Effects of Persuasive Arguments on Group Polarization and Risky Shift Matthew Bresnahan November 26, 2013 The following research article is about the group polarization theory and the risky shift phenomenon. Through my research, it can be concluded the most probable cause of group polarization and risky shift effects to be the persuasive arguments in group communication. Group Polarization and Risky Shift 2 Abstract This study served as a replication of Myers and Bishop (1971) with minor modifications to the methodology to account for additional variables. This study examined the persuasive arguments explanation of group polarization and the risky shift phenomenon, the tendency of a group to make a more extreme decision than its members have made, in individual and group discussion settings. Fourteen undergraduate students (7 female, 7 male) were given a questionnaire with two choice dilemmas from Kogan and Wallach (1964). Individuals indicated the probability of success of the subject in the dilemmas. Participants were divided into two male and female groups, consisting of three and four participants respectively. After the choice dilemmas were rated again, means for both the individual and group conditions were calculated. Results indicated risky shift conditions in the smaller 3-subject groups and cautious shifts in the 4-person subject group, while the individual condition produced a neutralization effect on choice shifts. The present study’s findings supported both of the researcher’s hypotheses. It can be concluded in group discussion context, a greater emphasis is placed on the quality of the novel persuasive arguments, rather than the mere quantity of novel persuasive arguments, while in the larger subject group discussion context, a greater emphasis is placed on the quantity of the novel persuasive arguments, rather than the quality of novel persuasive arguments. Implications of the present study’s findings can be applicable in peer pressure situations due to group polarization and risky shift effects strengthening attitudes and behavior. These contributions may be of significant value to teachers, parents, guidance counselors and individual group members charged with reducing these effects. Suggestions for future research include a larger sample with a wider variety of participants from all undergraduate campus communities and further attempts to differentiate the social comparison and the persuasive arguments hypotheses. Group Polarization and Risky Shift 3 Research Problem The goal of this paper is to examine the effects of group polarization and risky shift on communication. For years, researchers have been confounded to pinpoint the exact cause of the two phenomena. The conclusions of the present study will be done through a replication of the Myers and Bishop (1971) study with some additional variables. This paper will shed light on the absolute cause of these effects. In order to properly discuss these perplexing concepts, the consensus has to be reached about what constitutes the appearance of these effects. The basis for this will be covered by three definitions: choice shift, group polarization and risky shift. Boster and Mayer (1980) asserted, “a choice shift results to the extent that group members who had not previously considered these partially shared arguments are convinced of their merits" (p. 394). Group polarization can be defined as “an increase in the extremity of the average response of the subject population” (Rothwell, 1986, p. 187). The risky shift phenomenon occurs “when the average of individuals' decisions made prior to discussion was compared to the decision made by those same persons in a group discussion context, the group decision (tends) to be riskier” (Cline & Cline, 1980, p. 26). Although the census is that the effects, if apparent, occur in group communication, many researchers differ on what mechanism or prospective produces the effect of group polarization and risky shift. With no clear explanation of the effects, there are a myriad of perspectives, two of which will be detailed further in the review of the literature, that try to explain what yields group polarization and risky shift effects. As will be demonstrated, group polarization and the risky shift phenomenon play in an influential role in the power of persuasion that impact group communication and thought. Two hypotheses are proposed by the researcher about this experiment. Hypothesis 1: Subjects in smaller formed groups during discussions will produce better quality novel persuasive arguments and the addition of subjects to smaller formed Group Polarization and Risky Shift 4 groups during discussions will yield more novel persuasive arguments, creating risky and cautious shift conditions respectively in those settings. Hypothesis 2: Individuality in these conditions will produce a neutralization effect regardless of gender or the effects produced by the introduction of a confederate. Review of Literature The following review of literature focuses on three areas: First, background and misconceptions; second, an examination of the competing research perspectives as an ideological backing for the group polarization and risky shift effects, third, how