Group Polarisation Research Paper

AbstractThis experiment is a replication of Wallach, Kogan and Bem (1962) study on Group influence on individual risk taking. The aim for this experiment is to investigate the effects of shift in risk across the pre-discussion, group-discussion and post-discussion assessment on individual decisions and group consensus. There were 154 (both sexes) undergraduates students from the University of Sheffield used in this study.

The results were based on their opinions on the pre, group and post-discussion assessment and data were collected to examine whether individuals risk more as a result of group-discussion than when they were alone. Findings suggested that group-discussion does made an influence on decision-making on individual’s initial decision.

Many of us spend a significant portion of our lives making decisions individually or in groups. We make decisions about all sorts of risks differently when we are within a group or when we are alone. Wallach, Kogan and Bern (1962) found an averaging effect between individual decisions and the group consensus. It was believed that groups might seek to reach an agreement with group members that made group members move from their initial decisions. In 1961, Stoner identified a phenomenon that later became known as the “risky shift”.

The phenomenon explains that group discussion produced group decisions that chose riskier decision than the average of the individual group member prior decisions. However, the participants that were used in his study were all male graduate students of industrial management and this might cause a bias results. It was believed that they were more willing to take risks because it is a desirable attribute of a potential manager (Stoner, 1961).

Wallach, Kogan and Bern (1962) went on to replicate Stoner’s findings on “risky shift” and found out a much wider perspective that it is now referred to as “group polarization”. They did a similar study but with an addition of having participants for their individual opinions after the group discussion.

The result of their studies showed that there was a tendency that group consensus produce riskier decisions than the average individual opinion prior to discussion (Hogg & Vaughan, 2005). They also found out that after the group discussion, individual decisions were on average riskier than prior to group discussion. Through the post group discussion, individual’s risky shift on their decision was still maintained 2 to 6 weeks, this indicates that these group decisions were not just an overt conformity but as an acceptance of the new decision.

Furthermore, group consensus can sometimes be riskier or cautious in making decisions, in the direction favoured by the mean of the individual members’ initial positions. One possible explanation was proposed by Wallach, et. al (1964) that greater risks are chosen due to a diffusion of responsibility. Another explanation was explained in terms of social comparison/cultural values and persuasive arguments theory (Hogg &Vaughan, 2005)

The current study sets to replicate the main findings of the original Wallach, et. al (1962) study on group polarization. The purpose of this study is to examine at the relationship between the individual decisions and the group consensus based on a pre-discussion (individual decision), group-discussion and a post-discussion (individual decision) assessments.

All of the three assessments have to reach to a unanimous decision. The hypothesis is to test whether if group decisions show a shift in riskier decision compared with the pre-discussion decision and finally whether post-discussion decision will show a shift in riskier decision compared with pre-discussion decision. Method

ParticipantsThere were 154 male and female participants, ranged in age from 18 to 28, with a mean age of 19.8 years (SD = 3.2) in this study. All participants were second year undergraduate psychology students from the University of Sheffield. Participants were an opportunity sample from the Social Psychology lab class and were allocated to conditions haphazardly. Apparatus/ Materials

Pre, group and post opinion questionnaires that includes instructions and answer sheets were used in the first practical. There were 6 different opinion questions on the opinion questionnaires that describes different social dilemmas situation. Most of the questions were a replication from Wallach, et. al. (1962) experiment on “Group influence on individual risk taking”. The answer sheets were used to obtain data on the level of shift in risk across each assessment. Computers were used in the psychology building’s computer lab to key in the pre, group and post answers from each student. Design

The experiment had a one-way within subjects (repeated measures) ANOVA design. The independent variable was the opinion questionnaire within each assessment (pre, group and post discussions) and the dependent variable was the answers given by the participants about their decisions on risk as a result of pre group and post discussions. Procedure

Participants were given some instructions about the opinion questionnaires that describe different social dilemmas. Participants were then asked to consider a series of social dilemma questions from the opinion questionnaire. Firstly they were asked to answer individually (pre-discussion) then participants were split into smaller groups (group-discussion) and were asked to discuss each of the situations and reach a unanimous decision over what the answer should be. Finally, participants were then asked to give their individual opinion about the same problems once again individually (post-discussion). Results

As shown in Table 1, it demonstrates a clear difference in the average level of risk recorded across each condition. The table indicates that individuals in pre-discussion reported to risk less (M = 5.70, SD = 1.20). As for the group-discussion, with a unanimous decision, participants reported to risk more (M = 5.14, SD = .651). As a result of group-discussion, individuals reported to risk more in post-discussion (M = 5.20, SD = .822).

