others also engage in the beh Again, if dyads or small groups of students are engaging in cooperative learning activities like constructive controversy in a traditional classroom setting, the activity of neighboring dyads can be overheard, the engagement of the whole c lass gauged with a cursory look across the room. Students and their learning partners are there for the duration of the class with limited access to working - mediated scenario, not only are students exposed to numerous desirable alternative behavior options both online (e.g. surfing the web, attending to social news on Facebook, online sh opping, (e.g. Rosen, Carrier, & Cheever, 2013; Wang & Tchernev, 2012 ) and offline, such as taking a call, getting something to eat or drink, taking a nap, or indulge in similar typical desires ( Reinecke, Hofmann, & Klein, 2014 ), they also have less inform how they compare to the rest of the group in their engagement. Using behavioral prevalence information to impact behavior: Descriptive norms interventions Humans are social beings and care about what other behaviors provides clues for forming accurate perceptions of reality and helps develop and preserve social relationships and a favorable self - concept (Cialdini & Goldstein, 2004; Deutsch & Gerard, 1955). A large body of research exists on attempts to deliberately provide information about the behavioral prevalence of a certain target behavior in a specific reference group to 14 encourage individuals to conform to desirable behaviors , such as conserving water or electricity or, for experimental reasons, undesirable ones , such as littering . Conceptually, s uch normal p.1015, emphasis in the original) and can be distinguished from the other com monly descri bed variety of social influence, . The latter refers to in other words, what ought to be done. This distinction of social influence into descr iptive and injunctive elements is based on Deutsch informational social influence and normative social influence, with the former relying on what Cialdini (1988, p. 114) had famously described as a decisional shortcut : Unless manipulation such as that used by some salespersons is at play , the actions of the majority around us provide helpful evidence on. However, g iven the complexities of the many factors involved, results on the effects of strategically providing such behavioral prevalence information have been mixed, sometimes indicating strong conformity effects, sometimes showing no effects, and s ometimes showing effects opposite to the intended direction of influence . For example, an early experiment manipulating the prevalence of littering in the environment produced a three - fold likelihood of littering when this behavior was presented as common versus uncommon, an effect that was visible confederate (Cialdini et al. , 199 1 recycling behaviors, finding a twenty percent increase in materials recycled compared to households only receiving generic recycling information on their doorhangers. More recently, a field experiment by Goldstein, Cialdini, and Griskevicius (2008) compared the effects of 15 traditional ap peals made toward hotel guests to reuse their towels for the sake of water st of these studies have immediate practical implications for the generation of messages to influence people in desirable ways, often based on the presumed mental rule of thumb employed by recipients Based on these field experiments that usually relied on actual behavior as their key dependent variable, one may expe ct potentially beneficial effects of exposing online learners engaged in a task at risk for incomplete participation with potentially conformity - bolstering information about the engagement of the rest of the class. However, as the next sections will show, this expectation is less than straightforward. The social norms approach behaviors is the social norms approach, a theory stating that individuals may orient their behavior on incorrect perceptions of attitudes and behaviors of other members of their social groups (Berkowitz, 2005). Interventions based on the SNA are often geared at health promotion and prevention and may, for example, attempt to correct individuals drug use on one hand and their underestimates of health - oriented attitudes and behaviors on the other . A sample intervention based on the SNA approach might their drinking to 2 - correcting the mistake n perception that most college undergraduates partake in excessive drinking on a regular basis (Baer & Carney, 16 1993). Such normative restructuring aims at changing behavior as a result of changed (Lapinski et al., 2007). Of course, the applicability of a SNA to a behavior as speci fic and potentially novel as question by the need for SNA message targets to actually have a (incorrect but somewhat firmly held) perception about its normativi ty to allow for its correction in beneficial directions. As Miller and Prentice (1996) lay out, perceived norms have three origins: (1) observable behavior behav ior along with social projection of such self - knowledge onto others. In the natural context of online courses using cooperative small group or partner activities on a regular basis, such perceptions may indeed form over time and potentially form in ways th at under - represent actual engagement, in which the SNA would be directly applicable here. However, it is equally plausible that students do not form such perceptions, since they may not be able to observe behavior beyond that of their immediate work partne r and may not be in direct commun ication with other students, or if their communication results in the transmission of incorrect prevalence beliefs (Lapinski & Rimal, 2005). In the experimental context used here, a small pilot test asked students engaged i n the activity to estimate the percentage of others who were keeping up with the daily sub - tasks. I n the absence of any information about the rest of the participants , such estimates of the percentage of participants who were staying on track with t he dail y posting requirements averaged slightly under sixty percent and varied widely, as was expected . A SNA inspired intervention using realistic percentages in this context ( see table 10 in the appendix f or the actual baseline posting patterns ) would boost t he perceived behavioral 17 prevalence for some but also lower it slightly for others, just as it did in Glynn, Huge, and importance getting a clear baseline measure of perc eptions in order to accurately assess effects on one hand and to avoid unintended consequences of descriptive norm manipulations on the other, as discussed further below. But even with more familiar behaviors that message targets are likely to have infla ted perceptions on, such as college d rinking (Borsari & Carey, 2003), the SNA is not a straightforward success story. While normative restructuring strategies may have an effect on descriptive norms, such descriptive norms do not necessarily translate into modified behavior (Lapinski et al., 2007). For example, despite claims to the effectiveness of alcohol - related SNA interventions (e.g . Berkowitz, 2004), other researchers have failed to find any effects on a range of behavioral measures (Wechsler, Neslon, Lee, Seibring, Lewis, & Keeling, 2003). As Lapinski and Rimal (2005) have pointed out, this does not mean that descriptive norms do not matter, but that the many interacting pieces such as duration of the influence, potential moderators of the relationsh ip between norms and behavior, and the nature of the target behavior itself need to be considered. Their formulation of the theory of normative social behavior, described below, takes a step toward formalizing such considerations. The theory of normative s ocial behavior The theory of normative social behavior (TNSB; Rimal and Real, 2003; 2005) notes that descriptive norms frequently do not have clear - cut main effects on behavior, but that several factors, some of which are described by the theory, moderate such effects. Recognizing the complexities of predicting behavior, four moderators that research has linked with normative influence are included in the theory: outcome expectations, injunctive norms, group 18 identification, and ego - involvement. Figure 1: Theory of normative social behavior (e.g. Rimal et al., 2007) Outcome expectations are conceptually defined as beliefs that a certain behavior will provide significant benefits to the self or is providing significant benefits to others engagi ng in it (Bandura, 1986); and that the benefits of engaging in the behavior outweigh the costs of doing so (Rogers, 1975). Supporting this moderator of the TNSB in the context of college drinking, Rimal (2008) found that both perceived bene fits and the des criptive norms - perceived benefits interaction term explained a significant increment in variance in alcohol - related behavioral intentions. Further support for the interactive effect of descriptive norms and perceived benefits comes from a study crossing experimentally induced high - versus - low benefits of a target behavior (practicing yoga) with experimentally induced high - versus - low prevalence information (Rimal, Lapinski, Cook, & Real, 2005). Results indicated that without the expectation of significant b behavior of interest. However, for participants led to believe that there were numerous benefits to the practice, high descriptive norms were associated with high beha vioral intentions, in line with Descriptive Norms Behavior Moderators: Outcome expectations - benefits to oneself - b enefits to others - anticipatory socialization Injunctive norms - social approval Group identity - similarity - aspiration Ego - involvement 19 research on opportunity loss aversion. While fear on missing out is not very likely to occur in the context of online constructive controversy, the impact of descriptive norms indicating that a large majority is participa plausibly moderated by outcome expectations more generally. Here, perceived benefits could take the form of learning something valuable, while the time and effort required to log on to the and to research and formulate the next argument, pose costs of participation. Rimal et al. (2005) describe this moderating influence as multiplicative, indicating that in the absenc e of one factor, the other will not have an effect on target behavior, while in case of high outcome expectations the effect of norms on behavior is expected to strengthen. Similarly, injunctive norms (i.e., the perception of social approval for compliance with the descriptively - normative behavior and social sanctions or disapproval for noncompliance), group identification (i.e., perceiving the sources of the descriptive norms as similar enough to oneself for those norms to be relevant) and ego - involvement (i.e., the extent to - from descriptive norms to compliant behavior. As described previously, injunctive norms are conceptualized as what ought to be done in a particular situation and, rather than informing action as descriptive norms do, enjoining it 5) . In the experimental reliable participation in the activity is of little relevance and interest to others or society at large, and that few if any social sanct ions are to be expected if participation was to be stopped . The 20 to be participation and disapprove of its absence. Most research on the distinct effect of injunctive norms and their possible interactive effects with descriptive norms on behavior (e.g. Rimal & Real, 2003) had treated injunctive norms as a global measure of approval for a behavior, leaving norms are likely to apply here. It should be noted that Lapinski and Rim al (2005) have categorized influence exerted by a single partner for a private behavior (e.g. condom use) as interpersonal, not normative. Injunctive violation as a contributor to injunctive norms, their impact in the present context are uncertain: After all, according to Ajzen and Fishbein (1980), a second necessary element of injunc tive normative given the anonymous nature of the interaction at least in the current experimental context, this motivation is likely to be somewhat limited. However, an which the normative information is based is a plausible moderator of normative e ffects in the present context. Group identity i s conceptually defined within the TNSB as a degree of affinity for or the desire to connect with a reference group (Lapinski & Rimal, 2005) . Although reference groups that are close to and similar to the target individual are likely to exert more influence than more distal ones, as will b e discussed below, r esearch on nominal grou p (1990) experiments testing the hypothesis that in - groups exert more influence than do out - groups 21 s t hat assign trivial group membership to participants (e.g. based on the tendency to overestimate versus underestimate the number of dots presented on a slide) have shown that similarity perceptions do not need to be elaborate in order to impact social proce sses (Brewer & Kramer, 1986; Tajfel, Billig, Bunday, & Flament, 1971) . A more important distinction may be rooted din e arly work on social influence by both Kell e y (1952) Shibutani (1955) , who specified a key prerequisite f or a group to function as a refer ence group for an individual: The group needs to be perceived in such a way that its typical behaviors, attitudes, and other characteristics represent (1955) suggested distinction between normative groups (whose norms and attitudes are already accepted by the individual) and aspired groups (groups that the individual would like to join or maintain membership in) on one hand and comparative groups on the other ( groups that do serve as reference points for comparisons, including comparisons about personal qualities) highlights that it is the latter group concept that is most appropriate to consider in terms of reference group effects and therefore particularly app licable to the study of descriptive social norms. This point will be revisited in the section on social comparisons in academic contexts and the big - fish - little - pond effect further below. While most previous work on the TNSB has conceptualized group identi the reference group (Rimal, 2008, p. 108) as a combination of aspiration (i.e., the extent to which participants sought to be like these members) and similarity (the extent to which particip ants perceived a lack of psychological distance from members of the reference group), the present context is different from their original context: Incoming freshmen are likely to have a sense of 22 participants in the activity of interest here are less likely to have spent much time thinking about how much they wish to be like their classmates, whereas an assessment regarding their similarity with them is more ac cessible. Therefore, the choice was made to assume the more stringent conceptual definition of reference groups as suggested by Kelly and Shibutani with its primary focus on the similarity component. Previous research on behavioral intentions regarding al cohol consumption, exercise, and the use of sun screen showed that descriptive norms were most influential for individuals who strongly identified with the group on which this information was based (Rimal & Real, 2003; Rimal, 2008; Terry & Hogg, 1996). Imp ortantly, the moderating effect of group identification does not only hold the potential to affect the strength, but even the direction of the relationship between perceived descriptive norms and behavioral intentions: For example, a study on voting intent ions produced the expected positive relationship between perceived level of student voting for high identifiers but a significant negative relationship for low identifiers (Glynn, Huge, & Lunney, 2009). The need to consider the effects of moderators in the context of norms effects becomes evident when norms effects might cancel each other out this way, producing potentially misleading null results. Similarly, different levels of ego involvement are likely to shape the effects behavioral prevalence informa tion is likely to have. According to Lapinski and Rimal (2005, p. 136), ego involvement as originally postulated by Johnson and Eagly (1989) referred to the extent to which an individual self - concept is connected with their attitudinal position on a part icular issue, but - - of their self - concept (Conner & Armitage, 1998) and are thus likely to be highly ego - involved in 23 norms is stronger for individuals whose self - identity is closely aligned with the enactment of the behavior or th ose who are highly ego - involved in a behavio Lapinski & Rimal, 2005, p. 136 ). of college drinking intentions found that students whose self - identity was connec ted to the indeed more likely to be influenced by descriptive norms of drinking. In a later study on student drinking, Rimal (2008 , p. 106 ) used the label of behavioral identity for the construct of ego involvement as applied to a behavior, highlighting the role of prior typical behavior in shaping and his or her individuals have of themselves, develop ed primarily through the enactment of particular behaviors. Again, in this study, the behavioral identity of interest relates to conscientious engagement in an asynchronous online controversy task, and relevant self - perceptions are those developed from re student on one hand as a skilled communicator in the face of conflict on the other . Students who view their performance in these roles as relevant to their identity are expected perform as responsi ble students and communication partner s in this context as well, in contrast to their peers for whom such behaviors may be less relevant to their identity. Beyond this main effect, an interaction effect with descriptive norms is expected on behavior (Rimal et al. 2004; Lapinski & Rimal, 2005) in that descriptive norms have a stronger effect on individuals with higher behavioral identity involvement. Rimal (2008) explained this interaction effect with the proposition 24 106). Such effects go beyond the shared variance between previous and subsequent behavior (Connor, Warren, Close, & Sparks , 1999, cited in Rimal, 2008). Again, i n the present context, identity as a reliable work partner, a conscientious student, or strong (online) communicator may interact with descriptive norms in similar ways in that high behavioral identity al ong these relevant lines makes learners more alert to the extent to which the other participants are engaging in the task, while individuals whose identity is not closely aligned with such types of performance are more likely to be unimpressed by what the rest of the group seems to be doing. Additional TNSB research has identified group orientation (Lapinski, Rimal, DeVries, & Lee, 2007) as another moderator, further highlighting the role of individual difference variables. In this study, the authors exp erimentally induced in participants a perception of having either over - or underestimated the actual prevalence of water conservation behaviors. They found that individuals with high individualism (i.e., low group - orientation) scores on a version of Triand less positive attitudes and weaker behavioral intentions toward water conservation when exposed to information about high prevalence of water conservation behaviors within their university community. Somewhat in parallel to the moderation patterns of similarity found by Glynn and colleagues (2009) mentioned above, this finding further highlights the complexity of the norms - behavior link that is unlikely to be captured by focusing exclusively on main effects. While group orientation is beyond the scope of the present study, the general logic of the l imits of main effects does apply : It is conceivable that well - intentioned interventions geared at fostering engagement could produce reduced participation in students with low group identification or relevant levels of other moderators, an example of unintended possible norms 25 effects as discussed further below. Further possible moderators of norms effects The theory of normative social behavior calls for openness to addit ional moderators of the effects of normative restruc turing on behavior, including the behavioral attributes that are likely to make normative influences more or less relevant (Lapinski & Rimal, 2005 ; Miller & Prentice, 1994 ). Beyond perceived benefits and costs, such behavioral attributes that may be relevant are the degree of voluntariness of a behavior; its public - versus private nature; whether the effects of the behavior affect the individual engaging in them, the immediate environment , or all of these; whether effects are short term or long - term and incremental; whether it is socially motivated or not; whether it is pleasant in nature, and whether i t is socially approved or not. Supporting the need to distinguish behaviora l attributes , Manning (2009) found greater effects of descriptive norms on behavioral attitudes for behaviors that are not socially approved (e.g. drinking alcohol versus exercising), more pleasant, and for behaviors that were serving social needs, as coll ege drinking often does. Additional candidates for moderators of norms effects can be found in the nature of the reference group on which the descriptive norms are based: proximity to the message recipient is a factor that has likely contributed to mix ed results in research of descriptive norms effects. Borsari and Carey (2003, p. 332) illustrate the importance to consider proximity in the context of r member similarity, these very diverse reference groups differ in terms of familiarity and specificity to the participant, with the potential for normative misperceptions incre asing the less proximity there is. As Glynn and colleagues (2009) conjectured in the context of descriptive norms effects on 26 power that make compliance with the n orm more likely than descriptive norms based on a more distant, general reference group. Confirmatory factor analyses on participant responses regarding organ donation intentions and drinking behaviors have further confirmed the conceptual distinctiveness of personal and societal - level descriptive norms (Park & Smith, 2007; Park, Klein, Smith, & Martell, 2009). Further, i ndividual differences have been mentioned in the context of group orientation and behavioral identification above, and other authors hav e suggested attention to social comparison information (ASCI) as a possible factor (Yun & Silk, 2011). Finally , the context in which the message is communicated is frequently mentioned as a boundary condition for norms effects: If the context is marked by ambiguity, e.g. in the case of for clues about the nature of social reality (Cialdini, 2001; Festinger, 1954; Rimal & Real, 2005). The reliable and timel y engagement in the required subtasks of cooperative learning activities like constructive controversy within an online course as a target behavior is not a completely clear - cut case for classification along the identified attributes. For example, such eng agement is non - voluntary if the activity is required and graded, but learners have some latitude regarding the promptness and quality of their involvement in the activity to accommodate their grade aspirations, personal interest in the course, and other fa ctors , giving it some voluntary characteristics that might make the behavior amenable to descriptive normative influences. Likewise, the behavior is neither completely public nor completely private: In the context of the experimental procedure, participant s are anonymous to each other, but unlike in the case of, say, voting, there is another person who reads and responds to the behavior who may 27 or may not know the online peer through the shared course, and presumably there is the instructor who follows indi vidual group work to assist and potentially assess not only product but process as well. Similarly, as mentioned previously, the behavior does not have the potential for a very strong effect on the individual engaging in it (unlike, for example, health - ris k or prevention behaviors do), nor does it clearly relate to the greater good (since the eventual goal of constructive controversy and related activities to generate a better citizenry is unlikely to be perceived as such by the individual participant in th e moment, whereas the long - term purpose of recycling or energy conservation is more clear). But it is unique in that as a cooperative undertaking, it affects the assigned partner considerably in their ability to get engaged in the task to the extent they w ant. Similarly, most participants and, to some extent, online students engaged in this kind of task within a course , are likely to perceive effects as short term (earn ing the available points for the assignment and mov ing on) rather than as the incremental , long - term development of a more constructive approach to controversies, large and small, that was the underlying motivation for development of constructive controversy and cooperative learning more generally and the reason why instructors may wish to inc lude it in their courses. Even the degree of pleasantness cannot clearly be determined, since open - ended student feedback about the activity span s a broad range from complete boredom and annoyance to appreciation of gaining a broadened perspective. Therefo re, evidence that more pleasant activities tend to be more amenable to influence via descriptive norms (Manning, 2009; Rivis & Sheeran, 2003) cannot lead to specified predictions regarding the likely influence of such norms here. Similarly, the extent to w hich an activity is serving social needs, which tends to strengthen the effect of descriptive norms, is not easy to apply here. With all likelihood, most undergraduate students in the experimental scenario here would not consider interactions with an assig ned partner as 28 particularly relevant for their social lives. However, for geographically dispersed, fully online learners, that might be somewhat different. Finally and regarding exercising despite the low risk of s ocial sanctions for noncompliance in the experimental context, social approval for engaging in an educational task can be assumed to be fairly high, which may suggest a weakened effect of descriptive norms compared to socially less approved activities such as drinking, where descriptive norms tend to play a larger role. Descriptive norms can have unintended effects It should be noted that descriptive norms can have potentially counterproductive effects, caused largely via two mechanisms: psychological rea ctance and the potential for descriptive norms to actually point away from the desired behavior. The first mechanism, psychological reactance (Brehm, 1966; Brehm & Brehm, 1981), generally refers to the observation that individuals can be motivated to res ist or act counter to attempted social influence. To the extent that information about the prevalence of a supposedly decisions without such influence, it is pos sible that purposely not going along with the crowd serves to reassert that freedom. A few of the interaction effects mentioned above are directly consistent with the reactance mechanism , as when individuals with strong individualistic orient ations (Lapi nski et al., 2007) and individuals who perceived low similarity with the reference group from which descriptive nor ms stemmed (Glynn et al., 2009) actually reduce, rather than heighten, their behavioral attitudes and intentions to move away from, rather th an converge with, the perceived descriptive norm. Lapinski et al . engage in a behavior also means, of course, believing that the number of people who do not 29 engage in the behavior is small. It may well be that individualistically oriented people identified with this minority (or dis - identified with the majority) and thus expressed attitudes and The second mechanism is even more straightforward: Descriptive social norm interventions can backfire in cases when the perceived descriptive norm is actually not to perform the desired behavior (Goldstein et al., 2007; Cialdini et al., 2006). Thus, if the actual behavioral prevalence of desired behavior is not exactly desirable (e.g. there are low response rates, or heavy reliance on last - minute - minimal posting in online course discussion boards), normative information used in persuasive appeals can be counterproductive if it (a) draw s attention to the undesirable behavior and (b) normalizes the undesirable behavior. For example, visitors of a national park encountering signs highlighting removed the petrified wood from the park, changing the state of Demaine, Sagarin, Barrett, Rhoads , & Winter, 2006; p. 8) were found to focus less on the undesirability of such behavior than on its normalness. Especially when such a message is worded in negative terms (compared to the equivalent but less attention - Baumeister, Bratslavsky, Finkenauer, & Vohs, 2001, r ), visitors would follow the dysfunctional descriptive norm the most and engage in the highest levels of petrified wood theft. The use of descriptive norms regarding the desirable so m ewhat less salient behavior (i.e instructors online may recognize highlighting the negative (salient) normative behaviors in well - 30 mea about low voter turnout have been shown to produce such counterproductive behavioral energy conservation supports this concern by showing that individuals who actually are behaving in a relatively desirable way can end up reducing suc h desirable behavior to converge toward the less desirable typical patterns they are being informed about. Guarding against unintended norms effects Such counterproductive effects need to be guarded against to some extent . For one, the descriptive norm ca n be taken from a specific subgroup that is exhibiting a desirable norm, such a specific department in an organization in which a desired practice has spread more rapidly than in others . For example, in Schot (2011) social norms intervention the target b ehavior of updating employee profile page was more eagerly adopted by one department that could be used as a baseline for others. Secondly, actual numbers can be presented in a way that makes the desired behavior appear as frequent (and hence norm ative) as possible, for example by focusing on the growth rate of readership of the blog one