the competing research perspectives of the social comparison theory, the persuasive arguments theory coexist. History and misconceptions As Cline and Cline (1980) noted, the risky shift phenomenon was born when Stoner (1961) “discovered that, contrary to popular belief, groups take greater risks in decision making than do individuals (p. 26). This opened a new area of research, as researchers flocked to it trying to replicate and explain the phenomenon. In some attempts to explain the phenomenon, researchers’ findings have been easily conflicted. As Rothwell (1986) points out, while Teger and Pruitt conclude that "a risky shift is almost always found,” some studies have found shifts that favor conservatism (p. 182). It can be shown, then, that risky shift is only part of the more general phenomenon of group polarization. Even after thousands of experiments, researchers struggle to isolate the theoretical mechanism for group polarization or more specifically the risky shift phenomenon. As a result, there have been numerous perspectives, none more prevalent than Group Polarization and Risky Shift 5 the social comparison theory or the persuasive arguments theory, proposed to try to make sense of this decision making phenomenon. Social comparison theory In group discussion, individuals have different opinions on what or how something should be done. Every member of the group wants to fit in to avoid being blackballed for being different. The social comparison theory proposes “…any action that allows one to discover, either implicitly or explicitly, others' positions on an issue leads to a choice shift, with the direction of the shift dependent upon the way others respond” (Mayer, 1985, p. 92). Boster and Mayer (1980) explain that group members seek to compare themselves with other group members on characteristics such as risk or caution. Going into the group discussion, individuals believe that they are in agreement on those characteristics, however, if they are not, they view themselves as incorrect. Mayer (1985) argues under the social comparison theory, “risk is valued and that people perceive themselves as riskier than others” (p. 92). Lamm (1988) notes college students tend to underestimate how much risk the other members are going to take in their decisions. At that point in the group discussion, Kogan and Wallach (1964) argue that in attempt to save face and preserve their image in the group, individuals seek consistency in making a decision and their confidence in it. “Whichever is preferred, the person's introspections concerning judgmental confidence seem to follow suit in a reflective fashion“(Kogan & Wallach, 1964, p. 195). Lamm (1988) notes this is an example of the self-presentation theory, in which individual change their positions to present a more favorable image of themselves to the group. Whitney and Smith (1983) found “…bidirectional attitude polarization will follow interactions between groups in which cohesiveness is emphasized…” (p. 172). Pavitt (1994) argues social comparison theorists have disregarded “the basic psychological processes involved in what is basically a problem in Group Polarization and Risky Shift 6 attitude formation and change” (p. 642). With individuals going with the majority opinion of the group, social comparison creates “conditions [that] would reduce the individual member's receptivity of information (Whitney & Smith, 1983, p. 167). There also exists the release theory within social comparison. This theory asserts those groups members are afraid to voice radical opinions until they see others supporting it, then it becomes acceptable to voice such an extreme position (Lamm, 1988). Isenberg (1986) argues that under the guise of pluralistic ignorance, group members compromise so that their decision lands close to their decision, while still remaining close to the majority opinion of the group. Sometimes, individuals feel the effects of social comparison by just being exposed to viewpoints of others in the group to compare to under the mere exposure effect (Boster & Hale, 1988). Boster and Hale (1988) found mere exposure is “is assumed to be necessary and sufficient to produce choice shifts discussion were consistently larger than those resulting from mere exposure” (p. 182). They also noted a correlation between ambiguity and the impact social comparison had on decisions (Boster & Hale, 1988). The other research prospective examined in this review to explain the effects of group polarization and the risky shift is the persuasive arguments theory. Persuasive arguments theory Groups can make their members change their opinions, causing a choice shift, based on the persuasive arguments other group members make for their position. The persuasive arguments theory predicts “group members modify their initial opinions because they are persuaded by arguments presented during group discussions” (Boster & Hale, 1988, p. 180). Boster and Mayer (1980) claim these arguments start out from a larger group of arguments deciding between alternatives, and certain arguments are advocated during group discussion. Isenberg (1986) notes the presence of a choice shift and how a group polarization function is Group Polarization and