The error bars are small and approximately the same size across each assessment, this suggests that the data are grouped closely around the means. A one-way, within participants (repeated measures) ANOVA indicated that there was a significantly difference across all three discussion, F(2) = 30.1 , p < .001

Table 1Shift in Risk on Pre-Discussion, Group-Discussion and Post-Discussion (N = 154)

______________________________________________________________ VariableMean MedianModeSD______________________________________________________________ Pre5.708.702.331.20



The results shows that the average level of risk taken in the pre-discussion is significantly different from the risk recorded in the group-discussion F(1) = 32.23, p = .000. Moreover, the average level of risk taken in the post-discussion is also significantly different from the risk taken in the pre-discussion F(1) = 37.0, p = .000 . Therefore, the unanimous group decisions did show a shift towards greater risk taking when compared with pre-discussion, while there is also a shift towards greater risk taking in post-discussion compared with pre-discussion.

DiscussionThe results are consistent with the experimental hypothesis: The decisions that were made in the pre, group and post discussion did show a significantly difference from each other. The unanimous group decisions showed a shift in riskier decision compared with the pre-discussion individual decisions, while the post-discussion individual decisions also showed a shift in riskier decision compared with pre-discussion individual decision. The findings did support the concept of Group Polarization.

It showed that when individual were put into group discussion, they were willing to accommodate greater risks collectively than they would as an individual. Perhaps the most obvious explanation for this finding is that risky shift revolves around the notion that risk taking is a cultural value in itself (Hong, L.K., 1978), people tend to seek social approval and try to avoid social criticism (Hogg. M. A., Vaughan, G. M., 2005). Group discussions reveals which views that are socially desirable or culturally valued, so in order to gain approval, individuals will shift in the direction favoured by the group (Hogg. M. A., Vaughan, G. M., 2005).

Moreover, Forsyth (1990) said that studies shown that most people tend to exhibit feelings of admiration for others who perceived as being greater risk-takers. In the current studies shows that participants were willing to show greater risk than the average person, one possibility might be that they believe that showing themselves as a risk-taker might enhance their appearance in the group. However, the current study used students from the same university, same level, same department and also 80% of the class room students are British, this can cause the findings to be biased.

Even though that social comparison and cultural values is one of the reasons to explain about risk-shift, we have to keep in mind that culture provides a wide range of different beliefs and norms. Although Forsyth (1990) had made his point clear that there is evidence on the general idea of people in the society value risk, but this doesn’t necessary apply to other culture practices as well. Some culture for instance was brought up from an internalized oppressive culture that hinder their ability to take risk or give opinion.

Some might not even have the opportunity to go through the development of social-comparison, whereby to compare their own opinions with others in order to establish the socially approved way of thinking. Having to live through oppression, many people are forced to relinquish their sense of leadership and development to make any risky decision or opinions for themselves; this distinguishes the differences in cultural values on risk-shift behavior. In conclusion, the current study findings did provide evidence to support the hypothesis that there will be a difference in the average level of risk across the pre, group and post assessments.

Group Polarization phenomenon ties in well with social influence but future studies should investigate whether social influence does have an impact on shifts in risk within different social groups. It is important to recognize this issue and further research on risk-shift behavior in groups and individual with a diversity of people from different culture.

ReferencesBrown, R. (2000). Group Processes. (2nd Ed.) Oxford: Blackwell. pp. 197-212. Forsyth, D.R. (1990). Group Dynamics (2nd Edition). Pacific Grove, CA;Brooks/Cole Publishing. Stoner, J. A. F. (1961). A comparison of individual and group decisions involvingrisk. Unpublished master’s thesis, Massachusetts Institute of Technology,School of Industrial Management. Wallach, M. A., Kogan, N., & Bem, D. J. (1962). Group influence on individual risktaking. Journal of Abnormal and Social Psychology, 65, 75-86. Hong, L.K. (1978). Risky shift and cautious shift: some direct evidence on theculture-value theory. Social Psychology, 41: 342-346. Hogg, M. A., & Vaughan, G. M. (2005). Social psychology. 4th edition, pp. 341-343.