is trying to popularize rather than the actual number s of readers, if those are less impressive (Martin, Goldstein, & Cialdini, 2014). Th ird, in cases where information about the less - than - ideal prevalence of a target behavior induces individuals who were originally engaging in it to revert to a less desirable pattern (Gerber & Rogers, 2009; Schultz et al., 2007), the addition of personaliz ed injunctive energy conservation, the simple addition of a drawn smiley face on door hangers indicating that a 31 elow average effectively prevented the leading energy savers from stooping down to the level of their energy - hogging neighbors. Finally, aside from such applied studies geared at correcting perceived norms that are or the environment, basic research along the lines of the TNSB (e.g. Rimal, 2008) and focus theory (e.g. Cialdini et al., 1990 ) geared at exploring and testing mechanisms of normative influence without immediate practical application has relied on mild to severe over - or underrepresentation of norms. In experimental tests of the TNSB moderator effects, for example, descriptive norms are manipulated via fairly drastic versus high descriptive norms) to be crossed with different levels of the moderator of interest, providing taking steps to conserve water (Lapinski et al., 20 07), or that the practice of yoga had been on the (1990) often cited e xperiment a rtific ially created a very low behavioral prevalence impression regarding littering in public in unwitting participants (by cleaning the parking lot they would walk through of all litter) versus a participants would walk across it). Of course, for the formerly mentioned kinds of studies that do not focus on disentangling the moder ating mechanisms of descriptive norms effects for theory refinement, but where the goal lies in testing the feasibility of using truthful information about a desirable behavior, any deception is impractical. After all, repeated e xposure to 32 pronouncements of large majorities of people recycling their recyclables, conserving energy and water, paying their taxes, voting in elections, and making healthy choices that turn out to be fabricated would ultimately undermine the very mechani sms on which such potentially beneficial conforming motivations are buil t perceptions, decisions, and behaviors, and the motivation to create and maintain positive social relationships and gain a sense of bel onging (Cialdini, Martin, & Goldstein, 2015; Deutsch & Gerard, 1955; Martin et al., 2014). Therefore, for such interventions, unlike the experimental context, deception regarding the actual behavioral prevalence is usually avoided. However, precedents for studies in between these two clear - cut varieties exist. For example, guided by an interest in the potential of high descriptive norms to enhance the effectiveness of an information intervention regarding the importance of sunscreen use among Californian c d yet like in the present study, no definitive baseline norm data generated on a representative sample of the population existed, was based on pilot work that had indicated a prevalence of 60 - 78% of students claiming to use sunscreen at least 70% of the time. As such, the inflation of the percentage to more clearly depict a majority served to strengthen the informational value of the information of sunscreen us e as being clearly the sensible and also right thing to do. Like Mahler et a similarly interested in ensuring that presented norms are unlikely to lower descriptive norms for a message recipient. Therefore, rather than usin g the precise actual completion percentages per step and day, the percentages presented reflect ed 33 idealized percentages in terms of t heir punctuality: S tarting in the afternoon of the day th at a particular activity step was due (e.g. Step 3 on Day 3, with a suggested posting deadline of late condition receive d messages stating that between 70 and 85% of participants had completed this step. As table 10 illustrates, thi s percentage range corresponded with the actual range of completion percentages by the end of the next morning , as suggested by the pilot study and replicated by completion tracking for the third cohort in the present study. Pulling the percentage forward to the time frame during which it would have been reached if participants followed the instructions closely was geared at creating functional and guarding against dysfunctional descriptive norms without being fundamentally deceptive. Messages were clea rly displayed on activity website, and the precise number was randomly varied by the website to appear natural. Previous experimental work has shown that such normative messages can influence - norms experimental induction resulted in a significant drop in estimates of the prevalence of alcohol consumption. Lapinski et a conservation behaviors by suggesting that either 90% or 3% of people within the university community engaged in such behaviors found this norm induction was effective, with significantly higher prev alence estimates generated by participants in the high - prevalence condition (M=5.06, SD=1.56) than by those in the low - prevalence condition (M=2.25, SD=1.22). In a field experimental setting with voting as the target behavior, phone messages highlighting h igh - versus low expected turnout let to significantly higher (60.2%) versus lower (47.7%) estimates of reported turnout expectations (Gerber & R ogers, 2009). Similarly, 34 (2008) less drastic induction also led to a significant condition effec t in that participants who had received the descriptive norms on top of the basic information provided to all groups provided substantially higher estimates (81.9% versus 29.2%) of the percentage of college students regularly engaging in the target behavio r. To assess the effectiveness of this descriptive - norm induction approach in the current context , the first hypothesis is : Hyp1 a : Participants exposed to course - wide descriptive norms messages regarding the completion of constructive controversy subtasks will report higher perceived descriptive norms than participants not exposed to such messages. To the extent that the group exposed to such normative information reports higher perceived descriptive norms, a higher completion rate overall and more tim ely and reliable response behavior per step would be expected based on - analytical findings of a small but distinct effect of descriptive norms on behavioral intentions, with five percent of variance explained by descrip tive norms after taking into account theoretically older explanatory factors such as attitude toward the behavior , inducements by important others, and perceived behavioral control (Ajzen, 1991, as cited in Rivis & Sheeran, 2003 ). Similarly supportive of b ehavioral effects of descriptive norms is the rich body of field - experimental work done by e.g. Cialdini and colleagues mentioned at the very beginning of this section that depicts the often subconscious but poten t influence of - analysis noting a direct association of descriptive norms and behavior ranging from .11 to .35 depending on the level of relevant moderators such as usefulness and social motivation . However, factorial hypo theses regarding student engagement behavior in the context of online constructive controversy are posite d with more exploratory intentions than the former , g iven the opportunities for the described moderators 35 to counteract such effects on one hand and the presence of alternative accounts of norms effects through additional mechanisms germane to social comparisons in educational contexts (as briefly discussed in the next section) on the other . Hyp1 b, c : Participants exposed to course - wide descriptive norms messages regarding the completion of constructive controversy subtasks will have higher completion rates ( 1 b ) and post their responses in a more timely fashion ( 1 c ) than participants not exposed to such messages. Since the possible association of perc eived descriptive norms and behavioral outcomes is of interest even if the norms messages are unrelated to behavior, correlational hypotheses are posed as well for the present context, a - analysis of descriptive associ ations between perceived norms and behavior : Hyp2 a : Perceived descriptive norms are positively associated with completion rates. Hyp2 b : Perceived descriptive norms are positively associated with timely responses. As previously discussed, given the relev ance of group identification which has been shown to moderate the effect of descriptive norms on behavioral attitudes and intentions ( Rimal & Real, 2003; Rimal, 2008; Terry & Hogg, 1996 ), along with t he practical feasibility of cation with the reference group on one hand and the potential malleability of group identification in an online learning setting on the other , the following moderation hypothesis is proposed as well. It has similarly exploratory intentions given that some studies testing this aspect of the TNSB failed to confirm an interaction effect between descriptive norms and group identification (e.g. Rimal & Real, 2005) . Following Rimal and his colleagues, group identification is conceptualized as the extent to which themselves as similar to the referents on which the normative information is based. Hyp3 a : The effect of perceived descriptive norms on completion rate and/or timeliness of 36 responses is moderated by group identification in that the h group, the stronger the association between DNs and behavior . Finally, - related perceptions, moderation effects had been found for descriptive norms as discussed above ( Rimal, Real, & Morrison , 2004 ; Rimal, 2008) . In line with such previous work, ego - involvement is conceptualized as self - concept is connected with a particular behavior here, leading to the following hypothesis : Hyp3 b : The effect of perceived d escriptive norms on completion rate and/or timeliness of responses is moderated by ego - involvement - related ego involvement, the stronger the association between DNs and behavior. An alternative mechanism for effects of d escriptive norms in educational contexts: The big - fish - little - pond effect (BFLPE) B efore shifting the focus from the course - wide, normative influence to the intra - dyadic, peer based influence effects of interest, a look at a different literature further in forms the issue of possible effects of course - wide performance related updates . It is mentioned here after the first set of hypotheses since the existing research evidence is not directly geared at engagement variables but more closely related to variables discussed in the intra - dyadic influence section . H worth of education al psychology research on the so - called big - fish - little - pond effect (BFLPE, Marsh & Parker, 1984; Marsh, 1987). Classrooms provide a somewhat distinct context from some of the contexts in which norms research as described above has occurred : H ere, social c - concepts (e.g. Suls, & Wheeler, 2013) than for example neighborhood - based campaigns on recycling or energy conservation. 37 The BFLP individual achievement is positively linked to his or her academic self - concept, i.e., perceptions held about themselves regarding either their general academic or subject - specific abilities , cumulative achievement information about the school or class is self - concept (Marsh, Trautwein, Lüdtke , & Köller , 2008) 54) social comparison theory, the BFLPE describes how students in an academic environment with low achievement norms (i.e., small p onds, figuratively speaking) tend to evaluate themselves more theory, lead to higher self - evaluation. Vice versa, students of the same ability level placed in an academic environment marked by high achievement tend to adjust their academic self concept downward due to upward comparisons with more highly performing class - or schoolmates a small - fish - b ig - pond effect, so to speak. These consistently found effects o ccurred in many different cultural cont exts (Marsh & Hau, 2003) and have important implications as future academic choices, academic effort, and a host of subsequent achievement outcomes are linked to academic self - concept (Marsh et al., 2008). Many of t he features that mark the BFLPE and set it apart from more constructive self - other comparison processes, as when students use targeted, voluntary, and inspiring comparisons to motivate themselves ( Huguet, Dumas, Marsh, Régner, Wheeler, Suls , et al., 2009) are applicable to the presentation of completion percentages for daily sub - tasks in online constructive controversy. First, as Huguet and colleagues (2009) describe, assumed to result from forced p. 157, emphasis in the original), be it through the publication of grade distributions or the presentation of completion rates . Second, it is based on unfamiliar comparisons, i.e., comparisons with a 38 nce level of other students in the same - to - face classmates. Third, comparison ts are not explicitly instructed three attributes are present in the context of subtask completion percentages in o nline constructive controversy, suggesting that to the extent that the resulting comparisons are upward comparisons (which are, of course, the very comparisons that social norms interventions would have the highest hopes for, for fostering participation an d timeliness as lagging individuals catch up to the standards) might be counterproductive in other ways. Of course, academic self - concept is not a dependent variable of direct interest in this study, given that a single educational activity is unlikely t o affect an individual difference variable that builds over time. However, extensions of BFLPE research have considered the extent to which student interest in a particular area such as math can be affected by similar contrast effects to those observed wit h regards to self - concept, and found in a large representative sample of German ninth graders that as expected, higher performance at the class Lüdtke, Marsh, Köller, & Baumert, 2006). Such reductions in interest, like those in academic self - concept, are of course tied to a host of relevant educational constructs such as positive affect, persistence, perceived competence, and favorable learning outcomes (e.g. Ec cles & Wigfield, 2002), strengthening the case for a careful assessment of the risks of upward comparisons. In a test of the generalizability of the BFLP effect to other academic self - belief constructs that are either more likely to invoke a normative, soc ial comparison - based frame of 39 reference or more likely to invo ke an absolute frame of reference, Marsh and colleagues (2008) found support for their prediction that the BFLPE would generalize particularly to those constructs that tend to invoke social comp arison s . PISA data on 4000 fifteen - year old students confirmed the expectation that perceived self - efficacy and perceptions of control (over academic outcomes) were most affected by school average achievement data, while constructs more within the individu strategies used and effort persistence, were not statistically related to the performance level of others. However, given the indirect effects of constructs that are af fected, such as self - concept and interest, on subsequent motivation (Eccles & Wigfield. 2002; Trautwein et al., 2006), the risk of unintentional effects of high - achievement norms information is somethi ng to be kept in mind , both in the traditional FtF cont ext in which most of this research is based and in the ambiguity - enhancing (e.g. Mäkitalo, Weinberger, Häkkinen, Järvelä, & Fischer, 2005 ; Valaitis, Sword, Jones, & Hodges, 2005) online context, some key features of which are the focus of the next section of this chapter. Proximal Social Influence: The Role of Cooperative P artner M essages A second large area of interest regarding social influence in online cooperative learning relates to in fluence processes occurring within the dyad engaging in a task suc h as cooperative controversy . Thus, shifti ng from a consideration of course - based influence with a relatively unspecific, numer ic message to p artner - based influence processes , the next section will examine theory and research on how online learners engaged in cooperative learning activities are able to influence each other in more - and less productive ways. After laying out a background of research on the effects of asynchronicity and text - based interactions, two theories linking f the task to motivational and learning outcomes are introduced, along 40 with two additional sets of hypotheses regarding the effects of cooperation - promoting partner behaviors on such outcomes. Online argumentation as a situation at risk for individualistic and competitive goal construction The relative uncertainty regarding other course members activities discussed above extends to intra - dyad dynamics, operating here at a different level: learners may feel uncertainty about their partner in terms of their personality, ability, motivation, and their desire to constructively work together (e.g. Hsu, Ju, Yen, & Chang, 2007 ; Mäkitalo et al., 2005). As Nam group cohesion and relationships between members of online learning environments ar e more fragile than those of face - to - face learning environments because... there is no guarantee that other group members will behave as learners are expected to do (p. 237). Again, just as students are unlikely to abandon their interaction partner prior to task completion in a face - to - face version of constructiv e controversy, they are also less likely to ignore the need for social buffering of the argumentation task while working on it jointly in the classroom. From its beginning, the central concern of r esearch on computer - mediated communication in organizations and educational settings has been the absence of certain nonverbal and vocal cues that can reduce the quality of communication (summarized as the perspective, e.g. Short, Willi ams, & Christie, 1976; Sproull & Kiesler, 1986; Hallowell, 1999) along with the need and ability of interaction partners to effectively compensate for these shortcomings ( ues filtered in perspective, e.g. Walther, 1994; Angeli, Bonk, & Hara, 1998). While there has been much evidence for the cues - filtered - in - that students spend considerable portions of their exchanges on social interchange with the potential of fostering collegial bonds (e.g. McDonald, 1998; Kanuka & Anderson, 1998 ; 41 Gunawerdana & Zittle, 1997) in the case of open - ended online discussions, the picture is complicated by the introduction of more highly structured cooperative tasks designed to introduce constructive conflict into the online discussion and to guide a del iberate juxtaposition of viewpoints in hopes of improving learning outcomes. S hifting online learners focus on specific task - related activities in online learning environments can contribute to a neglect of important social aspects of cooperative learning , which is problematic given the that a sense of cohesiveness and feelings of trust and belonging serve as prerequisites for desirable learning outcom es of online argumentation (e.g. Davis & Rouzie, 2002; Kreijns , Kirschner, & Jochems, 2003; Nardi & Whitta ker, 2002; Northrup, 2001; Rovai, 2001) . Examining an intervention addressing such cognitive - socioemotional tradeoffs, Nam (2014) found that training online learners in the premises of con structive controversy (such as defining controversy as a problem - sol ving situation, exploring all different points of view, negotiating a resolution) was associated with lower achievement, less positive attitudes toward cooperative learning, and particularly openness and sharing (Johnson & Johnson, 1994) than training learners in m ore socio - emotionally oriented trust - focused behaviors. The latter intervention consisted of focusing learners on the need to openly share their input, communicate respectfully, express their c onfidence in the group to handle diversity and complexity efficiently, and to express their cooperative intentions. Scaffolding o nline c ourse i nteractions In response to the shortcomings of structured online argumentation activities in terms of vital socio emotional components as described , e.g., by Nam (2014), Kim (2009) made the case - emotionally enhanced task - type of interaction scaffolding is designed to complement typical epistemic argumentation 42 scaffolds as found in the broad range of available computer - supported collaborative argumentation platforms, such as Sc C omputer - supported Intentional Learning Environment (CSILE, Scardamalia & Bereiter, 1996) , B Knowledge Integration Environment (KIE; Hoadley, 2000) , and other approaches used in scaffolding studies ( e.g. Weinberger, Fischer, & Mandl, 2001). Kim (2009) encouraged learners in the experimental group to utilize both cognitive - and soci allyoriented prompts via drop - down menus available during message construction, while two control groups were presented with cognitive - only or no prompts, respectively. Sample cognitive prompts used were: Proposing initial argument, proposing counterargume nt, challenging, counterchallenging, and integrating, mirroring some of the formal steps of construc tive controversy. By contrast, socio - cognitive prompts designed to support both cognitive and social dimensions of interaction (p. 90). Examples incl ude d and presenting challenges politely (embedde d in points of agreement, I ). Dyads working with the socio - cognitive prompts produced more agreements than the control groups, along with slightly improved feelings of community. The link between socioemotionally enhanced behaviors and agreement lends further support to the reverse observation of tasks designed for disagre ement being at risk of engendering interaction patterns that are little conducive to feelings of cooperation, solidarity , By extension, as the theories presented in the next section will show, this is expected to have negative implications for motivational outcomes and achievement. - cognitive language had hig her category in assessments of the final products (a joint statement much 43 lik e the one required by constructive controversy), lending additional support to the value of socioemotional discussion enhancements in this context. However, there is evidence that there is such a thing as over - scaffolding learner interactions, as indicate d by reduced uptake of message starter suggestions in more structured cooperative online discussions (Bures, Abrami, & Schmid 2010). These authors also report that students preferred using their own language to fulfill cognitive - or sociocognitive function s, as opposed to having language inserted into their writing. Such observations make a case for training interventions that create an awareness of the need to compensate for lost cooperative perceptions and tone in structured argumentation tasks online, le aving it up to the student to carry out such compensati on in their own language. Plus, if students make dutiful use of the note starters and scaffolds provided, there is a risk of mindless overreliance upon them (Weinberger et al., 2001). A final, practica lly motivated reason to consider the effects of modeled behavior rather than the reliance on message scaffolds is the portability of such self - generated behaviors from course to course and beyond, independent of the course platform used and its specific di scussion board features. Early work on identifying peer behaviors that support the quality and sustainability of asynchronous online course discussions (e.g. Burge, 1994; Gunawardena, 1995) e), response (i.e., constructive - friendly messages. An experimental ass essment of o ne implementation of such an rules - based approach prescribed (and graded students on adherence to) a set of theoretically derived rules geared at deliberately remediating some of the potential problems encountered by virtual groups (Walther 44 & B unz, 2005). These rules included overtly acknowledging having read one to deadlines for accomplishing subtasks in order to reduce uncertainty that group members may feel given that the asynchronous online environment can create reduced perceived accountability. As hypothesized, self - was closely associated with trust in the grou p and , with correlations between .40 and .65. Achievement, operationalized by coding the , was significantly related to rule adherence as well, with correlations around .30. One online behavior of particular interest in the context of conducive group interactions were greetings, which strongly predicted adherence to particularly beneficial rules such as acknowledging others and being explicit. Walther and Bunz (2005) m that saying hi matters ( ) Indeed, in line with findings by Bunz and Campbell (2004), CMC partners appeared quite sensitive to (p. 842). This is in line with suggestions made by socio - emotional scaffo lding research , making greetings a promising candidate for inclusion in online constructive controversy interactions as well. Similarly , c orrelational studies of naturally - occurring interactions provide further support for observational study of an asynchronous online argumentation context, a more conversational communication style ng greetings, questions, acknowledgements, and agreements) was associated with more sustained argumen tation, and in particular with constructive argumentative moves such as argument challenge , challenge explanation , and challenge counterchallenge . Ag ain, the need for trust 45 and a feeling of cooperation seems particularly pressing when online course discussions are purposely designed in controversy form. Finally , a pilot study using the controversy website and instructions proposed for use here with 14 natural dyads coded for subcategories of socioemotional communication as they occurred naturally in this context provided sparse results. When present, such communication behaviors were largely limited to challenging politely, which was however less commo n than challenge s presented without any positive couching, with almost no observable instances of other subcategories of socioemotional communication behaviors described in the next chapter. Additionally, any such behaviors that were observed were frequent ly occurring within the same dyads, suggesting a mechanism of reciprocity, and were almost entirely absent from others. Overall, many more missed opportunities than used opportunities to use such language were observed. Theoretical bas es for the link betw een p erceptions of cooperation and productive a rgumentation Two theories, namely social i nterdependence theory and s ocio - cognitive conflict theory, point to specific pathways by which learners benefit from perceptions of cooperation and interdependence whi le engaged in online cooperative learning tasks such as constructive controversy. Both are in line with the premise that argumentation as a technique for promoting an understanding of divergent possible views on controversial issues and fostering integrati on skills can only be productive if learners are guided to approach the activity in a non - adversarial, co - constructive manner that allows learners to explore contrasting viewpoints while building their arguments, rather than focusing solely on advancing th eir initial (assigned or natural) position (e.g. Nussbaum, 2008; Nussbaum & Schraw, 2007) . 46 Social interdependence t heory Social interdependence theory (Deutsch, 1949 , 1962 ; Johnson & Johnson, 1994 ; 2005 ) differentiates between three main ways learners ma y view the relationship between the task at hand and their group members, and suggests that these perceptions then determine how learners will interact, creating different outcomes. Cooperative goal structures describe the view that the goal at hand is bes t achieved through cooperation with the partner, i.e., that there is positive interdependence between the involved actors in that a student goal attainment is contingent on SIT predicts that cooperative goal structur es cause individuals to seek outcomes that simultaneously benefit both partners. Constructive controversy as a cooperative learning activity is geared at engendering such perceptions by providing complementary starting points to guide the research efforts of students tasked with eventually synthesizing the best arguments from both sides. By contrast, competitive goal structures refer to the view that the goal is thwarted by the involvement of the other, i.e., that there exists negative interdependence in failure to obtain his or her goals, as in competitive debate where one side wins at the expense of the other. If goals are perceived to be structured competitively, individuals will seek outcomes that benefit themselves to the detriment of the partner. In a cooperative activity that is uld be indicative of perceived negative goal interdependence, as would the view that one already has the most relevant - consuming and mi sleading distraction. Finally, in dividualistic goal structures refer to the perception that achieving the goal is entirely indepe , in other 47 words, the perceive d absence of interdependence . Under such goal structures, individuals tend to seek outcomes that b Particularly for asynchronous online cooperative learning tasks, the risk of individualistic goal structures is a concern, as discussed previously in connection with ence with logistical and coordinative challenges and lost momentum in online discussions (also see Hewitt, 2005; Walther & Bunz, 2005). Social interdependence theory postulates that the nature of the interaction will differ depending on which of these th interactants, which, in turn, create s differences in achievement, the quality of the relationship, and motivational variables (Figure 2 ). Figure 2 : Soc ial interdependence t heory Since goal structures operate as individual - level variables and may differ between two members of a dyad, it is plausible that a participant with a more cooperative goal structure can pull a more competitively oriented partner up to more productive behaviors by modeling what Deutsch (1949, 1962) calls promotive interaction behaviors, as well as vice versa. Promotive interaction behaviors refer to behaviors that promote the success of others and include mutual help, sharing of resources and information, and acting in trustworthy and trusting ways, which in turn enhances the odds of goal achievement. In contrast to this, competitive goal structures result in what Deutsch labels oppositional achievement efforts by e.g. hiding resources and information from each other and acting in Task structure Perceived goal structures - cooperative - competitive - individualistic Interaction pattern - prom otive benign spiral - oppositional destructive spiral Achievement Relatio nship Motivation 48 distrustful and distrusting ways. These patterns are self - reinforcing in nature: Deutsch (1985) describes both a benign spiral by which (intermediate) goal achievement can feed back i nto more promotive behaviors, and a destructive spiral , by which an initial frustration at not reaching a goal fosters more oppositional interaction, decreased relational quality, and decreased motivation. For interactants subscribing to individualistic go al structures, the theory predicts only limited interaction with ffort when achieving the goal. S individual assignments over group work reflects this underlying goal orientation across tasks. S everal researchers have observed these mechanisms in online cooperative learning contexts, with learners holding cooperative views being more open about their views and more butions, less wording implying authority (Alexopoulou & Driver, 1996; Keefer et al., 2000). Revisiting versy task carried out asynchronously via CMC led to decreased cooperative perceptions and increased individualistic perceptions, along with the predicted declines in motivation (based on the belongingness, interest, and task value subscales of the i ntrins ic motivation inventory by Ryan, 1982) and reduced achievement (based on completion rates) in light of this theory adds support to the proposed linkages. Akin to belongingness manipulation attempting to stabilize cooperative perc eptions in the asynchronous CMC context, an even more direct intervention that explicitly invites such perceptions by inserting relevant cooperative is used here, described below and operationalized in the next chapter . 