Risky Shift 7 dependent how novel persuasive arguments are presented to the group. Isenberg (1986) states if the arguments are old information, a choice shift will occur in the direction of the one who is arguing for it, however, “If novel persuasive arguments are presented that are opposite to the direction initially favored by the group member, their position will shift in the opposite direction and depolarize” (p. 1145). Mayer (1985) argued persuasive arguments are more effective when they support the eventual choice shift than when not. Boster and Mayer (1980) found that the perception of quality of persuasive arguments influenced attitudes and strongly influenced post message attitudes. The persuasive arguments theory is not solely based on argument quality; it can also be based on the number of arguments. “Participants in group polarization research decide on a prediscussional recommendation based on the proportion of arguments on either side of the issue” (Pavitt, 1994, p. 640). “Subjects perceived the quality of risky arguments to be higher when the majority was risky than when the majority was cautious“(Boster & Mayer, 1980, p. 401). Research has found “that novelty leads to persuasiveness only when perceived validity is high” (Isenberg, 1986, p. 1148). The persuasive arguments theory takes into the psychological, but disregards the social factors that the social comparison theory covers (Pavitt, 1994). Even though the social comparison theory and persuasive arguments theory try to serve as standalone theories for explaining the effects of group polarization and the risky shift. Two theories intertwined On the surface, it would appear that the social comparison theory and the persuasive arguments theory, while similar, are able to be distinguished from one another. Isenberg (1986) argues the two theories coexist. His meta-analysis found a moderate composite relationship amongst the two theories and polarization. Boster and Mayer (1980) note the possibility for both theories to influence a choice shift’s direction and magnitude. Majority, which played a Group Polarization and Risky Shift 8 significant role in both theories, is interrelated. “…Majority manipulation induces a social comparison process. The social comparison process, in turn, affects perceptions of the quality of persuasive arguments (Boster & Mayer, 1980, p. 405). Boster and Hale (1988) found the choice shifts, both cautious and risky, were supported by the explanation that the social comparison and persuasive arguments theories operate at the same time in explaining group polarization and the risky shift. Mongeau and Garlick (1988) stated how Boster and Mayer found …”social comparison information became the basis on which to judge persuasive argument information” (p. 121). However, in their study they refuted Boster and Mayer’s (1980) findings because the participants disregarded the information when attitudes were formed. Critical Evaluation Ever since group polarization and the risky shift phenomenon were proposed, it has been a fertile area of research. Even though there are still the competing explanations of social comparison theory and the persuasive arguments, research within the perspectives has been commendable. In the social comparison perspective, Kogan and Wallach (1964) and Boster and Hale (1988) need to be recognized for their important contributions to this important perspective of group polarization and risky shift. In addition to their landmark study that created the choice dilemma questionnaire, which has become the standard instrument in group polarization and risky shift studies, their emphasis on the social cognition function of the social comparison theory is an area lacking sufficient research. With further research, the social cognition function might be what separates these two theories in this area of research. Bolster and Hale (1988) subcategorized what can cause a risky shift in the social comparison theory and how they differ from each other. Group Polarization and Risky Shift 9 In the persuasive arguments prospective, Boster and Mayer’s (1980) findings have to be recognized. Boster and Mayer (1980) found the important link of the perception of the quality of the arguments and post message attitudes. This validates the theory as a whole in the realm of group polarization and risky shift effects. This is due to quality of the arguments being a more reliable persuasive factor than merely the quantity of the persuasive arguments. The examination of group polarization and the risky shift phenomenon has become too muddled to isolate one explanation that some researchers have turned to trying to validate dual processes as a valid explanation for the confounding phenomenon. Boster and Hale’s (1988) examination use of dual models facilitates a need for the process to be replicated for future research. The dual model finally allowed researchers to examine real-world explanations of those results, rather than just theoretical backing, as made clear in previous research. The body of the group polarization and risky shift research provides limitations that need to be corrected for future studies. The proliferation of the Kogan and Wallach (1964) choice dilemma