49 Fostering promotive interactions online via cooperative partner messages Johnson, Johnson, and Holubec (1984) have formulated several levels of cooperative skills needed for promotive interaction in small groups which are particularly relevant for t he context of online cooperation with a partner: F unctioning skills are used in managing group activities and maintaining an effective working relationship, e.g. by asking for facts and reasoning and expressing support and interest; formulating skills buil d deeper - level understanding of the material being studied, e.g. by summarizing what has been said; and so - called fermenting skills stimulate reconceputalizations of the material by solving cognitive conflict, extending answers, and integrating ideas into a single posi tion . Both Walther and Bunz (2005) virtual group rules described above and socio - cognitive interaction moves as summarized by Kim (2009) cooperation skills: for example, acknowledging t relates to the functioning skill of encouraging participation; being formulating skill of inviting elaboration and accuracy checking, and presenting challenges poli tely directly relates to the fermenting skills of criticizing ideas without criticizing people, probing, generating alternative perceptions , and ultimately integrating different ideas. A set o f - promoting the purpose of this study is distilled from the intersection of these converging strands of research (also see Fahy, 2001, 200 2, 2003; Rourke et al., 1999 ) for use as the experimental stimulus as described in more depth in the next chapter. Morton Deutsch Before presenting the hypotheses derived for the current study from social interdependence theory, an extension of the theory needs to be mentioned. After all, further 50 clarification is needed regarding the origins of differ ent types of goal structures. The nature of the task of course plays a large role, with competitive debate creating negative interdependence and constructive controversy, if carried out as intended, creating positive interdependence. However, to explain va processes and effects elicited by a given type of social interdependence also tend to elicit that type of social interdependence. Thus, cooperation tends to induce, and be induced by, mutual help and assistance, exchange of needed resources, influence, and trust. Competition tends to ccess and striving to win conflicts. Individualistic efforts tend to induce, and be induced by, a n a voidance of other people. Each process tends to be self - confirming. Any part of the social interdependence process elicits the other parts of the proce Sheldon and Houser - - directionality of construct influence as proposed by the benign spiral in a two - cycle prospective design supports this line of reasoning. A metaanalysis on the relationships between cooperative vers us competitive or individualistic goal structures and the quality of peer relationships among middle schoolers (Roseth, Johnson, & Johnson, 2008) provides indirect support for the asserted linkages via substantial weighted mean e ffect s izes (ranging from . 42 to .55) when comparing associations of cooperative - and of less adaptive goal structures with a chievement and positive p eer r elationships . Based on SIT and the crude law, the next set of hypotheses is stated below (see Table 4 in the next chapter for in dividual hypotheses). Hyp4 a - c : The use of cooperative partner messages leads to increased cooperative perceptions and decreased competitive and individualistic perceptions . Hyp5: The use of cooperative partner messages leads to increased interpersonal 51 attr action . Hyp6 a - e : The use of cooperative partner messages leads to increased motivation . Hyp7 a - c : The use of cooperative partner messages leads to increased achievement . Socio - cognitive c onflict t heory Given the deliberately introduced cognitive conflict that forms the basis for cooperative learning activities such as constructive controversy, a second theory proposing an alternative pathway from cooperative perceptions to achievement is relevant in this context. Like Social interdependence theory, s ocio - c ognitive conflict theory (Doise & Mugny, 1984) suggests that cooperative perceptions influence achievement and motivation via the a daptivity of responses, but it includes distinct social - cognitive responses antecedent to the behavioral responses (Figure 3 ) . It posits that in such learning tasks, dissent among partners may stimulate task - related cognitive activity and result in constructive interactions and positive outcomes to the extent that such conflict is approached with a foc us on building understandin g a process labeled epistemic conflict regulation. On the other hand, such dissent approached with a focus on dominating the other in the discussion, or relational conflict regulation, will thwart beneficial outcomes on achievement, motivation, and the q uality of the relationship due to a counterproductive concern with social comparison s ( Butera & Darnon, 2010; Johnson & Johnson, 2009 ) as well as a misplaced effort in defending arguments of t he other (Keefer, 2002). Rourke and Kanuka (2007) observed that counterarguments to claims made in mediated debates were sometimes perceived as personal attacks, supporting the possibility that the first link of the theory below is particularly relevant in mediated contexts. 52 Figure 3: Socio - cognitive conflict t heory The beneficial effects of epistemic and detrimental effects of relational conflict regulation have been confirmed in laboratory studies in which a disagreeing confederate either refe rred back - to - face and computer - mediated - learning tasks, stronger conflict proved particularly beneficial or particularly detrimental based on whether or not epistemic reg ulation was being initiated. As a result, the nature of the interactions went along either upward or downward spirals, where deeper understanding was sought, points of view were integrated, and more learning gains realized, or where participants tried to a ssert themselves at the expense of careful consideration of all aspects of the issue, reducing learning gains, respectively ( Darnon, Buchs, & Butera , 2002; Darnon, Doll, & Butera, 2007). The present study adds to the literature by testing whether these soc io - cognitive processes are at play in computer - mediated, asynchronous controversy tasks. The following f ive hypothes es parallel those derived from s ocial interdependence theory but explicitly address the tendency to assume an epistemic - versus relational st ance in the context of (see Table 4 for individual sub - hypotheses) : Hyp8 a : Cooperative perceptions are positively associated with epistemic conflict regulation. Hyp8 b : Competitive perceptio ns are positively associated with relational conflict regulation. Under conditions of cognitive conflict: Task Perceptions - Cooperative - Competitive or Individualistic Adaptivity of responses - epistemic conflict regul ation - relational conflict regulation Achievement Relationship Quality Motivation 53 Hyp9: Epistemic conflict regulation is positively associated with perceived relatedness. Hyp10: Epistemic conflict regulation is positively associated with motivation. Hyp11: Epistemic conf lict regulation is positively associated with achievement. The chosen two - by - two factorial design described in more detail in the next chapter allows both for the investigation of main effects of the two independent factors the distal, course - wide norms - treatment and the more proximal, dyad - based partner messages treatment, as well as allowing an exploratory look at interactions between the two via corresponding research questions. After all, it is plausible that these two distinct avenues of social influ ence have more - or less than a linear - additive effect . For example, receiving encouragement to remain engaged in the task from multiple levels might enhance the effects of each level . Conversely, destructive interaction effects are possible , as when the re peated exposure to group level norms creates the undesirable side effect of feeling pressured in the absence of an engaging partner , or guilt about falling behind, which might actually exacerbated by the presence of a particularly engaging work partner. Ei ther scenario might then affect goal construction processes and motivation in undesirable ways. V arious such interactions, both constructive and destructive with regards to the outcome variables used and both enhancing/exacerbating and buffering/underminin g in nature , a re of potential interest (Fig. 4 a and Fig. 4 b ). Given the numerous possible interactions between the factors at both levels in this particular context, a set of open - ended research questions is posed: Figure 4 a : P otential distal - proximal infl uence interactions : CPMs as moderator Cooperation - promoting partner messages (CPMs) Descriptive norms messages Perceived descriptive norms Engagement 54 RQ 1 : Does the use of cooperation - promoting partner messages moderate the effect of descriptive norms on perceived descriptive norms or engagement? Conversely (and very closely related to the previous researc h question, albeit not completely interchangeable with it given the different combination of outcome variables of interest): Figure 4 b : P otential distal - proximal influence interactions : Norms messages as moderator RQ 2 : Does t he use of cooperation - promoting partner messages moderate the effect of cooperation - promoting partner messages on perceived goal structures, conflict regulation, motivation, engagement, or achievement? Cooperation - promoting partner messages (CPMs) Descriptive norms messages Goal structures, Conflict r egulation Motivation Engagement /Achievement 55 CHAPTER III: METHOD Overview of the R esearch D esign This study used a between - groups experimental approach as well as observational elements. Participants were assigned to four treatment groups based two - by - two factorial design. The two variables of interest were (1) descriptive norms present/absent and (2) cooperative partner messages present/absent. The four resulting conditions were presented to roughly fifty participants each, all of which were asked to complete a final survey assessing the constructs of theoretical interest referred to above. Both Analy sis of Variance and correlation/regression techniques were be used to examine the effects of the two manipulations and to test the relevant theories with this data set, respectively, as described in more detail in the analysis section below. Figure 5 : Fac torial design showing experimental and control groups SN on SN off CPM on SN+CPM n=58 50 analyzed - only 2) n =55 44 analyzed CPM off - only n =59 45 analyzed 4) n = 54 48 analyzed Notes: SN: Descriptive social norms manipulation (behavioral prevalence information), CPM: Cooperation - promoting partner messages Study C ontext and Participants Invitations to participate were disseminated to students eighteen years or older who were c urrently enrolled in one of the six participating undergraduate communication courses at a large Midwestern university. They were informed about the requirements to log on to a specific discussion website once per day for five consecutive days , requiring t he ability to go online at least once per day at a time that would be relatively flexible. The slight majority of students in 56 these classes are communication majors, but many other majors are represented as well given that some courses are required for bro ad ranges of majors and that some involved courses overlapped with curricula in business and other fields. Accordingly, 56.5% of participants in the final samples w ere communication majors , followed by a group of business and finance related majors (20.7%) . The remaining 22.8% indi cated one of almost forty different majors ranging from accounting to wildlife management . 95.9% (185) of participants ranged in age from 18 - 22, 4.1% (8) were 23 or older. A little over half (54.4%, 105) of participants were femal e. The majority of participants (76.2%, 147) reported their ethnicity as White. Participants were fairly evenly distributed across class levels (see Table 1 ). Table 1 : Sample characteristics Age 18 - 22 23 - 30 Over 30 95.9 % (185) 3.6 % (7) 1 (0.5 %) Gend er Female Male Did not report 105 (54.4 %) 86 (44.6 %) 1 (0.5 %) Ethnicity White Black or African American Asian or Pacific Islander Hispanic or Latino/a Native American Other 147 (76.2 %) 19 (9.8 %) 15 (7.8 %) 5 (2.6 %) 1 (0.5 %) 6 (3.1 %) Year 1 st year 2 nd year 3 rd year 4 th year 5 th year 36 (18.7 %) 28 (14.5 %) 65 (33.7 %) 57 (29.5 %) 7 (3.6 %) GPA 3.6 - 4.0 3.1 3.5 2.6 3.0 2.1 2.5 < 2.0 50 (25.9 %) 78 (40.4 %) 46 (23.8 %) 15 (7.8 %) 4 (2.1 %) Major Communication Business Finance Other* 109 (56.5 %) 26 (13.5 %) 14 (7.2 %) 44 (22.8 %) * Advertising, Anthropology, Engineering, Computer Science, Construction Management, Education, English, Human Biology, Human Development, Fisheries and Wildlife, Political Science 57 Procedures Students receive d an invitation to partic ipate in the study via an email and in - class announcement and were asked to submit their name and course information by email if interested in signing up. Further information about the details of the study as well as instru ctions on how to generate a needed user account on the interactive controversy website were provided in response. Students following these instructions were randomly assigned to one of the four experimental conditions by the system, ensuring that early - sta rting and last - minute tendencies as well as other interindividual differences would be allocated equally to conditions. Upon entering the site, participants were prompted to review a consent form that could be agreed to by clicking a link that would take t hem to the main site (see A ppendix A.3. for a screenshot). The main site contained a link to downloadable stimulus materials and the asynchronous discussion platform that scaffolds the 5 - step constructive controversy procedure. The topic of the controversy and the stimulus materials were chosen to be general enough for a range of different majors and individuals to connect to yet to have connections to organizational communication, workplace communication, the future of human resource management and thus bl end into most communication related courses (see Appendices A.3 and A.4 for sample materials and instructions). Over the course of the week after the invitation, participants engaged in the five interactive steps of the cooperative learning activity by fol lowing guiding prompts in the plat form. All participants were automatically notified via their provided email address whenever a partner response had become available, eliminating the need to log on to the site multiple times per day to check for a respons e. For participants in the two conditions that contain ed descriptive group norms, this information was prominently displayed on the main activity site, right above le norms message would say that , 58 e.g., these messages after noon of the day that the step was due, in order to ensure credibility of the message, i.e., allowing a sufficient time window for a majority of participants to have completed the activity step for that day . Par ticipants in the cooperation - promoting message condition receive d confederate - generated answers to each of their postings that contain per post a combination of two of those interactive/su pportive and cohesive partner behaviors (Fahy, 2001, 2002, 2003; Jeong, 2005; Kim, 2009; Rourke et al., 1999) most applicable to the cognitive scaffolding provided by the nature of constructive controversy and most aligned with Johnson, Johnson and Holubec previous chapter EA); challenging/ disagreeing politely (CP); using greetings (GT); and using group - centric language (GR). A particu lar cohesive behavior category wa s added to the li st provided by Kim (2009) based on Johnson et differentia ting ideas and reasoning of group members and integrating ideas into a single position . Messages of this added type made explicit references to the benefits of pooling information to broaden the range of available ideas and highlighting the goal of idea in tegration . Table 2 contains example scripts for each type of message . Participants in the SEPB - off condition receive d confederate - generated answers to each of their posts that contain ed none of the listed behaviors but simply provide d initial - and subsequ ent counterarguments as required by the CC instructions. In order to balance the need for keeping the messages both natural and parallel across the conditions, t he order for both substantive - and cooperative confederate input followed a systematic, randomi zed schedule (see Figure 10 in 59 Appendix B for a snapshot of the message generation schedule). Stock responses (in both conditions) were based on a pilot study and minimally modified for coherence. F would be referred to briefly. The systematic adherence to the message generation schedule led eleven of the 215 final activity participants to express suspicion regarding the partne r, as indicated by their answers to the survey exit questions, and their survey responses were excluded from the analysis. Table 2 : Prototypes of cooperative partner messages Interactive/supportive Behaviors AC Acknowledging/Complimenting message conte nt I really appreciated EA Expressing Agreement I can see where you are coming from with... CP Challenging/Disagreeing Politely MP Making Polite requests ( use for day after a step due date) What else do we need to know, from the YES side? Cohesive Behaviors GT Greeting Hi! GR Group - centric/inclusive language see what else we need to think about CC Cooperation - promoting language It looks like both of us have good points. My next point would be It will be interesting to put these arguments together later. For Source : adapted from Kim ( 2009 ) After completing the five steps of the co ntroversy acti vity, the student tasked with the initial draft of the integrative statement (the real participant) h ad to sign off on the equally systematically edited integration statement , and was prompted to click on the link to the final 60 survey containing measures of many of the constructs included in the hypotheses (see below). A link to d ebriefing information and raffle sign up were presented at the end of the survey. Participants completed the activity in groups of between 63 and 83 during the three weeks in which the study was offered in different courses. Measured V ariables Most variables of interest were assessed with well - established scales in the final questionnaire (see appendix 5 for items) , along with a few variables based on behavioral coding . Scales used are described first, followed by a description of c oded behavioral measu res. Unless otherwise specified , respondents were asked to rate their agreement to each item on Likert scales gher scores on all scales indicate higher self - reports of each construct. Reliability information listed below is based on the present study unless specified otherwise . For means and sta ndard deviations by group, see t able 5 . Motivation and attraction In line with previous research in this area, two subscales Motivation Inventory (IMI) were used , specifically, the interest/enjoyment subscale with items such as (4 items , perceived competence subscale ( 4 The activity in laboratory experiments (e.g., Ryan, 1982; Ryan, Mim s & Koestner, 1983; Ryan, Connell, & Plant, 1990). The subscales include interest/enjoyment and perceived competence as used here , along with effort, value/usefulness, felt pressure and tension, and perceived choice while performing a given activity. The i nterest/enjoyment subscale is considered the actual self - report measure of intrinsic motivation while th e competence subscale, like e.g. the perceived 61 choice subscale, is conceptualized as a positive predictor of intrinsic motivation (Ryan, 1982). Factor a nalyses during scale development showed coherent and stable factor solutions across a broad range of tasks and settings, and shortened versions have been used and found reliable, with four items per construct generally providing good reliabilities (e.g. Mc Auley, Duncan, & Tammen, 1989). (1985) classroom life scale. The classroom Life Scale (CLS) was developed on la rge samples of students from grades 5 through 9, and factor analyses confirmed the theorized factors, including perceived academic and personal support provided by teachers and other students. Internal e original 5 - item version ( Johnson & Johnson, 1983), and it correlated with liking of cooperative learning at r = .40, and only slightly so with liking of competitive learning and individualistic learning as described by the Social Interdependence Scale ab ove ( r = .15 and - .18, respectively). Follow - up studies using the measure confirmed the relation between this construct , preference for cooperative learning , and experience with cooperative learning (Johnson, Johnson, Buckman, & Richards, 1985). To assess perceived task value for the controversy activity, two subscales based on work by Eccles, Wigfield, and colleagues (e.g. Eccles & Wigfield, 1995) were adapted. Out of the four subscales of subjective task value Interest/Attainment/Utility/Cost value t hat were supported by confirmatory factor analysis by Conley (2012), the latter two were used here. Five items such as utility value e. the extent to which engaging in a task or learning abo ut a particular subject is perceived to be useful beyond the school context. In previous work, perceived utility value was positively correlated with a 62 beneficial, mastery - approach to learning, similar to the previously described epistemic conflict regulat ion construct ( r = .56) and competence beliefs ( r = .34) (Conley, 2012). Cost value , a construct that addresses the extent to which the target task is perceived as requiring sacrifices with regards to other desirable activities and therefore negatively con tributes to overall task value w as measured with .81). I n previous work, cost was uncorrela ted or very lowly correlated with competence b eliefs and mastery goals ( r = . 02 and .17, respectively, Conley, 2012). T hree items assessed work avoidance . F rk as possible in this activity ( Harackiewicz, D urik, Barron, Linnenbrink - Garcia, & Tauer, 2008). Exploratory factor analysis by Harackiewicz et al. (2008) using principal component extraction provided support for the measure s integrity. Support for its construct validity comes from negat ive path coeff icients found with self - - .19 and - .29, respectively) and positive coefficients college students in psychology course. Perceptions of social interdependence Goal stru ctures were assessed with the Social Interdependence Scale ( SIS, Johnson & Norem - Hebeisen, 197 9 ), with its three subscales assessing perceived cooperation ( 5 items, . 86, sample item: , competition ( 4 items, , 87, sample item: In this activity, I wanted to be the best in this activity and individualism ( 3 items, 89, sample item: we could have completed this activity on our own The social interdepend ence scale was originally developed as part of a larger instrument for evaluating affective outcomes of schools (the Minnesota School Affect 63 Assessment, MSAA), and tested on several large samples of students from grades 1 through college (Johnson & Norem - H ebeisen, 1997). Principal component factor analyses with varimax rotations followed by confirmatory factor analyses on later samples led to the three measures of cooperative, competitive, and individualistic interdependence, each of which contained items a ddressing liking to cooperate/compete/work individually as well as valuing cooperative/competitive/individualistic learning. Early iterations of the three measures with only 3 - 4 items produced Cronbach alpha reliabilities from .62 - .71 with later iteratio ns with 7 - 8 items producing alpha values from .84 - .88 as found by the scale authors and of .91 - .94 and .86 - .83 in college undergraduates (Roseth et al., 2011; Saltarelli & Roseth, 2014). A negative correlation tended to be found for cooperative and i ndividualistic , and a small positive correlation for competitive and individualistic scales. Supporting the scales construct validity, Norem - Hebeisen and Johnson (1981) found a theoretically predicted positive relationship between valuing cooperation and general measures of self - esteem, a negative relationship between individualistic preferences and self - esteem, and no correlation between valuing competition and self - esteem across four studies with seventh - to twelfth graders. Socio - cognitive conflict re gulation egulation s cale was used to assess the extent to which participants engaged in the controversy with a focus on issues and solutions ( epistemic conflict regulation), or with a focus on defending a certain position ( relational conflict regulation). The subscales assess When disagreements occurred, to what exte nt did (3 items, . 83) as well as the latter tendency, e.g., (3 items, 74). The se two scales were developed via factor 64 analyses using promax rotation on the six conflict regulation i tems (see appendix A.5.), which sociocognitive conflict regulation (e.g. Doise & Mugny, 1984; Mugny, Butera, Quiamzade, Dragulescu, & Tomei, 2003; Mugny, De Paolis, & Car ugati, 1984), with the first factor comprising the three relational regulation items and the second comprising the three epistemic regulation items. Internal reliabilities observed by the scale authors were .82 and .91 for epistemic and relational regulati on, respectively, and construct validity was supported via the theoretically expected prediction of epistemic regulation by mastery goals (i.e., the desire to acquire knowledge and develop competencies, r = .51) and of relational regulation by performance goals (i.e., the desire to demonstrate competenc i es relative to others, r = .51; Darnon, Muller, Schrager, Pannuzzo, & Butera, 2006 ). Despite their conceptual distinctiveness, epistemic and relational regulation were found to correlate positively ( r = .31) by these authors, and others found low positive correlations of .21 as well (Saltarelli & Roseth, 2014). Perceived descriptive norms The most common approach for measuring perceived descriptive norms within TNSB research has relied on a summative index o f behavioral prevalence estimates across a series of relevant situations, such as asking students to estimate how many drinks a typical student at their university would be drinking (a) when going to a bar, (b) when having friends over for drinks, (c) when going to a party and (d) during the weekend from Friday to Sunday evening, generating of target behaviors, behavioral prevalence assessment needs to be ada pted to align with the study on descriptive norms on water conservation behaviors (Lapinski, Rimal, DeVries, & Lee, 65 2007). These authors generated 5 items to as sess perceptions of the extent to which water - point Likert scale with higher numbers indicating higher perceived pre valence, and used confirmatory factor analysis (CFA) including tests of internal consistency and parallelism. The resulting 4 - item unidimensional solution served as the basis for the items used here, assessing the extent to which respondents perceived the target behavior to be carried out by the reference group. For the present study, f adapted from La pinski, Rimal, DeVries, and Lee (2007) to assess the extent to which students perceived the majority of other pairs to engage in the t arget behavior as instructed , including staying on track with the activity Potential moderators of norms - effects Two of the moderators proposed by TNSB were in cluded in the final survey, namely e go involvement and group identification . Ego involvement, i.e., the extent to which behavior under considerati on is linked to self - concept, was approach . The measure of behavioral identity as originally developed for the college drinking context friends, your parents, and members of your extended family) consider yourself a nondrinker (cod - identity related information lowered the reliability of the overall measure, h ence dropping it for the final index of behavioral identity which consisted of averaged responses to the two remaining questions. Internal consistency for this index was r = .85 and .92 (Rimal, 2008; Rimal & al., 66 items illustrates the need for the assessment source of behavioral identity information to be able the set of questions asked whet her they, their friends, or their teachers respectively would describe them as an , 93) and, in a second set of questions, an communicator ( 92) , based on the relev ance of these two behavioral identities to the target task, as described in the previous chapter . The measure for group identification used here was developed by Rimal and colleagues (Real & Rimal, 2007; Rimal, 2008; Rimal & Real, 2005), who followed Tajfe (1986) two - pronged to be like the referent others. As discussed previously, the present study focused on the similarity component, for which Rimal and colleagues te sted four items scored on a 7 - point scale with university intellectually/in their thinking/in their values/in their behavio 10) in a stud y applying the TNSB to incoming by perceived descriptive drinking norms of and identification with the student body of the large university. S upport for the reliability and validity of the similarity measu re was established via principal component factor analysis with varimax rotation, which indicated the expected factor structure of two identity factors along with injunctive norms related factors tested alongside it. 8 and .80 was found by these authors for similarity (Rimal, 2008; Rimal et al., 2004), along with a positive correlation of .49 with the second group identity factor, aspiration. In the context of college drinking behavior, similarity correlated with the s elf - reported target behavior at r = .25. In this study, g roup identification ( 89) was 67 assessed by following Rimal et al. (2004), order to get as specific as possible without excluding the actual referents, i.e., other participants in the activity who had to be recruited from additional classes. As stated previously , o utcome expectations a n d injunctive norm s, the two additional moderating factors outlined by the theory, were not assessed in this current study in order to reduce the burden on the participants and eliminate the need for a pre - activity survey on top of the used post - activity survey. Additional measures T he questionnaire concluded with two brief open - ended exit question s regarding what students (a) most liked and/or disliked about the activity and (b) what their best guess regarding the purpose of the study was . In all, 93% ( 203 ) of the 219 rece ived surveys contained a statement regarding the presumed purpose and 58% (126) of submitted surveys contain ed a statement regarding likes or dislikes. An important goal for these questions was to identify participants 11 participants ) or who guessed correctly that their motivational responses to different kinds of message styles were of interest ( 1 participant ). Their data was removed from the analyses . Behavioral E ngagement an d Achievement Behavioral engagement was quantified both in terms of number of completed steps out of the five prescribed steps of the acti vity (completion rate) as well as the number of steps out of these five that were completed within the prescribed post ing time frame for each step ( timeliness ). The length of the participant - generated draft of the final integrative statement was used as an indicator of engagement as well, given that adherence to minimal posting requirement 68 is a common observance in online courses and that writing longer text in a low - stakes task such T he quality of the participant - generated draft for the joint final essay was assessed using - point assessment of k nowledge integration (see Table 3 ). Inter - rater reliability was determined by comparing independent coding for a randomly selected subsam p le of 24 integrative statement drafts by the first author and a research assistant w ho was blind to the conditions and the hypotheses of the study. There was good agreement between th e raters' judgments , with an unweightened Cohen's . Table 3 : Levels of integration of arguments 4 Exceptional Joint statement is balanced, with integrative closing. It weighs evidence on two sides. 3 Well integrated Joint statement develops substantial counterarguments and rebuttals or a substant ial it - depends argument 2 Slightly integrated Joint statement has (a) a minor it - depends argument, (b) a minor reservation or (c) different conclusions at beginning and end 1 Unintegrated Joint statement discusses only one side of the issue or has no fin al conclusion Source : Nussbaum & Schraw ( 200 7) Finally, a short four - item content knowledge quiz consisting of multiple choice and true - or - false questions regarding the controversy topic yielded a knowledge score from 0 - 4 for each participant. Overview o f Hypotheses , Data Sources, and Analyses To investigate the research questions introduced above regarding the effects of social norms and socioemotional partner behaviors on asynchronous, online constructive controversy performance and motivation, and the plausibility of such effects being moderated via the proposed mechanisms, data were collected from three sources: Participants postings on the constructive controversy website, participants first drafts of the integrative essay, and 69 parti ci pants respons es to the final survey. Data sources and analyses for each hypothes e s cluster are summarized in Table 4 . As the table s hows, the proposed data collection can be analyzed on two levels: First, as a 2x2 factorial experimental design with random assignment to conditions allowing for ANOVA procedures to compare causal effects of the two manipulations and their possible interactions. Second, the linkages proposed by three selected theories can be te sted descriptively via correlation techniques. Table 4 : Research hypotheses, data Sources, a nalyses Hypothese s Data Sources, Instruments, Analysis: (see appendix A.5 for specifics on measures used) From the Theory of Normative Social Behavior: Effects of beh avioral prevalence info rmation on descriptive norms: Hyp 1 a : The presentation of descriptive norms messages regarding the completion of controversy steps within the participating cohort leads to higher descriptive norms as reported by learners. Effects of norms presentation on engagement: Hyp1 b - c : The presentat ion of descriptive norms messages leads to higher completion rates and higher timeliness of posting. Data Sources: Self - reported perceived descriptive norms (final survey); behavioral measure for completion/timeliness of responses Instrument: Descripti ve norms scale; observation of completion rate (# of steps completed/5 total steps); timeliness (# of steps completed within 24 hours of partner posting/5 total steps) Analysis: ANOVA (main effect on group means for perceived norms, completion, timelin ess) Associations between perceived norms and engagement: Hyp2 a : Higher perceived descriptive norms are associated with higher completion rates. Hyp2 b : Higher perceived descriptive norms are associated with more timely responses. Data Sources: Self - reported perceived descriptive norms; behavioral measure for completion/timeliness of responses Instrument: Descriptive norms scale; observation of completion rate and timeliness (see Hyp. 1) Analysis: Correlation analysis of completion rate/timeliness rate and perceived descriptive norms 70 Table Moderation effects: Hyp3 a - b : The effect of perceived descriptive norms on completion rate and/or timeliness of responses becomes stronger with increasing group identification and ego - involvement Data Sources: Self - report (final survey) , behavioral measure for completion/timeliness Instrument: Descriptive norms scale, ego - involvement scale; group identification scale Analysis: multiple regression with continuous predictors and multiplicative i nteraction terms From Social Interdependence Theory: Effects of Cooperative partner messages on perceived goal structures: Hyp4 a - c : Cooperative partner messages are associated with increased perceptions of cooperation, decreased perceptions of competi tion, and decreased perceptions of individualism. Data Source: self - reported social interdependence (final survey) Instrument: SIS subscales (cooperation, competition, and individualism) Analysis: ANOVA (main effects on group means for CPMs , with coo peration, competition, and ind ividualism as DVs) Effects of CPMs on relational quality: Hyp 5: Cooperative partner messages are positively associated with relatedness/attraction. Effects of CPMs on motivation: Hyp6 a - e : CPMs are positively associated w ith interest/enjoyment; perceived competence; perceived utility. CPMs are negatively associated with perceived cost and work avoidance. Effects of CPMs on engagement/achievement: Hyp 7 a - c : CPMs are positively associated with length of final statement draf t, integration level of final statement draft, and content knowledge. Data Source s : self - reported relatedness to par tner and task - related motivation, content knowledge questions (final survey); student draft of integrative final statement Instrument s : interpersonal attraction scale , short subscales for interest/ enjoyment ; perceived competence, perceived utility, perceived cost, and work avoidance; numeric count of characters written for final statement draft; coding scheme fo r final statement draft, s hort content quiz scores Analysis: ANOVA (main effects on group means for SEPB, with relatedness, interest/enjoyment, integration score as DVs) 71 Table From Socio - cognitive Conflict Theory: Associations of conflict regulation approach with perceived goal structures : Hyp8 a : Cooperative perceptions are positively associated with epistemic conflict regulation. Hyp8 b : Competitive perceptions are positively associated with relational conflict regulation. Data Sources: self - reported epist emic and relational conflict regulation and self - reported social interdependence (final survey) Instruments: Conflict regulation subscales, SIS subscales (cooperation, competition, and individualism) Analysis: Correlation analysis of conflict regulation subscale score and social interdependence subscale scores Associations of conflict regulation and relationship quality: Hyp9: Epistemic conflict regulation is positively associated with perceived relatedness. Associations of conflict regulation and mot ivation : Hyp10 a - e : Epistemic conflict regulation is positively associated with interest/enjoyment, perceived competence, perceived utility, and negatively associated with perceived cost and work avoidance. Associations of conflict re gulation and engageme nt/ achievement : Hyp11 a - c : Epistemic conflict regulation is positively associated with length of final statement draft, integration level of final statement, and content knowledge. Data Source s : self - reported attraction/relatedness and task - rela ted motivation, content knowledge questions (final survey); student draft of integrative final statement Instrument s : interpersonal attraction scale , short subscales for interest/ enjoyment ; perceived competence, perceived utility, perceived cost, and work avoidance; numeric count of characters written for final statement draft; coding scheme fo r final statement draft, short content quiz scores Analysis: Correlation analysis of both epistemic conflict regulation and relational conflict regulation and DVs l isted. 72 Table Interactions between the distal and proximal factors: RQ 1 : Does the use of cooperation - promoting partner messages moderate the effect of descriptive norms on perceived descriptive norms or engagement? RQ 2 : Does the use of coop eration - promoting partner messages moderate the effect of cooperation - promoting partner messages on perceived goal structures, conflict regula tion, motivation, or engagement/achievement ? Data Sources: see above Instruments: see above Analysis: Two - way ANOVA or nonparamteric equivalent (test for interaction effects of CPM and SN factors , with perceived descr. norms, timeliness, completion rate as DVs (RQ1) and cooperation, competition, individualism , epistemic regulation, relational regulation, Interest/ Enjoyment, Competence, Attraction, Utility, Cost, Work Avoidance, level of integration and content knowledge as DVs (RQ 2 ) ) 73 CHAPTER IV : RESULTS Participant Flow A total of 240 students were assigned to conditions across the three one - week periods in wh ich the experiment was offered for part icipation (58 - 61 per condition), corresponding to a n overall participation rate of .25. Thirteen of these potential participants did not receive the experimental manipulations by either creating an account but never s tarting the activity or by completing all five steps upon their initial log on, without waiting for partner responses. Two - hundred and twenty seven participants participated in the actual activity (54 - 59 per condition on day 1; 51 - 54 per condition by day 5 ). All students who had at least completed two of the five daily steps of the activity according to instructions were reminded via email to take the final survey. Two - hundred and nineteen surveys were received in total, 26 of which needed to be removed fro m the analysis, leading to a final set of analyzed cases of N=193 (44 - 51 per condition). Reasons for exclusion from the analysis included suspicions regarding the pen - ended exit questions of the final survey; disregard of the activity instructions that caused participants to miss the treatment; repeat surveys or surveys taken in less time than 95% of participants required to complete the survey. Table 9 in the appen dix contains additional details of the participant flow by condition and cohort. Reductions in sample size were distributed evenly across the four experimental conditions. Table 5 reports descriptive statistics for each dependent variable. To address betwe en - group variability due to history effects (such as the temporal proximity of final exams during the third and final iteration of the study), analyses were 74 c onducted on group - mean centered data, descriptive statistics for which can also be found in Table 5 . Table 5: Descriptive s tatistics __________________________________________________________________________________ _____ _ _____ Notes: DesNorms = Perceived descriptive norms; EgoCom/EgoStudent = Ego Involvement as self - perception as good communicator/go od student; GroupID = identification with classmates (similarity); Steps Completed and Timeliness = number of daily steps, out of five, that were completed or completed within the prescribed time window; FinalLength = length of final statement draft in cha racters; IntegratLevel = Integration level of final statement draft on 4 - point scale; Content Knowledge = score on a 0 - 4 point content quiz. All other variables assessed via 7 - point Likert scales with higher numbers indicating higher levels of the construc t. Manipulation Check To ensure that participants in the two groups that contained a descriptive norms treatment did in fact notice the normative information within the website, approximately one third of participants in all four conditions were asked to indicate whether the activity website had provided them with information ummy coded as 1 (participant recalls seeing normative information), while 75 information). 85.7% of the subsample of the two norms - containing conditions reported noticin g normative information, while 84.8% of the two norms - free condition reported not having noticed 2 (1, N = 68) = 33.86, p < .001 ) , providing evidence that this information was sufficiently visible. Factorial Analyses of Experimenta l Conditions Experimental effects were analyzed using a series of 2 ( completion percentages : provide d, not provided) x 2 (cooperation - promoting partner messages: provided, not provided) analyses of variance (ANOVAs) along with post hoc tests using Bonferro ni multiple applicable pairwise comparisons. While the necessary assumption of independence of observations was met by the design in which all participants responde d to daily input that was of error variances between groups was employed throughout to ensure this additional assumption for the procedure was met. Outlying c ases were removed for each dependent variable examined unless otherwise stated, as indicated in the degrees of freedom for each test. The robustness of findings was assessed by reanalyzing the d ata set with outliers included. Confirmatory factor analysis w as used to guard against potential confounding effects of overlapping factor loadings across different scales. Where overlap was found, derivation of constructs on non - overlapping items only was used to confirm results . For example, overlap between items f or perceived descriptive norms and cooperation could be eliminated by deriving norms scores from items 3 and 4, and cooperation scores from items 3, 4, and 5 only. The resulting effects ( F(1, 187 ) = 6.944 , p = .0 1 2 = .04 for an interaction effect of CPMs and completion percentages on 76 descriptive norms and F(1, 186 ) = 5.926 , p = .0 2 2 = .03 for a main effect of CPMs on cooperation) mirror the results based on full scales reported below. Perceiv ed descriptive norms For perceived descriptive norms, results showed a significant omnibus ANOVA , F(3, 178) = 4.372, p 2 = .07, and a two - way analysis of variance showed a significant interaction between factors, F(1, 178) = 6. 174, p = .0 1 2 = .03. Post hoc testing revealed t hat descriptive norms as perceived by participants in two of the three experimental groups were higher than in the control group with the third experimental group approaching significance: SN+CPM (p = .076), CPM_only (p = .045), SN_only (p = .006). Results provide support for hypothesis 1 a , which stated that the presentation of course - wide descriptive norms regarding the step completion would lead to higher perceived descriptive norms as reported b y learners. Interestingly, results further suggest that the presence of cooperative partner messages raised perceived course norms regarding regular engagement in the activity. Figure 6: Two - way ANOVA results for descriptive norms Social interd ependence Differences existed for perceived cooperative goal structure ( cooperation ), with an 2 = .07. A nonparametric Kruskal - Wallis test 77 for variance differences between groups and confirmed the ANOVA result, p = .02. A two - way Analysis of variance showed a significant main effect for cooperative partner messages, F(1, 185) = 13.352, p = .001, 2 = .07, supporting hypothesis 4 a . Po st hoc tests revealed significant differences favoring SN+CPM ove r the control (p = .018) and CPM_only over the control (p = .008) . Similarly, the data supported hypothesis 4 c of differences in perceived individualism , i.e. the preference to work on a simi lar task by oneself rather than with a partner: The omnibus one - w 2 = .09. A significant main effect 2 = .09, indicating that students exposed to such messages indicated a lower pr eference for working on a similar task by themselves compared to students not exposed to these messages. No statistically significant differences were found for competitive goal structure , with F(3, 199) = .032, p = .992 for the omnibus and F(1, 188) = .036, p = .85 for the two - way - on and CPM - off . Therefore, hypothesis 4 b was not supported in this study. Motivation and r elatedness Motivation variables examined for grou p differences included interest/enjoyment , perceived c ompetence, utility, cost, work avoidance (Hyp.6 a - e ), and attraction (Hyp.5) . For interest/ enjoyment , the omnibus ANOVA was significant with F(3,184) = 4.237, p = 2 = .07, indicating the presence of group differences. Post hoc testing re vealed that the CPM_only group reported higher interest - value than the control (p = .006), higher interest - value than SN+CPM (p = .038) and approaching significance of higher interest - value compared to the SN _only group (p = .07). 78 While the difference betw een groups exposed to cooperative partner messages over the control supports hypothesis 6 b , the difference between CPM_only and SN+CPM and the significant interaction term of Norms and CPM on 2 = .04) raise questions about the interplay of distal and proximal influences discussed below. Similarly, for perceived competence , a sig nific ant omnibus ANOVA with F(3,184) = 2 = .06 indicated the presence of group differences on this outcome. Post hoc testing revealed that the CPM_only group reported higher perceived competence than the control (p = .021) and higher p erceived competence that the SN+CPM group (p = .046). Here too, the significant interaction term of norms and CPM on 2 = .06) provides partial support for hypothesis 6 c , but also raises questions about the interplay of distal and proximal influences discussed below. Figure 7 : Two - way ANOVA results for i nterest/ e njoyment Figure 8: Two - way ANOVA for c ompetence F or utility , the omnibus one - way ANOVA was nonsignificant, but a two - way ANOVA indicated a s mall main effect for cooperative partner messages that approached statistical 79 significance, F (1, 188) = 2.913, 2 = .02, with students exposed to the messages reporting higher perceived utility than those not exposed. Results regarding work avoidance followed a similar pattern to utility as well, with a main effect for cooperative partner messages 2 = .02. Supporting hypothesis 6 a , attraction toward the partner produced a significant omnibus one - 2 = .17, and a significant main effect for C 2 = .16. Participants in the SN+CPM group perceived more attraction than participants in t he control (p = .001) or the SN_only group (p = .002), just as participants in the CPM_only group perceived more attraction than the control (p = .001) and SN _only group (p = .007). Figure 9 : Two - way ANOVA results for a ttraction 80 Behavioral engagement and achievement Contrary to hypothesis 1 b , completion rates did not differ significantly between groups that were exposed 2 (1, N=193) = .058, p=.810). Hypothesis 1 c , which stated that exposure to group norms would be 2 (1, N =193) = .036, p=.849). A main effect regarding behavioral engagement did emerge for CPM when considering the dependent variable length of final statement , which given the ratio - level measure was examined both via analysis of variance (F(1, N= 189) = 8.440, 2 = .05, and, given the problematic assumption of equality of error variances here too, nonparametically as well with a Kruskal - 2 (1, N=193) = 6.274, p=.012). Results of both tests indicate that students exposed to cooper ative partner messages wrote longer drafts of the final integrative statement, which is in line with hypothesis 7 a . Figure 10 : Two - way ANOVA results for l ength of final statement draft 81 Further, in support of hypotheses 7 b , level of integration of the final statement draft 2 (1, N=193) = 7.173, p=.007. However, the short 4 - item multiple choice/true - or - false measure of content knowledge did not generate difference s between groups, with F(3, 166) = 1.096, p = 2 2 (1, N=193) = 1.389, p=.237, for the nonparametric follow up to take into account the violation of normal distribution and equality of error variances for the four groups, failing to support hypothesis 7 c . Additional exploratory group comparisons Without any formal hypothes e s regarding group differences, self - reported time spent per step, preference for spending more or less time, and attitudes toward using the interactive controv ersy website were compared. There were no differences regarding reported time spent , F(3, 188) = 1.313, p = .27, confirmed by an ordinal (linear - by - 2 (1, 192) = .087, p = .77, since homogeneity of variances had to be rejected. Similarly, experimental groups did not differ regarding their reported preference for spend ing more or less time on the daily 2 (6, 192) = .087, p = .77. Reported enjoyment of using the activity website did not differ between groups either (F(3, 170) = .874, p = .456 for the omnibus ANOVA) . Correlational Analyses Several targeted correlations were examined based on theory - based observational hypotheses (Table 4 ). Table 6 shows the found correlations for all variables of interest. First, to explore the possible descriptive norms (2003) meta - analysis had found to average r = .44 in the context of health - related behaviors (using behavioral intentions instead of behavior, given the common reliance on the latter 82 measure in the summarized literature) in the context of online cooperative learning, descriptive was hypothesized to have a link with behavior. Specifically, hypothesis 2 a described a correlation between perceived descriptive norms and controversy - step completion, and hypothesis 2 b described a correl ation between perceived descriptive norms and timeliness of responses. Both hypotheses were supported albeit with small effect sizes. For step completion, r(191) = .236, p = .001; and for timeliness, r(191) = .237, p = .001, indicating that to some degree, higher perceptions of other participants as moving through the activity according to schedule was related to higher completion rates and timeliness of step completion. Based on sociocognitive conflict theory, hypothesis 8 a proposed a positive correlation between cooperative goal structure and epistemic conflict regulation, which was supported by the data, with r(191) = .505, p < .001. Hypothesis 8 b proposed a positive correlation between competitive goal structure and relational conflict regulation and wa s similarly supported by the data, with r(191) = .513, p< .001. Also based on sociocognitive conflict theory, positive correlations were hypothesized between epistemic conflict regulation and relationship quality as well as motivational constructs. Suppor t was found for interpersonal attraction (r(191) = .318, p < .001), interest/enjoyment (r(191) = .359, p < .001), and perceived utility (r(191) = .467, p < .001), and perceived cost (r(191) = - .220, p = .002, work avoidance and perceived competence (r(19 1) = .420, p < .001) (Hyps.9,10 a,c,d,e,b) . In line with expectations, relational regulation was neither positively correlated with the motivational constru cts cost and work avoidance, nor with interpersonal attraction . While relational regulation did produ ce a small positive statistically significant correlation with interest/value (r(191) = .420, p < .001), utility (r(191) = .260, p < .001), and compet ence (r(191) = .296, p < .001), these correlations disappeared or were reduced 83 to marginal statistical sig nificance when accounting for the substantial correlation between epistemic and relational regulation observed in this study, r = .47, with partial correlations or ( with interest/ enjoy (r(191) = .13, p = .07), utility (r(191) = .074, p = .31 ), and compet en ce (r(191) = .14, p = .052), Finally, sociocognitive conflict theory suggests positive linkages between epistemic conflict regulation and achievement (Hyp.11 a - c ), which the data supports. Small but statistically significant positive correlations between ep istemic conflict regulation and content knowledge (r (191) = .148, p = .04), length of final statement (r (191) = .177, p = .01) and integration level of final statement (r (191) = .283, p < .001) were found, whereas relational regulation was not positivel y related to any indicator of achievement. Table 6 : Intercorrelations of dependent variables Notes: DesNorm = Perceived descriptive norms; StepsCompl =steps completed; EpReg = Epistemic Conflict Regulation, RelReg = Relational Conflict Regulation; IntEn joy = Interest/Enjoyment; Attract = Attraction; WorkAvoid = Work Avoidance; FinalLength = Length of final statement draft (characters); IntegratLevel = Level of Integration of final statement draft. 84 Hierarchical Multiple Regression for M oderation of Nor ms E ffects Based on the theory of normative social behavior, hypothesis 3 predicted that any effect of perceived descriptive norms on completion rate and/or timeliness of responses would become stronger with increasing group identification (Hyp.3 a ) and inc reasing ego - involvement (Hyp.3 b ). Hierarchical multiple regression with a multiplicative interaction term of either group identification and perceived descriptive norms or of ego - involvement and perceived descriptive norms were used for both behavioral dep endent variables, timeliness of responses and completion rates. For both these dependent variables, perceived descriptive norms were entered in the first block, followed by group identification in the second block and the interaction term between descripti ve norms and group identification in a third block. This procedure was repeated for the second potential moderator, ego involvement (first ego involvement good student, second ego involvement good communicator). Support for the moderation hypotheses wa s not found for either of the dependent variable or either moderator (see tables 7 a,b,c and 8 a,b,c ) . In all cases, the full regression model was significant but this was due to the first term entered, descriptive norms only. For timeliness, the first model (descriptive norms only) explained a modest 5.6 % of variance in the dependent variable (F (1, 191) = 11.380, p = .001, R = .237; and for completion rate, the base model explained an equally modest 5.6% of variance in the dependent variable, F(1, 191) = 1 1.267, p = .001, R = .236. Adding the interaction term to the model led to only trivial and statistically nonsignificant increases in explained variance, with F - change values (1, 189) = 1.173, p = .19 for the potential moderator ego involvement (student) , F - change (1, 189) = 2.529, p = .11 for ego involvement (communicator), and F - change (1, 189) = .006, p = .94 for group identification levels for timeliness. 