questionnaire has created a problem that plagues the spectrum of group polarization and risky shift research. The proliferation creates two problems for studies using it. Due to the widespread use of the choice dilemma questionnaire as the universal instrument for such studies, with the majority of the field using it, there cannot be many distinctions between studies because the results are derived from the same universal instrument. The second problem the Kogan and Wallach (1964) choice dilemma questionnaire creates is its lack of consideration for the more complex psychological and social factors behind decision making and its lack of real-world variables in its scenarios, such as money. As Rothwell (1986) argues, since the participants in the studies are not active participants in the choice dilemmas, making conclusions based on data from hypothetical scenarios is rather difficult. The present study seeks to reduce these limitations Group Polarization and Risky Shift 10 by asking fewer choice dilemmas. With the reduction in the number of choice dilemmas, it will allow the participants to more critically evaluate the choice dilemmas using persuasive arguments. Method The experiment consisted of 14 participants (7 females, 7 males) that were recruited from Edinboro University of Pennsylvania. Ten of the fourteen undergraduate participants were part of the disabled community on the campus. Participants were volunteers and received no incentive for participation. All participants were native English speakers and had the physical ability to print. Participants were given an activity worksheet based off of the Kogan and Wallach (1964) choice-dilemma questionnaire. The questionnaire consisted of two real-world scenarios, in which individuals had to rate the chances of success for the person in the given scenarios. The students were then divided into two groups of three and two groups of four. The groups were divided by gender. In addition, one group member in the male subject group of three and the female subject group of four was a confederate. The groups engaged in a discussion on the choice dilemmas until a census was reached. The researcher encountered a problem in the collection of the data. First, the students did not follow the activity’s directions correctly. There was also a concern for the validity of the data collected given the small sample size. In regards to the sample size, the 14 subjects were randomly sampled. Even though the questionnaire has been previously developed, it still provided valid and reliable responses because the scenarios are realistic and had some reliability with the subject Group Polarization and Risky Shift 11 population. The study utilized scenarios derived from the Kogan and Wallach (1964) questionnaire because that is the standard measurement in similar studies. The subjects were monitored during the activity to ensure the activity’s instructions were followed properly. In this experiment, the controlled variable was the questionnaire. Several variables were being measured and manipulated to access their effect on group polarization and risky shift. These variables include: gender, group size and majority opinion. The gender variable was measured by having two male and female subject groups. The group size variable was measured by having two groups of three and two groups of four. The experiment also sought to determine if putting confederates in two of the four groups affected majority opinion and induced greater effects of group polarization and risky shift. The limitations on the methodology are the limited sample size and the inability to eliminate complete rival hypothesizes, but it reduced the effects of the rival hypothesizes with the use of the confederates in the use of experiment. The confederates tried to go against the majority opinion, opting for either more extreme or cautious decisions, which tried to reduce the effects of the rival hypothesizes during the experiment. Group discussion has a possibility of affecting external and internal validity. Subjects may shift their opinions during and after the group discussion. The confederated might have an impact on the external validity because they might skew the results and give group polarization and risky shift more pronounced effects. If there was to be any bias in this experiment, it would be personal bias towards subject or scenario. This will be controlled by the group discussion since discussion is made on the basis of group consensus. In this experiment, ethics might come into play when dealing with the confederates because they actively were trying to manipulate the group discussions, so the effects of group polarization can become more pronounced. This was controlled by two of the groups not having Group Polarization and Risky Shift 12 a confederate. They were debriefed when they are needed for an opinion questionnaire on social issues and that the questionnaire activity should take between 30-45 minutes. It remained ethical because the debriefing is not lying nor is restricting their right to consent. Data Analysis Fourteen undergraduate students were gathered for a group polarization exercise where they answered two choice dilemmas from Kogan and