85 Table 7 : Normative p redictors of t imeliness Group Identification: r (%) R2 (%) F cha nge Sig F Step 1: Descr Norms .237 .056 0.56 11.380 .001 Step 2: Group ID .241 .002 .058 .422 .517 Step 3 : DN x Group ID .242 .000 .058 .006 .940 Ego Involvement (student) r (%) R2 (%) F change Sig F Step 1: Descr Norms .237 .056 .056 11.380 .0 01 Step 2: Ego Inv St udent .240 .001 .058 .260 .611 Step 3 : DN x EI Student .257 .009 .066 1.731 .190 Ego Involvement (communication) r (%) R2 (%) F change Sig F Step 1: Descr Norms .237 .056 .056 11.380 .001 Step 2: Ego Inv Comm .247 .005 .061 .932 .334 Step 3 : DN x EI Comm .271 .012 .073 2.529 .113 Notes: Entries in the first column are the zero - order Pearson correlations between timeliness of posting (0 - 5 timely posts) and each predictor. Entries in the second and third column quantify the c hangein timeliness variance explained and the absolute value in variance explained by all variables in the current and previous block(s). Group ID = group identification (similarity); D N = perceived descriptive norms; EgoInv Student = Ego Involvement in t he activity via self - perception as a good student. ; EgoInv Student = Ego Involvement in the activity via self - perception as a good communicator . Similarly, for completion rate, adding the interaction term to the model did not explain additional variance , with F - change values (1, 189) = .07, p = .79 for the potential moderator ego involvement (student), F - change (1, 189) = 1.366, p = .244 for ego involvement (communicator), and F - change (1, 189) = .642, p = .42 for group identification. A ceiling effect f or step completion places strong limitations on the variance available for explanation (see Table 5). Interestingly, adding ego involvement (communicator) to the model for completion did lead to a small, but statistically significant increase in variance e xplained in this dependent variable, F - change (1, 190) = 5.693, p = .02, which might suggest that the extent to which students perceive themselves to be good communicators potentially is related to their completion of the 86 controversy steps. But given the l imited variability and the resulting deviation from normality in the completion rate variable, all results related to it need to be interpreted with caution. Table 8 : Normative p redictors of step c ompletion Group Identification: r (%) R2 (%) F c hange Sig F Step 1: Descr Norms .236 .056 .056 11.267 .001 Step 2: Group ID .238 .001 .057 .203 .653 Step 3: DN x Group ID .245 .003 .060 .642 .424 Ego Involvement (communication) r (%) R2 (%) F change Sig F Step 1: Descr Norms .236 .056 .056 11.267 .001 Step 2: Ego Inv Comm .288 .027 .083 5.693 .018 Step 3 : DN x EI Comm .300 .007 .090 1.366 .244 Notes: Entries in the first column are the zero - order Pearson correlations between step completion (0 - 5 timely posts) and each predictor. Group ID = group identification (similarity); DN = pe rceived descriptive norms; EgoInv Student = Ego Involvement in the activity via self - perception as a good student; EgoInv Student = Ego Involvement in the activity via self - perception as a good communicator. Ego Involvement (student) r (%) R2 (%) F change Sig F Step 1: Descr Norms .236 .056 .056 11.267 .001 Step 2: Ego Inv St udent .259 .012 .067 2.349 .127 Step 3 : DNx EI Student .260 .000 .068 0.070 .791 87 C HAPTER V: DISCUSSION Longstanding resea rch programs on the effectiveness of cooperative learning procedures such as constructive controversy have produced ample evidence of their potential to promote learning and a host of desirable related outcomes (Johnson & Johnson, 2009). As one of the olde st areas of research in American social psychology, research on cooperative versus competitive efforts in general , along with research theory specifically have provided strong support for some specific processes t hat drive these outcomes (Johnson & Johnson, 2005). However, these mechanisms do not seem to automatically translate into equivalent processes in asynchronous computer - mediated communication contexts, as commonly found in online courses (Roseth et al., 201 1; Saltarelli et al., 2014). Additionally, increased self - regulation is needed for asynchronous online learners who need to respond to their partner outside of a structured setting that provides visual and auditory the task. When present, such cues provide both informational influence (i.e., influence based on evidence about reality) and normative influence (i.e., influence based on the desire to conform to positive expectations of the environment, e.g. Gerard & Deut sch, 1955), making their absence in the asynchronous CMC setting a potential factor for reduced engagement. This study addressed these issues by experimentally manipulating both distal and proximal social influences on learners in such a context, the form er through explicit information ng to messages peers can use to foster promotive interaction patterns by tapping into a congruent set of cooperative skills, virtual group rules, and socio - cognitive interaction moves . 88 Behavioral Prevalence I nformation, Social N orms , and E ngagement Factorial analyses With regards to the social norms factor that was manipulated in the study, i.e., the daily overall completion percentage update s, exposure to such group - wide participation levels resulted in higher perceived norms as expected. But interestingly, the presence of cooperation - centric partner messages had the effect of raising perceived course norms regarding regular engagement in the activity to the same extent as the normative messages did . This was not originally hypothesized as these partner messages were not geared at altering perceptions of the typical cooperative nature. constant across the conditions. Various post - hoc explanations are possible. For example, Taggar ers in shaping team norms suggest that team regarding work norms . Although these authors focus on injunc often what is typicall added ; also see Deutsch & Gerard, 1955 ; Eriksson, & Strimling, 2015 ) , it is very conceivable that a partner using cooperation - centric messages acknowledging contributions, eliciting input, and pointing toward the shared goal to some extent takes on the role of a good leader and thereby affects . equally regular participation may appear less normative if t his partner behaves less leader - like, i.e. if they are simply doing their part without using any of the additional cooperation - promoting messages. A related explanation may lie in mechanisms along the lines of the human tendency to overestimate the represe ntativeness of small samples ( Tversky & Kahneman, 1971) here, the 89 partner and to apply availability heuristics ( Schwarz, Bless, Strack, Klumpp, Rittenauer - Schatka & Simons, 1991 ; Tversky & Kahneman, 1973): To some extent, the task may become more salie nt when a partner uses statements that add to the provided instructions and rationale, and therefore is perceived as more commonly carried out. Additionally, it is possible that the reliable presence by the confederate served as a particularly salient mode l for inference making regarding the rest of the pairs engaged in the task under the cooperation - promoting message condition. Someone displaying more engaging task behaviors is not only more likely to serve as a salient model for inferred performance by ot hers but may further increase the salience of the presented normative information, as compared to someone who is equally punctual and reliable but limits their contributions to the bare - bones substantive input to keep the activity going. Specifically, the former work sense of guilt about being behind those eighty or so percent of participants who have already completed the current step of the activity, making the percentage completion information more salient to respondents who feel a stronger sense of obligation to their engaging partner (M. Lapinski, personal communication, August 25, 2015). Obviously, t hese effects as well as possible alternative mechanisms explaining them would require carefu l empirical clarification. However, the evidence from the present study suggests a potential ly desirable side effect of students using cooperation - centric messages while working on asynchronous computer - mediated versions of cooperative learning tasks, as d iscussed further below. However, in the present context, t hese differences in normative perceptions did not translate into behavioral differences in terms of completion rates and timeliness of daily step completion in this study, reflecting Lapinski et al 7) statement that while modifying perceptions about descriptive norms via normative restructuring strategies tends to be possible, 90 the next step in the process a modification in behavior is a much less certain affair , as examined in some depth below in the implications section of this chapter. The possible moderat ion of effects of descriptive norms on behavior, including those that were not assessed in the present study, needs to be considered, and methodological limitations such as ceiling effe cts in both of the behavioral engagement measures, also expanded upon below, reduce the explanatory power of this first factor. However, given the role of the nature of the behavior and the context, there is reason not to discount norms effects on behavior in cooperative learning online just yet: After all, the situational ambiguity that usually makes us receptive to the informational and normative extent to which others are engaging in a behavior (Cialdini, 2001; Deutsch & Gerard , 1955; Lapinski & Rimal, 2 005) is very much present in the context of an asynchronous, text - based version of cooperative learning task. An additional point for consideration lies with the origin of the norms which in the present study can be less clearly categorized as proximal o r distal, with the former being more Borsari & Carey, 2003) with health related behaviors such as drinking alcohol. Yun and Silk (2011) classified college peers a classmates or peers who are currently also participating in the controversy activity as relatively oes little to explain the negative motivational effects in this context and the marginally reduced enjoyment of using the interaction website. Similarly, due the fact that participants in the activity had to be recruited from more than a single course in o rder to reach sufficient numbers, a slight other participants most students in my classes 91 be allowed, as a survey question about perceived similarity with other participants would have been difficult to address by the individual participant, unlike the more generic stand - in about classmates in general. The decision to allow for this mismatch was made based on the exp reference group used here, since participants knew that additional participants were recruited from similar courses. However, it is possible that this slight mis match attenuated any influence of perceived similarity on the effect of descriptive norms, and ideally, future work would avoid this mismatch by recruiting one or more intact large course(s). Of theoretical and practical importance in this context is the large role that reliably contributing partners seem to play, along with the system - generated reminder email of the of participants completed the activity if a ssigned to asynchronous conditions and to Saltarelli et al., (2014) where only 54.3% of students working asynchronously completed the activity , completion rates in the present study were much higher: 92.5% of students (210 out of 227 who started the activi ty) completed the five steps of the activity. Compelling post - hoc inferences can posting behavior as dwarfing effects of social norms that are based on the overall g roup that is participating in the activity, especially given the fact that opting out of the activity halfway As mentioned above, the reliable participat ion of a peer also engaging in cooperation - promoting interpersonal communication may have a particularly strong effect, but given the ceiling effects on the engagement measures in this study, this moderation hypothesis cannot be directly 92 a ddressed with the current data. However, the incidental evidence for this possible effect aligns with the relevance of peer - generated influence discussed after the next section. Correlational and moderation analyses Besides these factorial results of direct effects of the provision of norms on perceptions and behavior, correlational analysis of observed differences in perceived norms and these outcome variables were also considered. A small positive correlation was found between perceptions of other participants as moving through the activity according to schedule and completion rates and timeliness of step completion. Conscientiously performing participants (high completion and timeliness) tended to perceive the descriptive norms for engagement as higher than did participa nts with less conscientious performance, but of course the causal direction cannot be clearly established based on these findings. After all, besides being rooted in direct observation and communication with others, perceptions of the prevalence of a certa in behavior are formed in part based on social projection processes (Miller & Prentice, 1996): we are likely to think that others think and act as we do, as sometimes evident in a false consensus effect (Mullen & Hu, 1988). Therefore, the association betwe en perceived norms and completion/timeliness may be based less on beneficial norms - intervention outcomes and more on a heuristic process that r uns in the opposite direction, in that s tudents who complete task steps in a more timely and reliable manner gene ralize from their own behavior to create their perceptions this study, e.g. individual differences in conscientiousness or cynicism might drive both variables . The two moderators from the theory of normative social behavior tested here, i.e., group identification and ego involvement, were not found to moderate the strength o f this linkage. The 93 tradeoff needed for the operational definition of similarity has be en mentioned above. An additional reason to interpret the s tatistical ins ignificance of these moderators cautiously lies in the low power of observational designs for detecting interactions (e.g. McClelland & Judd, 1993; Trautwein, Marsh, Nagengast, Lüdtke , Nagy, & Jonkmann, 2012). The scarcity of extreme profiles, i.e. in this context, participants who scored very high on perceived norms and very low on group identification or ego involvement, or vice versa, is a likely factor here . After all, a small posi tive correlation of around r = .20 existed between perceived norms and each of the two measured moderator variables in this study, suggesting a relative scarcity of such extreme profiles as needed to generate sufficient variability on each factor per level of the other factor. The issue of compounded unreliability of the crossproduct term used in the analysis is further reducing statistical power to detect moderation via this regression approach (Busemeyer & Jones, 1983). Additional effects of b ehavioral p revalence information The hypotheses formulated for the descriptive norms factor centered mostly on behavior, specifically engagement adherence to the daily posting timeframe. However , results showed that the presentation of course norms did have unexpected side effects on a two of the motivational constructs that were hypothesized to be influenced mostly by the second experimental factor, cooperation - centric partner messages. Specific ally, interest/enjoyment and perceived competence were negatively affected by the presentation of norms, presumably negating the hypothesized positive effects of cooperation messages. a prescribed behavior can be interpreted might account for some of these unexpected effects: T he very 94 informational value that recipients of norms messages might appreciate as potentially useful for themselves (Cialdini, Reno, & Kallgren, 1991; Gerard & De utsch, 1955), might be more negative in the context of academic settings in which social upward comparisons with a high - - related self - concepts and subject - related inte rest ( Marsh et al., 2008 , p. 510 ; Trautwein et al., 2006 ) which is, on a very small and short - term sca le, consistent with the results from the current study. This big fish, little pond effect (BFLP E) (Marsh & Parker, 1984; Marsh, 1987; Marsh, Seaton, Traut wein, has a rich research base supporting that others can have detrimental effects on self - concept . While it is most often tested by using course - wide average grade information , it theoretically extends to mere task completion rates as well (R. Pekrun, personal communication, April 21, 2015). The detrimental effect s of being informed about a high prevalence of timely st ep - completion on both perceived competence and task interest/ enjoyment found in this study align with such effects, although more specific tests are needed. Complementing these considerations, aversive motivational effects of normative messages can be cons idered from the perspective of cognitive evaluation theory (e.g. Deci & Ryan, 1980; Ryan, 1982) as well: Rather than focusing on the mechanism s of guilt or upward social comparisons , this theory highlights the potential for feedback messages to be interpre ted as either informational (just telling the recipient where they stand, compared to others) or as controlling (e.g. as geared at getting them to do something). To the extent that the feedback is interpreted as having the latter intent , intrinsic motivati on is typically diminished , and longstanding research exists regarding the aversive effects of surveillance ( Lepper & Greene, 95 1975; Pittman, Davey, Alafat, Wetherill, & Kramer, 1980 ), deadlines ( Amabile, DeJong, & Lepper, 1976 ) and explicit competition (De ci, Betley, Kahle, Abrams, & Porac, 1981; as cited in Ryan, 1982) , all of which are to some extent supported by the percentage updates. Cognitive evaluation t heory explains this reduction in intrinsic motivation in that external constraints can shift the p erceived locus of causality for a behavior from i ntrinsic to extrinsic factors. By contrast, perceived choice in engaging or not engaging in a behavior supports a shift in perceived locus of causality from external to internal factors, increasing intrinsic motivation. Fittingly for the present context, changes in perceived competence are seen as a second process supporting competence strengthen, while conditions that r educe perceived competence weaken intrinsic motivation (Deci & Ryan, 1980; Ryan, 1982). In the context used here, participants exposed to am doing this to keep ecause it is explanation is similarly in line with the observed reduction of intrinsic motivation. Similarly, the extent to which participants may feel that they are less competen t at the task than the consistently punctually performing generalized other that they compare themselves to, competence evaluations and intrinsic motivation reports are expected to suffer from these comparisons, as specified by the BFLPE. Cooperation - P rom oting Messages and S ocial I nterdependence The use of partner - generated cooperation - promoting messages was hypothesized to be crude law of social relations , ex posure to messages that represent some fundamental (albeit 96 sa lient the shared goal, was hypothesized to increase perceptions of cooperation and decrease perceptions of competition and individualism. With the exception of the hypothesized reduction in competition, these predictions were supported. In the limited - cues environment that is created by asynchronous, text - base d computer mediate d communication, the deliberate, repeated use of the explicit cooperation - effects of cooperative processes equate some of the conditions that give rise to them hold true also in this context . Although unexpected, the fact that cooperation - centric messages did not actually lower competitive perceptions (which were correlated wit h cooperative perceptions, r = .56, p < .01) does align with similar findings by Saltarelli et al. (2014), who found no support for their prediction that an experimentally increased sense of belongingness would reduce competitive perceptions among college students working on a constructive controversy task asynchronously online, and with Tichy, Johnson, Johnson, and Roseth (2010) who found that participation in a constructive controversy activity actually heightened 3 rd - through 5 th graders perceptions of t he desirability of competition when compared to participation in an individual learning task. This co - occurrence of cooperative and competitive perceptions is considered further in the implications section below. The use of cooperation - centric messages by did produce the expected increase in attraction , a particularly relevant motivational construct in online learning contexts positive outcomes (e. g. Gunawardena & Zittle 1997; Hsu, Ju, Yen, & Chang, 2007). This is in - and social 97 It is reasonable to expect that the effect found here would be more pronounced if a more task - related measure of attraction had been used instead of the generic attraction measure, the former specifically - related attraction was e - adherence than the more generic social attraction measure also used. However, given that SIT makes predictions about the relational quality more broadly, and given the importance of social relationships and a sense of belonging for achievement, motivatio n, and engagement (Deci & Ryan, 2000; Furrer & Skinner, 2003; Martin & Dowson, 2009; Roseth, Johnson, & Johnson, 2009) , a focus on attraction generally seemed warranted here. The effect of C PMs on motivation constructs was more complex, as alluded to in t he previous section on norms. Interest/enjoyment was highest when partners used cooperation - this effect was complicated by a significant interaction with the no rms factor in that exposure to norms cancelled out any beneficial effect of cooperation - centric messages. This pattern was mirrored with perceived competence as the dependent variable. This interaction of these two influence factors will require follow up research to clarify. outcomes are fostered by cooperative interaction, perceived utility and work avoidance were affected in the expected directions, albeit with marginal st atistical support. Perceived cost was the least amenable to influence by the cooperative messages, but the proximity of the data collection to final exams for many participants likely places limitations to the strength of the 98 inference here. Cost, i.e. the perception that engaging in a target task is associated with sacrifices to be more firmly anchored in the external conditions of the participant than are perceptions of the potential utility o r interest value of the task. Of particular interest among the results are the behavioral engagement measure of final statement draft length and the behavioral achievement measure of integration level of final statement draft. Students exposed to cooperat ion - centric messages wrote longer integration statement drafts during the final step of the activity. Given the fact that participation fatigue is a real concern in this experimental design - this was the fifth time that participants had to take the time t o work on the task even small increases in length are noteworthy. Importantly, cooperation - centric messages also led to slightly higher integration level scores, mirroring herence . The final achievement measure, based on the four - item content knowledge quiz, did not produce differences between groups, but concessions were made regarding the quality of the measure due to space limitations of the already demanding final sur vey, to which a more reliable and valid measure would have added even more length. The fact that content knowledge assessed here was limited to relatively low levels of knowledge (Bloom, 1956; Anderson, 2000) may have contributed to a lack of differences b etween groups as well . Assessments of higher levels of knowledge (including accuracy and creativity of problem solving, higher - level reasoning) may be even more relevant and appropriate targets of cooperative - learning - based influences as identified in prev ious meta - analytical work (Johnson , Johnson, & Smith, 2006; 2014) and should be considered for use in future research . The lack of an effect on content knowledge found here should therefore not be interpreted as evidence against the potential for 99 differenc es on this area of achievement as proposed by social interdependence theory. However, more long - term research designs and more thorough assessment of content knowledge will be needed to address this linkage, for which the present study had limited capacity . The Role of Conflict R egulation In line with sociocognitive conflict theory, perceptions of the task as cooperative were positively related with epistemic conflict regulation, i.e., a focus on the issue at hand and on reaching the best solution rather than a focus on contradicting the opposing viewpoint, i.e., relational conflict regulation. Also in line with the theory, perceiving the task as competitive was related to the latter conflict approach. These findings replicate Saltarelli et al. (2014) , which s in the context of asynchronous online constructive controversy and produced similar correlation coefficients as the ones found in this study. Further support for the applicabil ity of sociocognitive conflict theory in the context of asynchronous online constructive controversy can be found in the expected association of epistemic regulation with perceived relational quality, the broad range of motivation constructs tested here, a nd three measures of achievement, along with the absence of such associations between relational regulation and these three outcome areas. While the correlation matrix provided in the results chapter at first glance seems to suggest some exceptions to the expected negative correlations between relational regulation and motivational outcomes ( Relationa l regulation actually correlates positively with some motivational variables , such as enjoyment of the task, perceived utility, and perceived competence ), thes e correlations become statistically insignificant when controlling for epistemic correlation , which was , unexpected ly , positive ly related with relational regulation and , expectedly, with motivational variables , and seemed to be 100 driving these unexpected ass ociations . This spurious correlation via a correlated, theoretically beneficial construct seems to mirror the positive correlations between competitive perceptions and some motivational outcome variables found here as well, which were gr eatly diminished by controllin g for cooperative perceptions. Limitations and D elimitations of the S tudy threats to internal and external validity, methodological concerns, and theoret ically and practically important questions that remain unanswered by the current design. These will be briefly considered along with their implications for further research. T radeoffs between internal and external validity ( Onwuegbuzie & Johnson, 2006; Ro e & Just, 2009 ) come in to play especially with regards to the second manipulated factor, presence versus absence of cooperation - centric messages. While this scripted, confederate - based approach can be seen as watering down the educational benefits of the procedure as designed by Johnson and Johnson in that stock responses cannot generate as coherent a flow of argumentation as would be ideal, observations on prior related research and the pilot for this project indicate that the use of stock responses appro ximates what an average participant would do, giving this approach a certain external validity. More importantly, the fact that both the cooperative partner message containing conditions and the control conditions follow the same argument order ensured tha t any compromising of the flow occurred to the same extent across conditions, strengthening the internal validity of the experimental approach. The viability of the confederate approach is supported by similar studies (e.g. Darnon, Doll, & Butera, 2007). B esides allowing for the systematic varying of the levels of cooperative partner behavior, the confederate approach 101 t have been paired with. Since chronemic provides rich grounds for inference making about the quality of the relationship (e.g. Walther, 1992 ; 2002 ) and therefore likely affects many o f the variables of interest in this study, it was decided to keep the partner reliability, depth of processing, and, to the extent that it is logistically possible, the response lag constant (the latter within a window of ca. 12 hours) in order to alleviat e some concerns about the partner - dependence of th e individual measurements used. Of course the use of confederates in dyadic research (and by extension, the use of response scripts) has its downsides that are more general than those pertaining to the natu re of the activity, such as a risk of violating interactional rules for the sake of treatment consistency (Duncan & Fiske, 1977). The fact that a few participants indicated doubts about the authenticity of their partner validates their concern. Experimenta l follow up work could involve training one half of participants to include the target behaviors at each step, followed by an imp lementation fidelity assessment/manipulation check and analysis of dependent variables via dyadic data analysis procedures that account for the lack of independence of measures based on real student pairs (Kenny, Kashy, & Cook, 2006). Extending on considerations of internal and external validity, the nature of the sample used here introduces limitations: For external validity, th e fact that even though it achieved fairly good representativeness of U.S. undergraduates in terms of demographics such as gender, age , ethnicity (U.S. Department of Education, 2015) , with the full range of class level s and a breadth of different majors be ing included . I ts recruitment from face - to - face courses, however, potential ly limits the generalizability of findings to the true context of interest the effects of these interventions and their mechanisms in fully online courses. Students currently enro lled in 102 online courses are particularly likely to already have previous experience with online study, and, especially if the online course is part of an entirely online program, to have very targeted, career - related motivations for taking the course (Cline felter & Aslanian, 2015), which might affect the way they interact within course assignments. We also know that over time, users of text - based online communication become better adapted to the limitations of the medium (Utz, 2000; Walther & Parks, 2002), a nd that anticipation of future interaction greatly affects relational communication (Walther, 1994), mechanisms that may affect and interact with this type of cooperation centric messages in important ways. Future research should either recruit typical onl ine learners to replicate effects of the factors tested here, or measure these potentially relevant dimensions when recruiting from traditional college courses. In terms of internal validity, the need to combine students from various courses for each itera tion of the activity introduced the similarity assessments to address potential moderation needed to be made with regards to a group that participants could concept Several methodological limitations have already mentioned above, such as shortcomings in one of the achievement measures (content knowledge) as a result of survey space limitations and ceiling effects in t wo of the engagement variables, completion of steps and timeliness of step completion . Future studies might be able to rely on trace data to better capture details of posting timeliness, and the use of actual dyads is likely to produce more variability in the completion rate variable to be explained by the independent variables of interest here . The needed compromise leading to a mismatch between referents for different measures has been discussed and is considered further below , a nd so have the shortcomings o f regression - based moderation tests in the absence of factorial results another possible effect of the limited variability in the 103 engagement measures used. Sampling error is a possible factor for some of the nonsignificant findings as well. For instance, for several motivational variables, such as perceived utility, cost, and work avoidance, means did differ in the predicted direction as a result of the cooperation - centric message factor, but the large variances of these variables necessitate additional c larification of such effects. It is important to note that most of the limitations on inference making are conservative in nature, i.e., they potentially weaken the power to detect differences and effects that may exist. Several of these shortcomings woul d be effectively addressed by using a more natural context, i.e. embedding the experimental task in a large - scale online course using a controversy that directly relates to the subject of study along with authentic assessment of achievement etc. The study delimitations , i.e., the related questions that are not being answered by this research, are equally informative in planning future work. First, while all hypotheses tested in the study are directly derived from relevant theory, the study does not attem pt to comprehensively test the many mechanisms proposed by these theories in the context of asynchronous online controversy. Additional analyses are needed to follow up on the link between cooperative perceptions and cooperation - reinforcing actions, psycho logical processes, and interactive benign spiral that can be contrasted with competition - reinforcing destructive spiral . Transcripts generated by text - based online versions of constructive controversy would lend themse lves to content coding to explore differences in interaction patterns email exchanges cogniti ons and actions behind the scenes would require additional self - report measures, with the usual caveats regarding limited accessibility of own mental processes ( Nisbett & Wilson, 1977). 104 For behavioral coding, a similar category system to the one used as a basis for the second independent variable here (Rourke et al., 1999; Fahy, 2001, 2002, 2003; Kim, 2009) could be used to assess the extent to which the presence versus absence of these types of behaviors are associated with different amounts of such inter active/supportive or cohesive statements in return. A second delimitation lies with the norms - related manipulation in this context. Information re the bare - bones information ab out daily completion rates used here , as completion rates alone some other interventions geared at boosting engagement through real - time updates on the rest of the class or team have done (e.g. Ganoe, Somervell, Neale, Isenhour, Carroll, Rosson, & of your peers have down have accessed t hree or more of with the task rather than its mere occurrence. To the extent that technical feasibility of the latter can be expected to grow, it will be intere sting to examine effects of such richer behavioral updates. Implications While further research is needed to clarify the open questions raised above, this study makes several contributions and has strengths that should be noted. First, the use of random assignment, the experimental control design, and the systematic use of scripts ensure that aside of the two factors of interest here, other variables such as e.g. the substance and amount of controversy input at each step or the response lag did not differ systematically between groups. Second, its hypotheses are directly derived from theories with strong records of validation - 105 particularly in the case of social interdependence theory, a long history of validation in the context of traditional use of const ructive controversy in the classroom allows us to explore to what extent those linkages that do not translate well to asynchronous computer - mediated versions of this kind of procedure are amenable to two kinds of communication - based intervention. Implicat ions for theory The theor y of normative social behavior According to the theory of normative social behavior, possible effects of descriptive norms on behavior are affected by outcome expectations for the behavior, social approval or disapproval for compl iance versus noncompliance, identification with the group that constitutes the norms , as well as ego - involvement in the task. In this study, the first two of these moderators (outcome expectations and injunctive norms) were not assessed for practical reaso ns, so that generally low values on either construct could have hampered the effect of the provided norms on behavior (e.g. Rimal et al., 2005). Given that outcome expectations for a pedagogical procedure aimed at fostering a dialectic approach to thinking and constructive, epistemic approach to conflict is less clearly beneficial to the individual than many other behaviors health - related and otherwise that are commonly examined within the social norms literature, the essential role of outcome expectati ons as a moderator of the descriptive norms - behavior link are indirectly highlighted by this study. Similarly, given the anonymous nature of the activity involved here , injunctive norms in particular are likely to be relatively low, compared to e.g. injunc profile online was something that was explic the role of interaction effects in this context. 106 More direct implicatio ns, however, exist with regards to the nature of the target behavior under consideration here. Research regarding the TNSB also highlights the importance of behavioral attributes, such as whether the behavior in question is enacted in public or in private and open to scrutiny by members of the reference group or hidden (Lapinski & Rimal, 2005) . The behavior examined here reliable and persistent engagement in the cooperative learning task can be less clearly assigned to an endpoint on these continua: Whi le the behavior is enacted individually and away from the rest of the group, the dyadic (and by design interdependent ) nature means that there is at least en a new post has been made, and While inferences regarding the role of behavioral prevalence information in shaping desired behavior in the context of cooperative learning via asynchrono us computer - mediated communication were limited by ceiling effects of the behavioral engagement measures here , the heightened perception of descriptive norms exhibited by participants whose partner used cooperation - promoting messages are of interest in lig ht of this behavioral attribute . As suggested above, these cooperation - promoting messages may serve to boost the salience of the reliably performing (confederate) partner as a model for general performance on one hand, which would suggest that perceived pr evalence of behaviors that involve another person can be determined by additional elements beyond those that are purely individual in nature (Miller & Prentice, 1996 ; Wenzel, 2005 ) . Perceived interpersonal influence to engage in a particular target behavio r may directly operate as a heuristic to assess behavioral prevalence when such information is less directly available. As such, there are opportunities for interactions between distal - and proximal sources of influence that are unique to behaviors involvi ng dyads (or, by extension, small groups 107 the behavior of which the target individual is familiar with or whose preferences for a particular behavior are clearly known ). Such interactions of the distal normative effects with proximal influence may also mak e the former (normative messages) harder to ignore , which speaks to the interaction effects of behavioral prevalence information and cooperation - promoting partner messages on motivational constructs: The reductions in intrinsic motivation (interest/enjoyme nt and competence) found for participants whose partner was using cooperation - promoting messages when also exposed to the completion percentage updates further suggest unique and theoretically important considerations regarding the behavioral attribute of interdependence in the context of descriptive norms effects. in the task may induce a higher sense of guilt when not conforming to behavioral norms in this study, a sense of guilt for being behind the alleged majority of participants who have completed the current step of the activity, making the percentage completion information not only more salient but also potentially emotionally aversive to respondents who feel obligated to the other person(s) involved . Both descriptions of social interdependence allude to such This is in line with t he reductions in enjoyment and perceived competence for those whose partner was particularly engaged and engaging via the use of cooperation - promoting messages. Therefore, the degree to which a behavior exhibits interdependence might operate as a moderator of the effects of descriptive norms as well. Future research is needed to assess the potential of such interdependence to impact the descriptive norms behavior link, and to validate the descriptive norms competence link (also see Rimal et al. (2005) f or existing 108 research on the effects of similarity on competence within the TNSB framework). The big - fish - little pond effect and aversive perceptions of being control led T he shift of a perceptions of norms messages as holding informational value (Cialdini, Reno, & Kallgren, 1991; Gerard & Deutsch, 1955) toward potentially more negative, self - concept harming perceptions in the context of academic settings provides an alternative or complementary account of the observed results: t h e data are consistent with t he expected reductions in interest and perceived competence described by previous research involving social upward comparisons with a high - Marsh et al., 2008 , p. 510; Trautwein et al., 2006 ). While this research examined eff ects on interest in school subjects, such as math, instead of interest in a cooperative learning task with generic content, and a semester - long time frame, instead of a one - week activity , i t appears that mechanisms proposed by these frameworks can operate on a smaller, more short - term scale as well . A caveat is that this aversive role of upward comparisons (with other participants, a majority of which is consistently ahead of the participant in the behavioral prevalence update condition) only explains the m ain effect s of reduced interest/enjoyment and perceived competence under this condition, not the interaction by which the reductions are specific to the cooperation - promoting messages condition. This main effect may also be relevant in that the mean levels of competence and interest/enjoyment for participants in the cooperation - promoting - messages - off condition may be at a floor level, so that the addition of normative information cannot do further damage, so to speak, as would be expected in the absence of interaction effects. The same goes for the possible role of interpreting the percentage updates as controlling, as suggested by cognitive evaluation theory: While the reduction in perceived intrinsic ism in this present context for the 109 cooperation - promoting - message condition ( Deci & Ryan, 1980; Ryan, 1982 ) , the absence of further reductions in motivation for the cooperation - promoting - messages - off conditions through potentially controlling completion pe rcentage messages is not further explained by this theory (but may, again, be rooted in a floor effect for motivation in CPM - off groups). Of additional interest here is the timing of the norms messages: Participants received them while they were logging i n to the activity website, i.e., at a time when they were already engaging in the target behavior (as opposed to, e.g., as part of the reminder emails that alerted most recent post). To the extent that being explicitly guided toward a specific behavior as one is already engaging in reduces intrinsic motivation to do so (Frey, 1997), the timing of the norms message needs to be varied systematically in future research in order to address the role of its interpretation as controlling an d inducing a change in perceived locus of causality from internal to external (Ryan, 1982, p. 450) , thereby reducing intrinsic motivation. Obviously, it is very plausible that the various proposed mechanism s that are supported by the data of this study guilt at falling behind, change from internal to external locus of causality, and upward social comparisons operate in concert in the context of online cooperative learning via constructive controversy , and other behaviors as well. Taken together, these considerations further highlight that different outcomes beyond behavior performance as dependent variables in the TNSB framework can be differently affected even by the same combination of core mediators due to the influence of additional factors such as these. Social interdependence theory A first implication for social interdependence theory lies in the replication of the beneficial effects of cooperative goal structures, related to numerous motivation constructs as 110 well as engagement and achievement outcomes in the limit ed - cues environment created by asynchronous CMC. Second, the role of competitive goal structures is more complicated in this con text, in that a substantial correlation between cooperative and competitive perceptions exist s when engagin g in constructive controversy asynchronously online. In contrast to previous constructive controversy studies that had divided participants into conditions either involving asynchronous - CMC or various oth er formats (Roseth et al., 2011; Salt a relli et al. , 20 14) , all participants in the present study communicated exclusively online and asynchronously, so that the higher correlation between cooperative - versus competitive goal structures found here may indicate that such goal structures are less mutually ex clusive when working with a partner in that context and that there is more overlap between antecedents and effects of either kind of perception as a result . Results of this study do not directly explain the mechanisms driving such co - occurrence, but existi ng research describes so me possibilities. For example, quasi - experimental comparison s of anonymous - online versus FtF competition on a game designed for social studies and science practice produced higher self - reported motivation and satisfaction with the a ctivity in students participating anonymously ( Yu, 2003; Y u, Han, & Chan, 2008) , leading networked team competition where opponents are not identified may . In their overview of social interdependence theory, Johnson and Johnson (2005) describe some of the findings regarding side - effects of competitive goal structures that account for the frequently observed losses in self - protective strategies such as self - worth protection, self - withholding of effort in order to be able to attribute failure to not having tried rather than a lack of competence. It is plausible that such mechanisms lose their detrimental effects on achievement 111 experience a reduced need to self - protect in such detrimental ways. Similarly, Pareto, Haake, Lindstr n, & Gulz, (2012 ) describe how students engaging in a virtual math learning activity spontaneously generated several different creative approaches to blending cooperation and competitio n within an computerized learning task , supporting the notion that even i n non - anonymous dyadic interactions, competitive elements could be productively integrated with cooperation . after engaging in constructive controversy activity (com pared to engaging in an individual learning task) also speaks to the cooperation - competition association . Their study illustrates the fact that constructive controversy, a task that is cooperative by design , does not preclude favorable perceptions of compe titive approaches to it in the face - to - face context they cooperative activity aimed at agreeing on the best possible joint decision (i.e., a cooperative outc 781), and speculate that o much fun arguing with each other that a halo effect was created toward intellectual conflict that extended to plausible that participants in this activity are not fully aware of the conceptual distinction (1949, 1962) social interdependence theory. While the theory sees the two as independent from each other, in that intellectual co nflict can be approached under a competitive ( where success for one partner 112 necessitates failure for the other ) or a cooperative goal structure (where success for one partner facilitates success for the other as well), in their everyday usage , these two te rms are probably less clearly seen as orthogonal. Therefore, working with a partner under very explicit cooperative instructions as provided by the constructive controversy activity, it is very likely that the survey , within th e cooperative task structure , and therefore more in the sense of a friendly, playful competi ti ons, along the lines of the next round of the activity. Competition, here , seems to be about in the final analysis, rather than about single - handedly achieving the overall goal while making the partner lose. This corresponds to one of the boundaries of social interdependence theory its original conceptualization (Deutsch, 1949) its assumption o f a single goal (cooperative/ competitive / person always has multiple goals and, therefore, situations are always mixed motive, with individuals often having coop erative, competitive, and individualistic goals in the same . The results of this study and of the related work cited here seem to support that even under the umbrella of the same (cooperative) goal , competitive elements can exist, but de - emphasized - Such a c ompetitive approach to sub - goals in service of the cooperative main goal element needed for cooperative efforts to be successful in that it ensures that both sides of the 113 Slavin, 1988), while at the same time representin g one of the conditions for competition to be that important ). osition , and jointly finding (see Appendix A.3.) in themselves illustrate the em beddedness of competition within the cooperatively structured goal context. The correlation of cooperative and competitive preferences in this task was found to equal extents both for participants who were exposed to the cooperation - promoting partner messa ges and the neutral, facts - only partner messages. An interesting test of the extent to which the constructive controversy procedure supports such a correlation of perceptions would involve a comparison with a group of participants exposed to cooperation - re ducing cues, along the lines of intentionally induced relational conflict regulation For example, rather than highlighting the value of without any such acknowledgment, confederate - generated messages could actively discount , rather than simply ignore, , to delimit t he extent to which it is the structure of the task and to what extent it is the interpersonal communication of the partner balancing out competitive with the needed cooperative perceptions. After all, description of the risk of slipping fr om cooperation into competition in the absence of active and sustained effort to prevent such deterioration is a strong concern within the context of asynchronous CMC where any entrance point into a oppositional pattern is exacerbated by the 114 reduced cue en vironment and the overreliance on the limited cues that are available ( Walther & Parks, 2002; also see Johnson & Downing, 1979). Sociocognitive conflict theory S imilar implication s exist for sociocognitive conflict theory . Replicating findings from previou s work (e.g. Saltarelli et al., 2014), the two conflict regulation approac hes were found to be positively correlated , somewhat mirroring the correlation found between cooperation and competition. For epistemic conflict regulation as for competition, agreem ent to items such as potentially reflects constructive behaviors - turn , - goal level, by thoroughly evaluating each proposed argument , rather than at the level of the overarching goal of creating a mutually satisfying synthesis. The similarities in correlations between competition and r elational conflict regulation and cooperation with epistemic conflict regulation (with Saltarelli et al., 2014) extend beyond the predictions directly a rising from sociocognitive theory in that individualism showed a slight negative association with episte mic regulation , suggesting that the distinction between independent (individualistic) perceptions and interactive - perceptions (both cooperative and competitive) is as at least as important as the original distinction between cooperative and competitive per ceptions that the theory focuses on. As Johnson and Johnson (2005, p. 297) maintain a cooperative system. These costs may be prohibitive when individual action is feasi ble internet literacy skills who are working online on a cognitive conflict that involves information they can potentially gather 115 individually from online sources, the advantages of working with a partner who specializes on a particular side of the issue may be outweighed by the costs of coordinating turn - taking. T herefore, on e implication for sociocognitive conflict theory (Doise, & Mugny, 1984) of the present study lies with the distinctio n of individualistic - versus interactive (both competitive/relationally regulated, in terms of sub - goals, and cooperative/epistemically regulated, in terms of the overarching goal) orientations toward learning . For larger, long - term cooperative learning ta sks that require deeper levels of research into the different sides of the issue, the cost - benefit balance of interactivity versus individual work is likely to shift in favor of interactivity. The role of low - stakes conflict in SIT and SCCT A lthough very aligned with the induced conflict that is at the core of constructive controversy, the mechanism s desc ribed by sociocognitive theory are of course not the only plausible mediator s differentiation of effective versus bungling actions, corresponding psychological processes and promotive versus oppositional interaction patterns allows for many additional mechanisms producing differential outcomes beyond the mechanism of conflict regula tion (Johnson & Johnson, 2005). Relevant to the assumption of competitive goal structures and relational conflict regulation may be the fact that constructive controversy relies on an artificially induced conflict rather than a conflict that naturally aris es out of clashes between deeply held convictions, and that it occurs in relatively small doses of approximately ten minutes per day for five days, which may help further explain why cooperative perceptions did not preclude, nor were they precluded by, com petitive perceptions, as evidenced by the substantial correlation of r = .56 between these two variables in this study . 116 attachment to the position they were arguing for, it is possible that i n those cases were identification with the assigned position was moderate to low, high scores on the competitive perceptions measure may be more indicative of a playful, friendly competition than a verbal fight to the death, so to speak, and a similar atte nuation of relational conflict regulation is plausible as constructive controversy as boosting positive perceptions of competition as well. Implications for prac tice via elaborate technological solutions involving audio, video and real - time interaction (e.g. Aragon, 2010; McLellagn, 1999); however, asynchronous, text - based cour se discussions are likely to continue to play a substantial ro le in online learning. This is due to their high flexibility and low technological demands that make them suitable for learners who are dispersed across different time zones, who juggle various other professional and personal demands, and who join the course with varying degree of technological savvy and equipment (De Wever, Schellens, Valcke, & Van Keer, 2005; Tallent - Runnels et al., 2006), their advantages for discussions that benefit from refl ection and information processing afforded by their asynchronous nature (Harasim, 1993; Pena - Shaff & Nicholls, 2004), and their proven potential to promote critical thinking and social construction of knowledge (Garrison et al., 2000; Ruberg, Moore, & Tay lor, 1996). Therefore, practical implications for such learning contexts can be derived from the two experimental factors used here. First, the study highlights the need to explicitly coach online learners in how to best approach team tasks that either b y design or unintentionally involve conflict potential in an asynchronous, text - based context. Creating an explicit awareness in 117 online students of the frequently unproductive (but easy to fall into) communication patterns is a first and necessary step f or learners to avoid detrimental trajectories likely to hamper their own motivation, engagement, and ultimately, achievement. Providing awareness of productive versus destructive spirals postulated by social interdependence theory and the evidence for simp le yet empirically supported protective behaviors (including, but not limited to cooperation - promoting messages as tested here; see also Kreijns, Kirschner, & Jochems, 2003; Lin & Laffey, 2006; Walther, 1997; Walther & Bunz, 2005; Strijbos, Martens, & Joc hems, 2004) can be seen as another element in the toolbox online learners need to make cooperative elements of their online courses productive for their partners and, via reciprocal effects, themselves. Specifically geared toward constructive controversy , a constructive interactions among six - grade students engaging in the task : More support, more ques tions, better attention to the partner and higher quality of cooperation within dyads overall were observed, suggesting that social support boosting interventions can be effective without consuming large amounts of instructional time or learner attention ( Golub & Buchs, 2014). Pertinent for the context of constructive controversy is research on the effects of self - fulfilling prophecies in anonymous computer - mediated interaction showing that use of disagreements by participants instructed to enact affinity - or disaffinity goals with a discussion partner online not only affected perceptions of attraction by the receiver of such disagreeing messages. Instead, the his or her own behavi or toward the partner, despite the fact that they were fully aware of their goal of intentionally inducing affinity or disaffinity ( Walther, Van Der Heide, Tong, Carr, & Atkin, 2010; Walther & Tong, 2014). To the extent that online versions of constructive 118 controversy with their heavy reliance on disagreements are therefore doubly at risk for relational deterioration, the important role of CPMs is further supported. Clearly, the aforementioned toolbox for cooperative learning success also requires good tim e management, self - discipline, and external factors such as the organization and accessibility of the learning mana gement system used Graham, Cagiltay, Lim, Craner, & Duffy, 2001). However, peer influence is par ticularly important in that task partners are more likely than course instructors to be perceived as ingroup members whose expressed attitudes and views , including those toward a cooperative learning task, are likely to hold more influential power over ind ividuals than those of outgroup members (e.g. Mackie & Cooper, 1984; Turner, 1991). And as Walther and Bunz (2005) suggest, conformity to behavioral rules here , adherence to certain asyn chronous - cooperation etiquette may have special potency in an envi ronment where virtual partners know little else about one another besides their task - related online behavior. As mentioned in the context of sociocognitive conflict regulation above, in the asynchronous online context in particular , individual perception s may be potentially more detrimental than competitive perceptions, as they were significantly and negatively correlated with all motivational and achievement constructs except for content knowledge. Exploratory mediation analyses suggested that reduced ep istemic conflict regulation contributes to these negative associations. Especially to the extent that having to wait for a partner to respond in the asynchronous setting is experienced as a cost, individualistic perceptions are likely to enter the picture. Calls to guard against problems created by a resulting disinterest in engaging productively with peers via course design or participant training, although nothing new (e.g. 119 Burge, 1994; Kreijns, Kirschner, & Jochems, 2003), are highly relevant in light of the present findings. Less clear but equally important are implications for descriptive group norms in this context. In contrast to previous intervention studies that indicated increased intention to engage in socially and individually beneficial target behaviors, normative information was mostly associated with reductions in desirable motivation variables and not related to behavioral engagement variables as predicted. Practitioners considering the use of such norms should tread lightly and carefully con sider the context, the balance of extrinsic and intrinsic motivational factors, the informational value of norms, and the potential inte ractions between these elements in order to avoid the risk for directly - stated completion rate information to be perceiv ed as controlling or, however subtly, threatening as a social upward comparison as described by the BFLP effect. The findings of the present study suggest alternative ways of employing normative information that may be more effective and less susceptible t o unintended side effects at the same time. Three possible ways are suggested here, but many related options are worth exploring. First, as described as one of the delimitations of this study, providing better observability nt does not need to be limited to the mere meeting of completion of task benchmarks. Richer updates regarding that include are likely to create a high performing course norm as well while being perceived as more informative and less controlling (see existing interventions geared at boosting engagement through real - time updates on the rest of the class or team have done , e.g. Ganoe et al., 2003). Again, m ore substantive 120 highlight the nature and the q uality of enga gement with the task rather than its mere occurrence. As mentioned previously, updates regarding completion rates can be seen as double - edged swords at risk of generating extrinsic pressures, a sense of falling behind and a distraction from potentially mor e intrinsically motivating content - based example setting. As technical feasibility of the latter can be expected to grow, it will be interesting to implement and examine effects of such richer behavioral updates. Secondly, the dose of normative informatio n is a point that requires careful consideration. need for repeated exposure to behavioral prevalence information to process and internalize descriptive norms, a activity used within this study, such repeated updates by necessity led to a very dens e timeline that would be less extreme in more naturalistic contexts, where e.g. a weekly update about other without risking a potential overdose. Either way, instr uctors should be sensitive to the need to balance consistency and repetition for norm creation on one hand with the risk for perceived heavy - handedness on the other. rests on the c orrelational finding that perceived descriptive norms were related to positive outcomes overall, and the surprise observation that the cooperation - promoting partner messages effectively boosted perceived descriptive norms. Although follow up work is needed to confirm any potentially productive effects of such perceptions given the purely correlational results in this study, it 121 remains likely that fostering high descriptive norms regarding course engagement remains a worthwhile goal . This claim is in line wi th Cialdini and colleagues thoroughly backed findings regarding that descriptive social norms have on subsequent social findings with venerable roots reaching back to work by, among others, Fest inger ( 1954 ) and Le Bon (1895; 1960 ), and Milgram, Bickman, and Berkowitz ( 1969). However, given detrimental comparison effects along the lines of the BFLP effect considered in this study, as well as the risk for directly perceivable efforts to control beh avior to be perceived as just that controlling, and to produce negative feelings regarding (Brehm & Brehm, 1981; Cialdini, 2007, p. 264) on the other, it seems safe for practitioners to bet e, of which the cooperation - promoting messages may be one example. 