Wallach (1964) on the basis of probability of success in the given scenarios. Ten of the fourteen undergraduate participants were part of the disabled community on the campus. A copy of the questionnaire can be found in Appendix B. The choices provided to the participants were 1, 3,5,7, and 9, with 1 being the most limited probability of success and 9 being the most probable to happen. Let it be stressed that there is also a choice for there to be zero chance for it working. Subjects were first given the questionnaire individually, and then they were divided into groups, two male and female groups with group sizes of three and four subjects. After the first section of the experiment, the individual totals were calculated. This was also done for the group discussion totals. Results of the individual and group mean calculations can be seen in Table 1 in Appendix A. Anything that resulted above 5 indicated varying degrees of risk. Conversely, anything below 5 indicated varying degrees of caution. The individual mean was 5.2. This data indicates no shift in either the cautious or risky direction. This is because the mean lies just beyond the neutral point of 5 of the scoring. This supports the second hypothesis. Hypothesis 2 stated: Individuality in these conditions will produce a neutralization effect regardless of gender or the effects produced by the introduction of a confederate. The male and female groups, as well as the results based on group size and natural vs. manipulated groups, where compared against each other to see if the variables have any correlation with group polarization and risky shift. The three subject male Group Polarization and Risky Shift 13 group mean was 6.7 and the four person male group mean was 4.5. The data indicated a moderate shift in the risky direction for the three subject male group. This is due to the three subject male group mean being noticeably greater than the individual mean in the risky direction. The three person female group mean was 7.6 and the four person female group mean was 3.5. The data indicated a shift in the risky direction for the three subject female group. This is due to the three subject female group mean being noticeably greater than the individual mean in the risky direction. The data also indicated cautious shifts for both the four subject male and the four subject female groups. This is due to the four subject male and the female group ranges being more cautious than the individual mean in the risky direction. The data indicate that gender did not play a role in signifying a risky or cautious shift. While the data signal a risky shift for the three subject male and female groups and a more cautious shift in the four subject male and female groups, based on the research and empirical observation, this can be attributed to the group size and its effects on novel persuasive arguments in group discussion. The data indicated that the natural vs. manipulated group conditions did not play a role in signifying a risky or cautious shift. This is due to the manipulated groups, which were the three subject male group and the four subject female group, had confederates, who advocated or the minority opinion position based on individual means, had the three subject male group moderately shift in the risky direction, while the data for the our subject female group indicated a cautious shift. Therefore, there was no correlation between the manipulated groups. The data was used to test the first hypothesis due to the shifts in this experiment that were affected more by the persuasive arguments theory. Hypothesis 1 stated: Subjects in smaller formed groups during discussions will produce better quality novel persuasive arguments and the addition of subjects to smaller formed groups during discussions will yield more novel Group Polarization and Risky Shift 14 persuasive arguments, creating risky and cautious shift conditions respectively in those settings. In the two three subject groups, one of which was male and the other female, the responses became more polarized, causing a risky shift. Based on empirical observation and the research, this is due to an individual member’s novel arguments carrying more weight in a smaller subject group. Since there were shifts indicated in these situations, the explanation indicated persuasive arguments and not social comparison theory. The statistics used in this experiment were the initial (individual) mean for the choice dilemmas and the group mean for the choice dilemmas. Discussion The results of the present study indicate a significant correlation between persuasive arguments theory and the group polarization phenomenon. The results become more polarized in the two, three subject group conditions in relation to the individual mean of the two Kogan and Wallach (1964) choice dilemmas provided in the questionnaire located in Appendix B. Participants rated the probably of success of the given scenarios significantly higher in the three subject group condition than in the individual or four subject group conditions, which indicated a risky shift. Conversely, participants rated the probable success of the given scenarios significantly lower in the four subject group condition