122 APPENDICES 123 Appendi x A : Constructive Controversy site description with instructions Day 1: Initial statements Figure 1 1 : Screen shot of the controversy website, Day 1 Day 2: Open Debate # 1 Today you will engage in an open debate about the topic with your opposing part ner. The goal here is to freely exchange information and ideas, arguing forcefully and persuasively for your their thinking, and be sure to ask for support for th eir assertions. You also want to present counter arguments and defend your claims by rebutting the attacks on your position. Remember under the most critical of anal ysis. For this reason, you want to avoid any temptation you might 124 feel to drop your assigned position. Have fun with the debate while recognizing that only critical thinking about both sides of the argument will strengthen your final consensus essay. As yo u engage in an open debate, remember also the rules for constructive controversy. Specifically, you want to be critical of ideas, not people. Remember also that your goal is not to trongest argument. Thus, you want to understand both sides of the issue and only change your mind when the evidence clearly indicates that you should do so. Given the asynchronous nature of this activity, your open debate involves two steps over two cons Statement (see above). In the second part of the open debate (tomorrow), you will send a second Your text statement should be about 100 to 125 words long. Directions for Submitting Rebuttal Statement 1: 1) 2) Type your response below. Day 3: Open Debate #2 The next step in the Open Deb ate is sending a second rebuttal statement in response to the evidence and reasoning, present counter arguments, and defend your assigned position. Your text stat ement should be about 100 to 125 words long. Directions for Submitting Rebuttal Statement 2: 1) 2) Type your response in the appropriate box below. Day 4: Reverse Perspectives In Step 4 each of yo u will reverse positions. This means that you will present your opposing notes as you do this, again being as sincere and forceful as you can. Your goal is t o show that or elaborate on those points that you feel can be strengthened. 125 Your text statement should be about 100 to 125 words long. Directions for Submitting the Reverse Perspectives Statement: 1) 1 and 2. 3) Type your response in the appropriate box below. Day 5: Synthesis In Step 5, the final step, you and your opposing partner will now drop all advocacy of your assigned positions, working together to find a mutually agreeable synthesis of the best and most valid argumen ts on both sides of the debate. Working together you will write a brief, 125 - to 250 - word statement that (a) summarizes the best and most valid arguments on both sides of the debate and (b) synthesizes these arguments into a consensus position that is more rational than the two originally assigned positions. Each of you will be asked t o sign this statement, indicating that you agree with it and con sider it ready to be evaluated. draft of the Consensus Statement. Second the student assigne draft and provid e a final Consensus Statement. Directions for Drafting Consensus Statement Please follow the directions clo sely and make your statement around 125 - 250 words. 126 Appendix B : Constructive controversy materials Can organizational socialization practices as found at places like Google be too intrusive into (Sample arguments to get you started feel free to add your own ideas to the discussion. ) Perks and more: Google has been ranked at the top of Fortune 's list for the past two years. This is not simply a result of extravagant employee benefits such as on site bowling alleys, Lego rooms, billiard tables and gym member ships to free haircuts, Wi - Fi - outfitted shuttle rides to work and pet - friendly policies. Instead, these perks are combined with an authentic reverence for employees and the belief that people truly thrive in their jobs and remain loyal to them when they feel fully supported and authentically valued (Crowley, 2013). Morale drop prevention: Right after starting a new job, employee motivation tends to be high, but according to some studies, in as little as six months the level of morale can drop dramatically if employees do not feel valued. Google and organizations like it use generous perks and a strong company culture to prevent such drop offs during the accommodation phase of organizational Continued success: The sheer success of the company seems to indicate that the carefully engineered Human Resource practices are producing a happy and profitable workplace: Google´s stock appreciated by over 650% since the fir Dow Jones average that is up by just 44% in the same timeframe. The speed of innovation is stunning and may indicate well - supported minds at work: Numerous other products most of us take for granted such as Google Maps, Gmail, Translator, Google Earth, and Android were created or refined since 1998 when Larry Page and Sergey Brin founded the firm (Crowley, 2013). The famous 20% rule: employee regardless of job title or pay level has the opportunity to devote up to 20% of their workweek to a project of their choice. Typically, Googlers choose to help out on some other company venture, but the pursuit is ultimately up to each employee (Crowl ey, 2013). Inspiring organizational goals: Google has uncommonly aspirational ambitions (its mission employees find especially motivating and inspiring (Crowley, 2013) . Employee voice: The firm solicits employee feedback on everything from how they prefer to be compensated, to the design of new bicycles used throughout the expansive headquarters campus (Crowley, 2013). Crowley, M. C. (2013). Not a happy accident: How G oogle deliberately designs workplace satisfaction. Fast Company . Retrieved from http://www.fastcompany.com/3007268/where - are - they - now/not - happy - accident - how - google - deliberately - designs - workplace - satisfaction Perk Place: The benefits offered by Google and o Knowledge@Wharton podcast (2007, March 21). Retrieved from http://knowledge.wharton.upenn.edu/article/perk - place - the - benefits - offered - by - google - and - others - may - be - grand - but - theyre - all - business/ 127 Can organization al socialization practices as found at places like Google be too intrusive into ( S ample arguments to get you started feel free to add your own ideas to the discussion. ) Play some, work a lot: Descriptions of workplac e perks, culture and socialization practices of High Tech firms such as Google tend to draw colorful pictures of a playground - like atmosphere selective hiring processes (Bulygo, 2013). Life in the Googleplex = Life in a bubble? otal institutio once you check in, you do not leave, usually referring to prisons or mental hospitals. Is a workplace that also becomes a place to eat, socialize, do laundry, even sleep, a mild version of a total institution, disconnecting employees Integrators versus Segmentors: Blurred lines between work and home may appeal to people integrators for people who like to main tain distinct walls between work and home , who are losing your job rk Groupthink - Risks: N umerous firms including Kodak, Xerox, and HP, were extremely s and cause them to act on emotions rather than facts (Sullivan, 2012). Workforce inequalities: Companies like Google rely on large portions of highly skilled temporary labor - up to 50% of workers at peak times. These workers are excluded from the famous generous benefits. Hopes of eventually converting to full - time employee status are rarely realized. This two - class system is not often mentioned in discussions of human resource philosophy at places like Google. Data - driven optimization gone wild? Human re Google. Google uses all sorts of data, from surveys to psychological profiles, to gauge cafeteria tables and the l ength of the optimal lunch line is carefully calculated and engineered (3 - 4 minutes, short enough to not waste time yet long enough to talk to others). (Bulygo, 2013, BBC V ideo (May, 2012 ), retrieved from https://www.youtube.com/watch?v=PA54HWLZ2e4 http://blog.kissmetrics.com/googles - culture - of - success/ ong With Employee Engagement? The Top 20 Potential Problems. Retrieved from http://www.ere.net/2012/02/23/what%E2%80%99s - wrong - with - employee - engagement - the - top - 20 - potential - problems/ Kunda, G. (2009). Engineering culture: Control and commitment in a high - t ech corporation. Philadelphia, PA: Temple University Press. 128 Appendix C : Final survey items Exploratory item for manipulation check/baseline norm estimates: At the end of each day, about how many pairs would you guess were done with that day's step? (0 - 2 0%; 20 - 40%; 40 - 60%; 60 - 80%; 80 - 100%) Perceived descriptive course norms (adapted from Lapinski, Rimal, DeVries, and Lee, 2007): 1. Other students were regularly engaged in the controversy activity. 2. Other pairs were moving through the controversy steps accor ding to schedule. 3. Most other pairs stayed on track with the activity. 4. Most students were doing their part to complete the controversy activity. Interest/Enjoyment (IMI, Ryan, 1982) 1. I enjoyed doing the activity very much. 2. The activity was fun to do. 3. This activity was exciting to me. 4. I liked this activity. Perceived Competence (IMI) 1. I think I was pretty good at this activity. 2. I think I did pretty well at this activity, compared to other students. 3. I am satisfied with my performance in this activity. 4. I was p retty skilled at this activity. Interpersonal Attraction (CLS, Johnson, Johnson, Buckman, & Richards, 1985) 1. I liked the peer I was working with. 2. The peer I was working with liked working with me. 3. During the activity I felt warmly towards the peer I was wo rking with. Utility Value (adapted from Conley, 2012) 1. Being good at activities like this could be important for my future (like when Iget a job or go to graduate school). 2. This activity was valuable because it could help me in the future. 3. This activity co uld be useful for me later in life. 4. This activity is practical for me to know. 5. This activity will help me in my daily life outside of school. Cost Value (adapted from Conley, 2012) 1. Considering what I want to do with my life, this activity was just NOT wor th the effort. 2. Success on this activity required that I give up other activities I enjoy. 3. When I think about the hard work needed to be successful in this activity, I am not sure that it was worth it in the end. 4. This activity required more effort than I wa s willing to put in. 5. I had to give up a lot to do well on this activity. 129 Work Avoidance (Harackiewicz, Durik, Barron, Linnenbrink - Garcia, & Tauer, 2008) 1. I tried to do as little work as possible in this activity. 2. One of my goals was to get through this acti vity by doing the least amount of work possible. 3. One of my goals was to not work hard on this activity. Perceptions of Cooperation (SIS, Johnson & Norem - Hebeisen, 1979) 1. I liked sharing my ideas and materials with the other student in this activity. 2. I lear ned important things from the other student in this activity. 3. I liked to help the other student learn in this activity. 4. I like working with the other students. 5. I believe working with onather student is better than working alone. Perceptions of Competition (SIS) 1. 2. I wanted to do better than the other student in this activity. 3. I wanted to be the best student in this activity. 4. I believe competing with the other student is a good way to work in this a ctivity. Perceptions of Individualism (SIS) 1. I wish we could have completed this activity on our own. 2. I would have rather worked on this activity alone than with other students. 3. I wish we completed this activity on our own. Relational Conflict Regulation (CRS, Darnon et al., 2006) 1. try to show you were right? 2. try to resist by maintaining your initial position? 3. try to show your partner was wrong? Epistemic Conflict Regulation (CRS) When disagreements occu 1. try to think about the arguments again in order to understand them better? 2. try to examine the conditions under which each point of view could help you understand? 3. try to think of a solution that could integrate both points of view? Content Knowledge 1. Perks prevent morale drop - TRUE (1) is correct 2. 20% Rule allows - Option d (4) is correct 3. - FALSE (2) is correct 4. Common criticism - Option c (3) is correct Group Identity measure (Rimal & Real, 2003) 1. I am similar to most students in my classes intellectually 2. I am similar to most students in my classes in my thinking. 3. I am similar to most students in my classes in my values. 130 4. I am similar to most students in my classes in my behaviors. Ego Involvement (adapted from Rimal, Real, & Morrison, 2004) a) Good Student identity: 1. I would describe myself as an excellent/good/average/poor student. 2. My friends would describe me as an excellent/good/average/poor student. 3. Most of my instructors would descr ibe me as an excellent/good/average/poor student. b) Good Communicator identity: 1. I would describe myself as an excellent/good/average/poor communicator. 2. My friends would describe me as an excellent/good/average/poor communicator. 3. Most of my instructors wou ld describe me as an excellent/good/average/poor communicator. Website Enjoyment adapted from Venkatesh) 1. I liked working with the Controversy website. 2. Working with the Controversy Website was fun. 3. The Controversy website made work more interesting. Ave rage time per step On average, how much time did you spend working on each step of the constructive controversy activity? Time preference All things being equal, would you have preferred more/same/less time to work on the Constructive Controversy activity ? Website Norms percentage How often did the website show you the percentage of participants already done with the step for that day? Additional screening/exit questions Is there anything you either particularly liked or disliked about this activity? In your own words, what is your best guess on what this research is trying to find out? Demographics Age Ethnicity Gender GPA Major Year in school 131 A ppendix D : Participant Flow Table 9 : Participant f low . D ifferent lines within a row represent the thr ee diff erent weeks of data collection. n per group 1) SN+CPM 2) CPM_only 3) SN _only 4) Neither/ control All conditions Assigned to group: 18 21 22 61 18 21 22 61 17 21 22 60 17 20 21 58 70 83 87 240 Received exp. Manipulations: 17 20 21 58 16 19 20 55 16 21 22 59 14 21 20 55 63 81 83 227 Did not receive Never started activity (1) Never started activity (1) Never started activity (1) 3 Never started the activity (1); Did all steps on Day 1 (1) Never started activity (2) Did all steps on day 1 (1); Never started activity (1) 6 Never started activity (1) 1 Never started the activity (1); Did all steps on Day 1 (1) Never started activity (1) 3 Never started activity (10); Did all steps on Day 1 (3) 13 Participation flow: n / group 1) SN+CPM 2) Coop 3) SN _only 4) Neither/ control All conditions Step 1 17 20 21 58 16 19 20 55 16 21 22 59 14 20 20 54 63 80 83 226 Total enrollment of students across all courses in which study was offered: 909 students, n=280 emailed with request for sign - up instructions Assigned to condit ions by creating account on activity website (n=240) 132 Table Step 2 17 20 21 58 16 19 20 55 16 20 22 58 13 20 20 53 6 2 79 83 224 Step 3 17 19 21 57 15 19 20 54 16 19 22 57 13 20 20 53 61 77 83 221 Step 4 16 19 21 56 15 19 20 54 16 19 20 55 13 19 20 52 60 76 81 217 Step 5 15 19 20 54 14 19 19 52 15 18 18 51 12 19 20 51 56 75 77 208 Survey 17 19+1 18 55 14 17 19 5 0 16 +2 18+1 21 58 15 19+1 21 56 64 76 79 219 Analyzed (n=193) : n per group 1) SN+CPM 2) CPM_only 3) SN _only 4) Neither/ control All conditions Analyzed 16 18 16 50 13 16 15 44 16 16 19 51 13 16 19 48 58 66 69 193 Excluded from analysis (due t o reasons below) 1 2 2 5 1 1 4 6 2 3 2 7 2 4 2 8 6 10 10 26 1 - suspicion re: partner 1, 2 2 1, 2 2, 1 11 2 - close guess re: study purpose 1 1 3 - had not participated or steps upon first logon 1 1 1, 1 4 4 - repeat survey taker 1 2, 1 1 5 5 - survey duration <3minutes 2 1, 1 1, 1 6 133 A ppendix E : Actual completion percentages Table 10: Actual step completions and completion percentages by day and time for sample week Note: Completion percentages for 1pm are highlighted in red because they allow for a direct comparison with the provided inflated percentage updates that ranged from 70 - 85%. Throughout the course of the day in question, this inflation becomes less and less pronounced. The normative range provided in the treatment is equiv alent to actual completion percentages for that 134 A ppendix F : Message generation schedule Figure 1 2 : Message generation schedule b alancing substantive content across conditions 135 REFERENCES 136 REFERENCES Ajzen, I. (1991). The theory of planned behavior. Organizational behavior and human decision processes, 50 (2), 179 - 211. Ajzen, I., & Fishbein, M. (1980 ). Underst anding attitudes and predicting social behavior . Englewood Cliffs, NJ: Prentice Hall. Allen, I., & Seaman, J. (2015). Grade level: Tracking online education in the United States. Babson Survey Research Group . Retrieved from: http://onlinelearningconsortiu m.org/read/survey - reports - 2014. Anderson, L.W., Krathwohl, D.R., Airasian, P.W., Cruikshank, K.A., Mayer, R.E., Pintrich, P.R., Raths, J., Wittrock, M.C. (2001). A taxonomy for learning, teaching, and assessing: a revision of bloom's taxonomy of educational objectives . New York: Allyn & Bacon. Angeli, C., Bonk, C., & Hara, N. (1998). Content analysis of online discussion in an app lied educational psychology course . Retrieved from http://crlt.indiana.edu/publications/crlt98 - 2.pdf Aragon, S. R. (2010). Creating social presence in online environments. New Directions for Adult and Continuing Education (San Francisco: Jossey Bass, 2003) , 57 - 68. Baer, J. S., Stacy, A., & Larimer, M. (1991). Biases in the perception of drinking norms among college students. Journal of Studies on Alcohol, 52 , 580 586. Baer, J.S. & Carney, M.M. (1993). Biases in perception of the consequences of alcohol use among college students. Journal of Studies on Alcohol 54 , 54 - 60. Baumeister, R. F., Bratslavsky, E., Finkenauer, C., & Vohs, K. D. (2001). Bad is stronger than good. Review of General Psychology, 5 , 323 370. Berkowitz, A.D. (2004). An overview of the socia l norms approach. In L. Lederman & L. Stewart (Eds.), Changing the culture of college drinking (pp. 193 214). Cresskill, NJ: Hampton Press. Bloom, B.S., Engelhart, M.D., Furst, E.J., Hill, W.H., & Krathwohl, D.R. (1956). Taxonomy of Educational Objectives, Handbook I : The Cognitive Domain . New York: David McKay Co Inc. Bobek, D. D., Roberts, R. W., & Sweeney, J. T. (2007). The s ocial norms of tax compliance: Evidence from Australia, Singapore, and the United States. Journal of Business Ethics, 74 (1), 49 - 64. 137 Brehm, J. W., (1966). A theory of psychological reactance. New York: Academic Press. Brehm, S. S., & Brehm, J. W. (1981). Ps ychological reactance: A theory of freedom and control . New York: Academic Press. Journal of Personality and Social Psych ology 49(50), 543. Bunz, U., & Campbell, S. W. (2004). Politeness accommodation in electronic email. Communication Research Reports, 21 , 11 - 25. labeling depends upon t he type of activity. International Journal of Computer - Supported Collaborative Learning, 5 (1), 103 - 116. Bures, E. M. , Abrami, P. C. , & Schmid, R. F. (2010). Fostering Quality Online Dialogue: Does Labeling Help? Journal of Interactive Learning Research, 21 , 187 - 213. Journal of Distance Education, 9 (1), 19 - 43. Butera, F., & Darnon, C. (2010). Socio - cognitive conflict and learning: Past and present. ICLS '10 Proceedings o f the 9th International Conference of the Learning Sciences, 2, 212 - 213. Byrne, S., & Hart, P. S. (2009). preliminary theoretical framework. In C. Beck (Ed.), Communication Yearbook 33, Mahwah, NJ: Lawr ence Erlbaum Associates. evaluations, knowledge content, social cues and personal information on later messages. Computers & Education, 50 , 678 - 692. Christensen, P. N ., Rothgerber, H., Wood, W., & Matz, D. C. (2004). Social norms and identity relevance: A motivational approach to normative behavior. Personality and Social Psychology Bulletin , 30, 1295 1309. Cialdini, R. B. (1988). Influence: Science and practice (2nd e d). Glenview, IL: Pearson . Cialdini, R. B. (2003). Crafting normative messages to protect the environment. Current Directions in Psychological Science, 12 , 105 - 109. Cialdini, R. B. (2007). Descriptive social norms as underappreciated sources of social cont rol. Psychometrika, 72 (2), 263 - 268. 138 Cialdini, R. B., & Goldstein, N. J. (2004). Social Influence: Compliance and Conformity. Annual Review of Psychology, 55 (1), 591 - 621. Cialdini, R. B., Kallgren, C. A., & Reno, R. R. (1991). A focus theory of normative conduct: A theoretical refinement and reevaluation of the role of norms in human behavior. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 24, pp. 201 - 234). San Diego, CA: Academic Press. Cialdini, R. B., Martin, S. J., & Goldstein, N. J. (2015). Small behavioral science - informed changes can produce large policiy - relevant effects. In C. R. Fox & S. B. Sitkin (eds.), Challenging assumptions about behavioral policy (pp. 21 - 29). Durham, NC: Behavioral Science & Policy Association. Ciald ini, R. B., Reno, R. R., & Kallgren, C. A. (1990). A focus theory of normative conduct: Recycling the concept of norms to reduce littering in public places. Journal of Personality and Social Psychology, 58 , 1015 - 1026. Cialdini, R. B., & Trost, M. R. (1998 ). Social influence: Social norms, conformity, and compliance. In D. Gilbert, S. Fiske & G. Lindzey (Eds.), The handbook of social psychology (4 ed., Vol. 2, pp. 151 - 192). New York: McGraw - Hill. Clinefelter, D. L., & Aslanian, C. B. (2015). Online college students 2015: Comprehensive data on demands and preferences. Louisv ille, KY: The Learning House . Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2013). Applied multiple regression/correlation analysis for the behavioral sciences . London: Routledge. Co nley, A. M. (2012). Patterns of motivation beliefs: Combining achievement goal and expectancy - value perspectives. Journal of Educational Psychology, 104 (1), 32 - 47. Conner, M., & Armitage, C. J. (1998). Extending the theory of planned behavior: A review an d avenues for future research. Journal of Applied Social Psychology , 28, 1429 1464. Corriero, E. F., Tong, S. T., & Sopory, P. (2015, January). Behaviors, perceptions, responsiveness, and presence: The dyadic model of mediated communication. System Scienc es (HICSS), 2015 48th Ha waii International Conference . IEEE. Cramton, C. D. (2001). The mutual knowledge problem and its consequences for dispersed collaboration. Organization Science, 12, 346 - 371. Darnon, C., Buchs, C., & Butera, F. (2002). Epistemic and relational conflicts in sharing identical versus complementary information during cooperative learning. Swiss Journal of Psychology, 61 (3), 139 - 151. 139 Darnon, C., Doll, S., Butera, F. (2007). Dealing with a disagreeing partner: Relational and epistemic conf lict elaboration. European Journal of Psychology of Education, 22 , 227 - 242. Darnon, C., Muller, D., Schrager, S. M., Pannuzzo, N., & Butera, F. (2006). Mastery and performance goals predict epistemic and relational conflict regulation. Journal of Educatio nal Psychology, 98 , 766 776. Davis, M., & Rouzie, A. (2002). Cooperation vs. deliberation: Computer mediated conferencing and the problem of argument in int ernational distance education. International Review of Research in Open and Distance Learning, 3 (1). De Wever, B., Schellens, T., Valcke, M., & Van Keer, H. (2005). Content analysis schemes to analyze transcripts of online asynchronous discussion groups: a review. Computers and Education, 46 , 6 28. Deci , E. L., & Ryan, R. (2000). The what and why of goa l pursuits: Human needs and the self - determination of behavior. Psychological Inquiry, 11 , 227 268. Deutsch, M. (1949). A theory of cooperation and competition. Human Relations, 2 , 129 - 152. Deutsch, M. (1962). Cooperation and trust: Some theoretical notes . In M.R. Jones (Ed.). Nebraska symposium on motivation (pp.275 - 319). Lincoln: University of Nebraska Press. Deutsch, M. (1973). The resolution of conflict: Constructive and destructive processes. New Haven, CT: Yale University Press. Deutsch, M. (1985). Distributive justice: A social psychological perspective . New Haven, CT: Yale University Press. Deutsch, M., Coleman, P. T., & Marcus, E. C. (2011 ). The handbook of conflict resolution: Theory and practice . Hoboken, NJ: John Wiley & Sons. Doise, W., & Mug ny, G. (1984). The social development of the intellect . Oxford, UK: Pergamon Press. Duncan, S. D. Jr., & Fiske, D. W. (1977). Face - to - face interaction: Research, methods, and theory . Hillsdale, NJ: Erlbaum Eccles, J. S., & Wigfield, A. (2002). Motivational beliefs, values, and goals. Annual Review of Psychology, 53 , 109 132. 140 Eriksson, K., & Strimling, P. Injunctive versus functional inferences from descriptive norms : C omment on Gelfand and Harrington. Journal of Cross - Cultural Psychology . Prepublished Sempt ember 9, 2015 , DOI: 10.1177/0022022115605387 Fahy, P. (2002a). Epistolary and expository interaction patterns in a computer conference transcript. Journal of Distance Education , 17(1), 20 - 35. Fahy, P. (2002b). Use of linguistic qualifiers and intensifiers in computer conference. The American Journal of Distance Education , 16(1), 5 - 22. Fahy, P. (2003). Indicators of support in online interaction. International Review of Research in Open and Distance Learning , 4(1). Retrieved from: h ttp://www.irrodl.org/conte nt/v4.1/fahy.html. Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7 , 117 140. Fishbein, M., & Ajzen, I. (1975 ). Belief, attitude, intention and behavior: An introduction to theory and research. Boston, MA: Addison - Wesley. F orsyth, D. R. (2000). Social comparison and influence in groups. In J. Suls & L. Wheeler (eds.), Handbook of Social Comparison (pp. 81 - 103). New York: Springer US. en gagement and performance. Journal of Educational Psychology, 95, 148 162. Ganoe, C. H., Somervell, J. P., Neale, D. C., Isenhour, J.P. L., Carroll, J. M., Rosson, M. B., & McCrickard, D. S. (2003). Classroom BRIDGE: Using collaborative public and desktop timelines to support activity awareness. CHI Letters , 5 (2), 21 - 30. Garrison, D. R., T. Anderson, & W. Archer (2000). Critical inquiry in a text - based environment: Computer conferencing in higher education. The Internet and Higher Education, 2 (2 3): 1 19. Garrison, D. R., Anderson, T., & Archer, W. (2001). Critical thinking, cognitive presence, and computer conferencing in distance education. American Journal of Distance Education, 15 , 7 23. Gerber, A. S., & Rogers, T. (2009). Descriptive social norms and m otivation to vote: Everybody's voting and so should you. The Journal of Politics, 71 (01), 178 - 191. Gilbert, L., & Moore, D. R. (1998). Building interactivity into web courses: tools for social and instructional interaction. Educational Technology , 38(3), 2 9 35. Gilroy, K. (2001). Collaborative E - learning: the right approach. ArsDigita Systems Journal. Retrieved from: http: //www.arsdigita.com/asj/elearning/ 141 Goldstein, N. J., Cialdini, R. B., & Griskevicius, V. (2008). A room with a viewpoint: Using social n orms to motivate environmental conservation in hotels. Journal of Consumer Research, 35 (3), 472 - 482. Golub, M., & Buchs, C. (2014). Preparing pupils to cooperate during coope rative controversy in grade six: A way to increase positive interactions and learn ing? European Journal of Psychology of Education, 29 , 453 - 466. Graham, C., Cagiltay, K., Lim, B. R., Craner, J., & Duffy, T. M. (2001). Seven principles of effective teaching: A practical lens for evaluating online courses. The Technology Source, 30 (5), 50 . Gunawardena, C. N. (1995). Social presence theory and implications for interaction and collaborative learning in computer conferences. International Journal of Educational Telecommunications , 1(2/3), 147 166. Guanawardena, C. N., Lowe, X., Constance, A., & Anderson, T. (1997). Analysis of a global debate and the development of an interaction analysis model for examining social construction of knowledge in computer conferencing. Journal of educational computing research , 17 , 397 431. Gunawardena, C., & Zit tle, F. (1997). Social presence as a predictor of satisfaction within a computer mediated conferencing environment. American Journal of Distance Education, 11 (3), 8 - 26. Guzdial, M., & Turns, J. (2000). Effective discussion through a computer - mediated an ch ored forum. The Journal of the Learning Sciences, 9 , 437 - 469. Hallowell, E. M. (1999). The human moment at work. Harvard Business Review, 77 , 58 - 66. Hara, N. M., Bonk, C. J. M., & Angeli, C. M. (2000). Content analysis of online discussion in an applied ed ucational psychology course. Instructional Science, 28 , 115 - 152. Harackiewicz, J. M., Durik, A. M., Barron, K. E., Linnenbrink - Garcia, L., & Tauer, J. M. (2008). The role of achievement goals in the development of interest: Reciprocal relations between ach ievement goals, interes t, and performance. Journal of Educational P sychology, 100 , 105 - 130. Harasim, L. M. (1993). Networks: Computers and international communication . Cambridge, MA: MIT Press. Hattrup, K., & Jackson, S. E. (1996). Learning about individua l differences by taking situations seriously . In K. R. Murphy (Ed.), Individual differences and behavior in organizations (pp. 507 - 547). San Francisco: Jossey - Bass. 142 Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analys is: A regression - based approach . New York: Guilford Press. Hewitt, J. (2005). Toward an understanding of how threads die in asynchronous com puter conferences. The Journal of the Learning Sciences, 14 , 567 - 589. Hoadley, C. M. (2000). Teaching science throu gh online, peer discussions: SpeakEasy in the Knowledge Integration Environment. International Journal of Science Education, 22 , 839 - 857. Hsu, M. H., Ju, T. L., Yen, C. H., & Chang, C. M. (2007). Knowledge sharing behavior in virtual communities: The relat ionship between trust, self - efficacy, and outcome expectations. International Journal of Human - Computer Studies, 65 , 153 - 169. Huguet, P., Dumas, F., Marsh, H., Régner, I., Wheeler, L., Suls, J., Seaton, & Nezlek, J. (2009). Clarifying the role of social c omparison in the big - fish little - pond effect (BFLPE): An integrative study. Journal of personality and social psychology, 97 (1), 156. Jaccard, J., & Turrisi, R. (Eds.). (2003). Interaction effects in multiple regression (Vol. 72). Newbury Park, CA: Sage. J accard, J., Wan, C. K., & Turrisi, R. (1990). The detection and interpretation of interaction effects between continuous variables in multiple regression. Multivariate Behavioral Research , 25 , 467 - 478. Jaggars, S. S. (2014). Choosing between online and fac e - to - face courses: Community college student voices. American Journal of Distance Education , 28 (1), 27 - 38. Jeong, A. C. (2006). The effects of conversational language on group interaction and group performance in computer - supported collaborative argumentat ion. Instructional Science , 34 , 367 - 397. Jeong, A., & Joung, S. (2007). Scaffolding collaborative argumentation in asynchronous discussions with message constraints and message labels. Computers & Education , 48 , 427 - 445. Johnson, D. W. (2003). Social int erdependence: interrelationships among theory, research, and practice. American Psychologist , 58 , 9 34 - 945. Johnson, B. T., & Eagly, A. H. (1989). The effects of involvement on persuasion: A meta - analysis. Psychological Bulletin , 106, 290 314. Johnson, D. W, & Johnson, R. T. (2009). Energizing learning: The instructional power of conflict. Educational Researcher, 38(1), 37. 143 Johnson, D. W. & Johnson, R. T. (2010). Cooperative learning and conflict resolution: Essential 21st century skills. In J. Bellanca & R . Brandt (eds.), 21st Century skills: Rethinking how students learn (pp. 201 - 219). Bloomington, IN: Solution Tree Press. Johnson, D. W., & Johnson, R. T. (1983). Social interdependence and perceived academic and personal support in the classroom. The Journ al of Social Psychology , 120(1), 77 - 82. Johnson, D. W., & Johnson, R. T. (1994). Learning together and alone: Cooperative, competitive, and individualistic learning (4th ed.). MA: Allyn and Bacon. Johnson, D. W., & Johnson, R. T. (2005). New developments i n social interdependence theory. Psychology Monographs, 131 , 285 - 358. Johnson, D. W., & Johnson, R. T. (2007). Creative controversy: Intellectual challenge in the classroom (4th ed.). Edina, MN: Interaction Book Company. Johnson, D. W., & Johnson, R.T. (19 89). Cooperation and competition: Theory and research. Edina, MN: Interaction Book Company. Johnson, D. W., Johnson, R. T., Buckman, L. A., & Richards, P. S. (1985). The effect of prolonged implementation of cooperative learning on social support within th e classroom. The Journal of Psychology, 119 , 405 411. Johnson, D. W., Johnson, R. T., & Holubec, E. J. (1984). Cooperation in the classroom . Edina, MN: Interaction Book Co. Johnson, D. W., Johnson, R. T., & Smith, K. A. (2014). Co operative learning: Impro ving university instruction by basing practice on validated theory. Journal on Excellence in College Teaching , 25(3&4), 85 - 118. Johnson, D. W., Johnson, R. T., & Tjosvold, D. (2000). Constructive Controversy: The Value of Intellectual Opposition. M. Deutsc h and P. T. Coleman, (eds.), The handbook of conflict resolution: Theory and practice (pp. 65 - 85). San Francisco: Jossey - Bass. Johnson, D. W., Johnson, R., & Smith, K. (2006). Active learning: Cooperation in the university classroom (3rd ed.). Edina, MN: Interaction. Johnson, D. W., & Norem - Hebeisen, A. A. (1979). A measure of cooperative, competitive, and individualistic attitudes. The Journal of Social Psychology, 109 , 253 - 261. Joo, Y. J., Bong, M., and Choi, H. J. (2000). Self - efficacy for self - regulat ed learning, academic self - efficacy, and Internet self - efficacy in web - based instruction. Educational Technology Research & Development 48(2): 5 17. 144 Kanuka, H., & Anderson, T. (1998). On - line social interchange, discord and knowledge construction. Journal of Distance Education, 13 (1), 57 - 74. Kelley, H. H. (1952). Two functions of reference groups. In G. E. Swanson, T. M. Newcomb, & E. L. Hartley (Eds.), Readings in social psychology (2nd ed., pp. 410 - 414). New York: Holt. Kenny, D. A., Kashy, D. A., & Cook, W. L. (2006). Dyadic Data Analysis. New York: The Guilford Press. Kim, B. (2009). The effects of prompts - based argumentation scaffolds on peer - led interactive argumentation . Doctoral dissertation, University of Missouri Columbia. Kirschner, P. A., Beers, P. J., Boshuizen, H. P. A., & Gijselaers, W. H. (2008). Coercing shared knowledge in collaborative learning environments. Computers in Human Behavior, 24 , 403 420. Klein, A. G., & Moosbrugger, H. (2000). Maximum likelihood estimation of latent interact ion effects with the LMS method. Psychometrika, 65 , 457 474. Köller, O., Baumert, J., & Schnabel, K. (2001). Does interest matter? The relationship between academic interest and achievement in mathematics. Journal for Research in Mathematics Education, 32 , 448 470. Kreijns, K., Kirschner, P. A., & Jochems, W. (2003). Identifying the pitfalls for social interaction in computer - supported collaborative learning environments: A review of the research. Computers in human behavior, 19 , 335 - 353. Lapinski, M. K., & Boster, F. J. (2001). Modeling the ego - defensive function of attitudes. Communication Monographs , 68, 314 324. Lapinski, M. K. & Rimal R. N. (2005) An explication of social norms. Communication Theory. 15 , 127 - 147. Lapinski, M. K., Rimal, R. N., DeVries , R., & Lee, E. L. (2007). The role of group orientation and descriptive norms on water conservation attitudes and behaviors. Health Communication, 22 , 133 - 142. Mäkitalo, K., Weinberger, A., Häkkinen, P., Järvelä, S., & Fischer, F. (2005). Epistemic cooper ation scripts in online learning environments: Fostering learning by reducing uncertainty in discourse? Computers in Human Behavior, 21 , 603 - 622. Marsh, H. W. (1987). The big - fish - little - pond effect on academic self - concept. Journal of E ducational P sycholo gy , 79 , 280 - 293 . 145 Marsh, H. W., & Hau, K. T. (2003). Big - Fish - Little - Pond effect on academic self - concept: A cross - cultural (26 - country) test of the negative effects of academically selective schools. American P sychologist , 58 , 364 - 371 . Marsh, H. W., Hau, K. - T., & Wen, Z. (2004). In search of golden rules: Comment on hypothesis - testing approaches to setting cutoff values for fit indexes and dangers in overgeneralizing findings. Structural Equation Modeling, 11, 320 341. Marsh, H. W. , & Parker, J. W. (1984). Determinants of student self - concept: Is it better to be a relatively large fish in a small pond even if you don't learn to swim as well? Journal of Personality and Social Psychology , 47 , 213 - 225 . Marsh, H. W., Trautwein, U., Lüdt ke, O., & Köller, O. (2008). Social comparison and big - fish - little - pond effects on self - concept and other self - belief constructs: Role of generalized and specific others. Journal of Educational Psychology, 100 , 510 - 523 . Martin, A. J., & Dowson, M. (2009). Interpersonal relationships, motivation, engagement, and achievement: Yields for theory, current issues and educational practice. Review of Educational Research, 79, 327 365 . Martin, S. J., Goldstein, N., & Cialdini, R. (2014). The small BIG: small changes that spark big influence . New York : Grand Central Publishing. McAuley, E., Duncan, T., & Tammen, V. V. (1989). Psychometric properties of the Intrinsic Motivation Inventory in a competitive sport setting: A confirmatory factor analysis. Research Quarterly for Exercise and Sport, 60, 48 - 58. McClelland, G. H., & Judd, C. M. (1993). Statistical difficulties of detecting interactions and moderator effects. Psychological bulletin, 114 , 376. McDonald, J. (1998). Interpersonal group dynamics and development in co mputer conferencing. American Journal of Distance Education, 12 (1), 7 - 25. Meece, J. L., Blumenfeld, P. C., & Hoyle, R. H. (1988). Students' goal orientations and cognitive engagement in classroom activities. Journal of educational psychology, 80 , 51 - 66 . Me non, G., Raghubir, P., & Schwarz, N. (1995). Behavioral frequency judgments: An accessibility - diagnosticity framework. Journal of Consumer Research , 212 - 228. Miller, D.T. & Prentice, D.A. (1996). The construction of social norms and standards in Higgins, F .T. and Kruglanski, A.W. (Eds.). Social Psychology: Handbook of Basic Principles . (pp. 799 - 829). New York: Guilford Press. Mishra, P., & Koehler, M. (2006). Technological pedagogical content knowledge: A framework for teacher knowledge. The Teachers Colleg e Record, 108 , 1017 - 1054. 146 Monteil, J. - M., & Huguet, P. (1999). Social context and cognitive performance: Towards a social psychology of cognition . Hove, United Kingdom: Psychology Press. Moore, J. L., & Marra, R. M. (2005). A comparative analysis of online discussion participation protocols. Journal of Research on Technology in Education , 38 , 191 - 212. Mugny, G., Butera, F., Quiamzade, A., Dragulescu, A., & Tomei, A. (2003). Comparaisons [Social comparison of competences and social influence dynamics in aptitude tasks]. , 104, 469 496. Mugny, G., De Paolis, P., & Carugati, F. (1984). Social regulations in cognitive development . In W. Doise & A. Palmo nari (Eds.), Social interaction in individual development (pp. 127 146). Cambridge, England: Cambridge University Press. Mullen, B., & Hu, L. (1988). Social projection as a function of cognitive mechanisms: Two meta - analytical integration. British Journal. Social Psychology. 27 , 333 - 356 Nam, C. W. (2014). The effects of trust and constructive controversy on student achievement and attitude in online cooperative learning environments. Computers in Human Behavior , 37 , 237 - 248. Nardi, B., & Whittaker, S. (200 2). The place of face to face communication in distributed work. In P. J. Hinds & S. Kiesler (Eds.), Distributed work: New research on working across distance using technology (pp. 83 - 110). Cambridge, MA: MIT Press. Norem Hebeisen, A. A., & Johnson, D. W. (1981). The relationship between cooperative, competitive, and individualistic attitudes and differentiated aspects of self esteem1. Journal of Personality , 49 , 415 - 426. Northrup, P. (2001). A framework for designing inte ractivity in web - based instruction. Educational Technology , 41(2), 31 39. Nowak, K., Watt, J. H., & Walther, J. B. (2009). Computer mediated teamwork and the efficiency framework: Exploring the influence of synchrony and cues on media s atisfaction and outc ome success. Computers in Human Behavior, 25 , 1108 - 1119. Nussbaum, E. M. (2008). Collaborative discourse, argumentation, and learning: Preface and literature review. Contemporary Educational Psychology , 33 , 345 - 359. Nussbaum, E. M., & Schraw, G. (2007). P romoting argument - counterargument integration in students' writing. The Journal of Experimental Education , 76 (1), 59 - 92. Palloff, R. M., & Pratt, K. (2011). The excellent online instructor: Strategies for professional development. Hoboken, NJ: John Wiley & Sons. 147 Pareto , L., Haake, M., Lindstr m , P., Sj den, B., Gulz, A. (2012). A teachable agent game engaging primary school children to learn arithmetic concepts and reasoning. International Journal of Artificial Intelligence in Education, 24 , 251 - 283. Pena - Shaff, J. B., & Nicholls, C. (2004 ). Analyzing student interactions and meaning construction in computer bulletin board discussions. Computers and Education, 42 , 243 265. Reinecke, L., Hofmann, W., & Klein, S. A. (2014). Media - use and self - control: An experience sampling study of media use , goal conflicts, and psychological well - being. Paper presented at the 64th Conference of the International Communication Association (ICA), 22 - 26 May 2014, Seattle, WA, USA Reno, R. R., Cialdini, R. B., & Kallgren, C. A. (1993). The transsituational influence of social norms. Journal of Personality and Social Psychology , 64, 104 - 112. Rimal, R. N. (2008). Modeling the relationship between descriptive norms and behaviors: A tes t and extension of the theory of normative social behavior (TNSB ). Health Communication, 23 , 103 - 116. Rimal, R. N., & Real, K. (2003). Understanding the influence of perceived norms on behaviors. Communication Theory, 13 , 184 - 203. Rimal, R. N., & Real, K. (2005). How behaviors are influenced by perceived norms: A test of the theory of normative social behavior. Communication Research, 32 , 389 414. Rimal, R. N., & Real, K., & Morrison, D. (2004, May). Expanding the theory of normative influences: The role o f ambiguity and identity . Paper presented at the annual conference of the International Communication Association, New Orleans, LA. Rimal, R. N., Real, K., & Morrison, D. (2004, May). Expanding the theory of normative influences: The role of ambiguity and identity. Paper presented at the Health meeting in New Orleans, LA. Rimal, R. N., Real, K., & Morrison, D. (2004, May). Expanding the theory of normative influences: The role of ambiguity and identity. Paper presented at the annual conference of the International Communication Association., New Orleans, LA. Rimal, R . N, & Real , K. (2005). How behaviors are influenced by perceived norms: A test of the theory of normative social b ehavior. Communication Research, 32 , 389 - 414. Rivis, A., & Sheeran, P. (2003). Descriptive norms as an additional predictor in the theory of planned behavior: A meta - analysis. Current Psychology, 22 , 218 - 233. 148 Roe, B. E., & Just, D. R. (2009). Internal and external validity in economics research: Tradeoffs between experiments, field experiments, natural experiments, and field data. American Journal of Agricultural Economics, 91 , 1266 - 1271. Rosen, L. D., Carrier, L. M., & Cheever, N. A. (2013). Facebook and texting made me do it: Media - induced task - switching while studying. Computers in Human Behavior, 29 , 948 - 958. achievement and peer relationships: The effects of cooperativ e, competitive, and individualistic goal structures. Psychological Bulletin, 134, 223 246. Roseth, C. J., Saltarelli, A. J., & Glass, C. R. (2011). Effects of face - to - face and computer - mediated constructive controversy on social interdependence, motivation , and achievement." Journal of Educational Psychology, 103 , 804 - 820. Rourke, L., Anderson, T., Garrison, D. R., & Archer, W. (1999). Assessing social presence in asynchronous text - based computer conferencing. Journal of Distance Education, 14 (2). Rovai, A. P. (2001). Classroom community at a distance: a comparative analysis of two ALN - based university programs. Internet and Higher Education, 4, 105 118. Ruberg, L. F., Moore, D. M., & Taylor, C. D. (1996). Student participation, interaction, and regulation i n a computer - mediated communication environment: a qualitative study. Journal of Educational Computer Research, 14 , 243 268. Ryan, R. M. (1982). Control and information in the intrapersonal sphere: An extension of cognitive evaluation theory. Journal of Pe rsonality and Social Psychology, 43 , 450 461. Ryan, R. M., Mims, V., & Koestner, R. (1983). Relation of reward contingency and interpersonal context to intrinsic motivation: A review and test using cognitive evaluation theory. Journal of Personality and S ocial Psychology , 45 , 736 - 750. Saba, F., & Shearer, R. L. (1994). Verifying key theoretical concepts in a dynamic model of distance education. American Journal of Distance Education, 8 (1), 36 - 59. Saltarelli, A. J., & Roseth, C. J. (2014). Effects of synchr onicity and belongingness on face - to - face and computer - med iated constructive controversy. Journal of Educational Psychology, 106, 946 - 960. Scardamalia, M., & Bereiter, C. (1996). Computer Support for Knowledge - Building Communities. In T. Koschman (Ed.), CS CL: Theory and practice of an emerging paradigm (pp. 249 - 268). Mahwah, NJ: Lawrence Erlbaum Associates. 149 Schot, V. (2011). Social norms to motivate IT use. Netherlands. Schultz, P. W. (1999). Changing behavior with normative feedback interventions: A field experiment on curbside recycling. Basic & Applied Social Psychology, 21 , 25 - 36. Schultz, P. W., Nolan, J. M., Cialdini, R. B., Goldstein, N. J., & Griskevicius, V. (2007). The constructive, destructive, and recons tructive power of social norms. Psychological Science, 18 , 429 - 434. Schwarz, N., Bless, H., Strack, F., Klumpp, G., Rittenauer - Schatka, H., & Simons, A. (1991). Ease of retrieval as information: Another look at the availability heuristic. Journal of Person ality and Social psychology , 61 , 195 - 212 . Sheeran, P., & Orbell, S. (1999). Augmenting the theory of planned behavior: Roles for anticipated regret and descriptive norms. Journal of Applied social Psychology, 29 , 2107 - 2142. Sheldon, K., & Houser - Marko, L. (2001). Self - concordance, goal attainment, and the pursuit of happiness: Can there be an upward spiral? Journal of Personality and Social Psychology, 80, 152 - 165. Sherif, C. W., Kelly, M., Rodgers, H. L., Sarup, G., & Tittler, B. I. (1973). Personal invol vement, social judgment, and action. Journal of Personality and Social Psychology, 27 , 311 328. Shibutani, T. (1955). Reference groups as perspectives. American Journal of Sociology , 60, 562 - 569. Short, J., Williams, E., & Christie, B. (1976). The social p sychology of telecommunications . Toronto, ON: Wiley. Skinner, E. A., & Pitzer, J. R. (2012). Developmental dynamics of student engagement, coping, and everyday resilience . In S. L. Christenson, A. L. Reschly, & C. Wylie (Eds.), Handbook of research on stud ent engagement (pp. 21 - 44). New York: Springer. Slavin, R. E. (1988). Cooperative learning and student achievement. Educational Leadership, 46 , 31 - 33. Sproull, L., & Kiesler, S. (1986). Reducing social context cues: Electronic mail in organizational comm unication. Management Science, 32 , 1492 - 1513. missing from online learning: Interpretations through the Community of Inquiry framework. International Review of Res earch in Open and Distance Learning, 7 (3), 1 - 24. 150 Suls, J. M., & Wheeler, L. (2000). A selective history of classic and neo - social comparison theory. In J. M. Suls & L. Wheeler (Eds.), Handbook of social comparison: Theory and research (pp. 3 22). Dordrecht , Netherlands: Kluwer Academic. Taggar, S., & Ellis, R. (2007). The role of leaders in shaping formal team norms. The Leadership Quarterly, 18 , 105 - 120. Tajfel, H., & Turner, J.C. (1986). The Social identity theory of intergroup behavior. In S. Worchel & W .G. Austin (Eds.), Psychology of intergroup relations . (pp. 7 - 24). Chicago: Nelson - Hall. Tajfel, H., Billing, M., Bundy, R., & Flament, C. (19 71). Social Categorization in i ntergroup b ehaviou r s. European Journal of Social Psychology 78(1), 149. Tallent - Ru nnels, M. K., Thomas, J. A., Lan, W. Y., Cooper, S., Ahern, T. C., Shaw, S. M., et al. (2006). Teaching courses online: A review of the research. Review of Educational Research, 76 , 93 135. Tichy, M., Johnson, D. W., Johnson, R. T., & Roseth, C. J. (2010). The impact of constructive controversy on moral development. Journal of Applied Social Psychology , 40 , 765 - 787. Toma, C., & Butera, F. (2009). Hidden profiles and concealed information: Strategic information sharing and use in group decision making. Perso nality and Social Psychology Bulletin, 35 , 793 - 806. Trautwein, U., Lüdtke, O., Marsh, H. W., Köller, O., & Baumert, J. (2006). Tracking, grading, and student motivation: Using group composition and status to predict self - concept and interest in ninth - grade mathematics. Journal of Educational Psychology, 98 , 788 - . Trautwein, U., Marsh, H. W., Nagengast, B., Lüdtke, O., Nagy, G., & Jonkmann, K. (2012). Probing for the multiplicative term in modern expectancy value theory: A latent interaction modeling study. Journal of Educational Psychology , 104 , 763 - . Tversky, A., & Kahneman, D. (1971). Belief in the law of small numbers. Psychological bulletin , 76 , 105 - 132 . Tversky, A., & Kahneman, D. (1973). Availability: A heuristic for judging frequency and probability. Cognitive psychology, 5 , 207 - 232. U.S. Department of Education, National Center for Education Statistics (2015). Digest of Education Statistics, 2013 (NCES 2015 - 011), Table 311.10. Retrieved from http://nces.ed.gov/fastfacts/display.asp?id=60 151 Valaitis, R. K., Sword, W. A., Jones, B., & Hodges, A. (2005). Problem - based learning online: perceptions of health science students. Advances in Health Sciences Education , 10 , 231 - 252. Vygots ky, L. S. (1978). Mind in society: The development of higher psychological processes. Cambridge: Harvard University Press. Walther, J. (1992). Interpersonal effects in computer mediated interaction: A relational perspective. Communication Research, 21 , 46 0 - 487. Walther, J. B. (1996). Computer - mediated communication: Impersonal, interpersonal, and hyperpersonal interaction. Communication Re search, 23 , 3 - 43. Walther, J. B. (1997). Group and interpersonal effects in international computer - mediated collaboration. Human Communication Research, 23 , 342 - 369. Walther, J. B. (1999). Die Beziehungsdynamik in virtuellen Teams (The relational dynamics in virtual teams). In M. Boos, K. J. Jonas, & K. Sassenberg (Eds.), Computervermittelte Kommunikation in Organisationen ( Virtual communication in organizations , pp. 11 - 25). Göttingen, Germany: Hogrefe. Walther, J. B. (2002). Time effects in computer - mediated groups: Past, present, and future . In P. Hinds & S. K iesler (Eds.), Distributed work (pp. 235 - 257). Cambridge, MA: MIT Press. Walther, J. B. (2008). The social information processing theory of computer - mediated communication. In L. Baxter & D. O. Braithwaite (Eds .), Engaging theories in interpersonal commun ication (pp. 391 - 404). Newbury Park, CA: Sage. Walther, J. B. (2009). Theories, boundaries, and all of the above. Journal of Computer - Mediated Communication, 14 , 748 - 752. Walther, J. B., & Bunz, U. (2005). The rules of virtual groups: Trust, liking, and performance in computer - mediated communication. Journal of Communication, 55 , 828 - 846. Walther, J. B., & Parks, M. R. (2002). Cues filtered out, cues filtered in: Computer - mediated c ommunication and relationships. In M. L. Knapp & J. A. Daly (Eds.), Handbook of interpersonal communication (3rd ed., pp. 529 - 563). Thousand Oaks, CA: Sage. Walther, J. B., & Tong, S. T. (2014). Inventing partners in computer - mediated communication: how CMC sustains self - fulfilling prophecies and relational attributions. In Social Computing and Social Media (pp. 517 - 527). Springer International Publishing. Walther, J. B., Van Der Heide, B., Tong, S. T., Carr, C. T., & Atkin, C. K. (2010). The effects of i nterpersonal goals on inadvertent intrapersonal influence in computer - mediated communication. Human Communication Research, 36 , 323 - 347. 152 media multitasking, person al needs, and gratifications. Journal of Communication, 62 , 493 - 513. Weaver, C. M., & Albion, P. R. (2005). Momentum in online discussions: The effect of social presence on motivation for participation. Proceedings of Australasian Society for Computers in Learning in Tertiary Education Conference (ASCILITE 2005), Balance, Fidelity, Mobility: Maintaining the momentum? (pp. 703 - 706). Sydney, Australia. Weinberger, A., Fischer, F., & Mandl, H. (2001). Scripts and scaffolds in problem - based computer supported c ollaborative learning environments: Fostering participation and transfer (Research report No. 144). Munich, Germany: Ludwig - Maximilians - University, Institute for Empirical Pedagogy and Pedagogical Psychology. Weisband, S., & Atwater, L. (1999). Evaluating self and others in electronic and face - to - face groups. Journal of Applied Psychology, 84 , 632 - 639. Wenzel, M. (2004). An analysis of norm processes in tax compliance. Journal of Economic Psychology, 25 , 213 - 228. Wenzel, M. (2005). Misperception of social norms about tax compliance: From theory to intervention. Journal of Economic Psychology, 26 , 862 - 883. Wenzel, M. (2005). Motivation or rationalization? Causal relations between ethics, norms and tax compliance. Journal of Economic Psychology, 26 , 862 - 883. Wickham, K. R., & Walther, J. B. (2009). Perceived behaviors of emergent and assigned leaders in virtual groups. In N. Kock (ed.), Virtual team leadership and collaborative engineering advancements: Contemporary issues and implications (pp. 50 - 64). Hershey , NY: Information Science Reference. Wilder, D. A. (1990). Some determinants of the persuasive power of in - groups and out - groups: Organization of information and attributions of independence. Journal of Personality and Social Psychology , 59, 1202 1213. Wi lliams, P. E., & Hellman, C. M. (2004). Differences in self - regulation for online learning between first - and second - generation college students. Research in Higher Education, 45, 71 - 82. Yamagishi, T. & Kiyonar, T. (2000). T he group as the container of gene ralized reciprocity. Social Psychology Quarte rly, 63, 116 - 132 . Young, J. D. (1996). The effect of self - regulated learning strategies on performance in learner controlled computer - based instruction. Educational Technology, Research and Development , 44 (2), 1 7 27. 153 Yu, F. Y. (2001). Competition within computer - assisted cooperative learning environments: Cognitive, affective, and social outcomes. Journal of Educational Computing Research, 24 , 99 - 117. Yu, F. Y. (2003). The mediating effects of anonymity and proxi mity in an online synchronized competitive learning environment. Journal of Educational Computing Research, 29 , 153 - 167. Yu, F. Y., Han, C., & Chan, T. W. (2008). Experimental comparisons of face - to - face and anonymous real - time team competition in a networ ked gaming learning environment. CyberPsychology & Behavior, 11 , 511 - 514.