than in the individual or three subject group conditions, which indicated a cautious shift. These findings suggest that an individual’s novel persuasive arguments have more influence on group discussion in smaller subject groups than in larger subject groups. In other words, in the smaller subject group discussion context, a greater emphasis is placed on the quality of the novel persuasive arguments, rather than the mere quantity of novel persuasive arguments. Conversely, an individual’s novel persuasive arguments have less influence on group discussion in larger subject groups than in smaller subject groups. In other words, in the larger subject group discussion content, a greater emphasis is placed on the Group Polarization and Risky Shift 15 quantity of the novel persuasive arguments, rather than the quality of novel persuasive arguments. The hypotheses for this experiment were that smaller subject group discussions would produce better quality novel arguments and larger subject group discussions would produce a greater number of novel persuasive arguments leading risky and cautious choice shift conditions respectively. The experimenter also theorized that individuality would have a neutralization effect on choice shifts. The present findings support these hypotheses as participants in the three subject groups ranked a significantly higher probably of success for the given scenarios than the individual mean results, while participants in the four subject groups ranked a significantly lower probably of success for the given scenarios than the individual mean results. While the participants in group discussion produced risky and cautious shifts, individual mean results indicated no choice shift in either direction. The present findings are in agreement with Boster and Mayer (1980), who found that the perception of quality of persuasive arguments influenced attitudes and strongly influenced post message attitudes. Their findings also indicate that social comparison also played a factor in the experiment because “…majority manipulation induces a social comparison process. The social comparison process, in turn, affects perceptions of the quality of persuasive arguments (Boster & Mayer, 1980, p. 405). This is concurrent with the present findings because in the three subject male and female group where the risky shift was induced, the majority position already leaned in the risky direction due to the group members in those subject groups initially had risky individual means. This underscores the difficulty of isolating a research prospective that fits the group polarization and risky shift phenomenon. The present study also supports Mayer (1985) that persuasive arguments are more effective when they support the eventual choice shift than when not. The present findings further support hypothesis 2 and Boster and Mayer (1980) who Group Polarization and Risky Shift 16 found that “Subjects perceived the quality of risky arguments to be higher when the majority was risky than when the majority was cautious“ (p. 401). The present findings support hypothesis 2 and Lamm (1972) that group means were higher and more polarized than individual means. The present findings also support Billig (1976) because “It has been found repeatedly that group attitudes and judgments will be more polarized than individual attitudes and judgments “(p. 495). The present findings indicated that there were no gender effects that contributed to either a risky or cautious shift, which supports Myers and Bishop (1971). However, the present findings refute Myers and Bishop (1971) and Boster and Mayer (1980) assertions that there were no significant group size effects on choice shifts. In the present study, the results indicated group size effects influenced the persuasive argument quality, which the greater the argument quality the riskier the shift was. The present findings do not support the Boster and Mayer (1980) finding that persuasive argument manipulation was successful in causing and determining the direction of a choice shift. In the present study, there was no correlation found between argument manipulation and choice shift because the two groups, the three subject male group and the four subject female group, which were being manipulated by a confederate indicated a risky and a cautious shift for the two groups respectively. Therefore, no correlation could be made between persuasive argument manipulation and the direction of the choice shift. Conclusion Two potential problems with this study are the limited population from which the participants were sampled from and the coexistence of both the social comparison and persuasive arguments theories as explanation for the group polarization and risky shift phenomenon in the present study’s findings. An experiment consisting of fourteen participants is not a representative sample of a more general population. Ten of the Group Polarization and Risky Shift 17 fourteen undergraduate participants were part of the disabled community on the campus. This affected the perspective during the group discussion. The relational status between participants affected the diversity of novel persuasive arguments during group discussion. A way to eliminate this confound is by doing the experiment again using a larger sample with a wider variety of participants from all undergraduate campus communities. One of the present study’s findings indicated the coexistence of the social comparison and persuasive arguments theories as an explanation for the effects of group polarization and risky shift. The coexistence of these theories makes it difficult to for future research to isolate the theories to determine their effects in causing group polarization and the risky shift phenomenon. A way to eliminate this confound is by doing the experiment again using pre-taped group discussions for participants to listen to in place of authentic group discussions. In general, the present study’s findings prove to be applicable for groups engaging in peer pressure. First, there would be stronger peer pressure effects in smaller groups due to an emphasis on the quality of persuasive arguments. As Boster and Mayer noted, individuals view risky persuasive as higher when the majority position of the group shift towards riskier decisions. This would apply to the real-world context by the majority position of the group is to engage in juvenile delinquency behavior so, individual members of the group think it is a good plan due to it being the majority position of the group. Second, the lack of gender effects on choice shifts indicates that the implications of the present study’s findings on peer pressure can be applied universally because gender will not significantly affect peer pressure effects. A third implication of the present study‘s findings on peer pressure is being in a group will consistently polarize group decisions more than individual decisions, as consistent with Billig (1976) findings. The Group Polarization and Risky Shift 18 more bipolarized decision has a tendency to lead to more extreme behaviors because polarization reinforces and strengthens attitudes. This is applicable in peer pressure settings because what starts out as juvenile delinquency behaviors evolves into more serious criminal offenses due to group attitudes being strengthened as a result of the polarization effects in group discussion and behavior. The influence of this process on peer pressure is such that these contributions may be of significant value to teachers, parents, guidance counselors and individual group members charged with reducing the effects of group polarization and risky shift on peer pressure. References Boster, F. J., & Mayer, M. (1980). Choice shifts: Argument qualities or social comparison. In D. Nimmo (Ed.), Communication Yearbook (4th ed., Vol. 2, pp. 393-410). New Brunswick, N.J.: International Communication Association. Group Polarization and Risky Shift 19 Billig, M. (1976). Judgments of values and group polarization: Tests of the value-for-risk hypothesis. European Journal of Social Psychology, 6(4), 495-501. Boster, F. Hale, J. (1988). Comparing effect coded models of choice shifts. Communication Research Reports, 5(2), 180-186. Cline, R., & Cline, T. (1980). A structural analysis of risky-shift and cautious-shift discussions: The diffusion-of-responsibility theory. Communication Quarterly, 28(4), 26-36. Isenberg, D. J. (1986). Group polarization: A critical review and meta-analysis. Journal of Personality & Social Psychology, 50(6), 1141-1151. Kogan, N., & Wallach, M. (1964). Risk taking: A study in cognition and personality. New York: Holt, Rinehart & Winston. Lamm, H. (1988). A review of our research on group polarization: eleven experiments on the effects of group discussion on risk acceptance, probability estimation, and negotiation positions. Psychological Reports, 62807-813. Lamm, H. (1972). Self-image, perception of peers' risk acceptance and risky shift. European Journal of Social Psychology, 2(3), 255-272. Mayer, M. E. (1985). Explaining choice shift: An effect’s coded model. Communication Monographs, 52, 92-101. doi:10.1080/03637758509376097 Mongeau, P., & Garlick, R. (1988). Social comparison and persuasive arguments as determinants of group polarization. Communication Research Reports, 5(2), 120-125. Myers, D. G., & Bishop, G. D. (1971). Enhancement of dominant attitudes in group discussion. Journal of Personality & Social Psychology, 20(3), 386-391. Pavitt, C. (1994). Another view of group polarizing: the “reasons for” one-sided oral argumentation. Communication Research, 21625-642. doi:10.1177/009365094021005004 Group Polarization and Risky Shift 20 Rothwell, J. (1986). Risk-taking and polarization in small group communication. Communication Education, 35(2), 182-185. Whitney, J. C., & Smith, R. A. (1983). Effects of group cohesiveness on attitude polarization and the acquisition of knowledge in a strategic planning context. Journal of Marketing Research (JMR), 20(2), 167-176. Appendix A. Table 1: Ind. & Group Means Subjects Individual Male Group Female Group Male Group Female Group (Means)* 3** 3* 4* 4* Group Polarization and Risky Shift 21 1 4 2 4 3 5 4 5 5 6 6 6 7 8 8 3 9 3 10 3 11 6 12 8 13 8 14 7 6.7 7.6 OVR IND AVG=5.2 *Data represents means of two choice dilemmas combined **Included a confederate 4.5 3.5 Group Polarization and Risky Shift 22 Appendix B. Group Polarization and Risky Shift 23