As discussed in my post yesterday, that group polarization occurs is a fact. Why it occurs is more puzzling. What drives this phenomenon is not as simple as an inherent desire to conform that we all have. If a desire to conform were governing group polarization, we should expect the end result of the group to be no more extreme that the most extreme viewpoints of the majority of group members. But we know that group polarization can happen even when the majority of group members start out with rather moderate opinions.
The psychological processes driving group polarization require that there is a pre-existing tendency towards a particular moderate viewpoint that is held by the majority of the group members or by the most powerful among the group members. Call this the “initial condition.” Once the initial condition is in place polarization tends to occur via one of two processes: Social comparison or persuasion. The processes are often presented as different theories of group polarization but there is no good reason to think that they can’t both be in effect on different occasions and sometimes perhaps simultaneously during deliberation.
Social comparison occurs when people in a group are exposed to each others’ viewpoints or deliberate on a particular subject matter. We tend to assess ourselves by comparing ourselves to others, and we have a desire to be perceived favorably by relevant others, which makes us likely to assert a less extreme viewpoint, until we discover that others hold the same or a more extreme view. This, however, is only the beginning of the explanation of group polarization. If the viewpoint of other group members is more extreme than the opinion you originally asserted, then social comparison explains how you come to hold a more extreme opinion. But it doesn’t explain the emergence of a more extreme outcome for the group as a whole compared to any initial perspective of individual members. The main factor here is subpersonal one-upmanship (Isenberg 1986). Group members will tend to outdo each other as long as the more extreme viewpoint is considered desirable implicitly or explicitly. So, the group as a whole will often end up with a viewpoint that is more extreme than any starting point, when that general direction appears to group members to be in-group desirable. This is also known as a choice shift. Sanders and Baron describe it as follows:
This realization either “releases” the moderate members from their fear of appearing extreme or motivates moderate members to “compete” with the extreme members to see who can come closest to espousing the most admirable position. In either case, the moderates are motivated to adopt more extreme positions, while there is no corresponding pressure on extreme members to moderate their opinions (although, of course, simple conformity pressure may lead to some small amount of moderation by extreme members.) The net result is an overall polarization of opinions, that is, a choice shift. (Sanders and Baron, 1977: 304)
Misperceived group norms can also mistakenly be seen as the ideal that one should strive toward to be perceived favorably. This, together with subpersonal one-upmanship, can move a group to a more extreme viewpoint. What’s interesting is that once the group polarization has taken place, the initially misperceived group norms have automatically become the new norm. So, this is an example of a kind of normative bootstrapping. By adhering to a misperceived norm, that norm becomes the true norm for the group.
A striking example of how group opinions can be affected by misperceived group norms is the long-term abidance by the trans-woman-exclusion policies governing large queer and feminist events such as the Michigan Womyn’s Music Festival. For many years festival organizers and festival-goers have abided by an informal policy that only women who were assigned the female sex at birth should attend the event. This is also known as “the intention.” The intention concerning the intended attendees is captured by the slogan “womyn-born-womyn.” As many trans* activists, most notably transsexual activist Julia Serano, have pointed out on several occasions, the long abidance by this intention strongly turns on traditional norms in queer and feminist groups as well as plain old sexism. Traditionally, discrimination against trans* people from these corners consisted in derogatory anti-trans-women remarks or “witch hunts” by prominent feminist writers, such as attempts to prohibit hormone therapy and sex reassignment surgery.
The opinions underlying the exclusion policies of the Michigan Womyn’s Festival, however, moved toward extreme versions of some of the traditional viewpoints. The official explanation of why trans* women were not welcome at the festival was that they had experienced the male privilege during their upbringing and that some of them still have a penis. It was argued that experiencing the male privilege, albeit involuntarily, made trans* women fundamentally different from “womyn born womyn.” Having festival attendees with a penis at the event, furthermore, was asserted to be an impending threat to the women there both physically and psychologically.
While these opinions are bad enough, they don’t reflect the actual extremist viewpoints that group members came to hold over time as a result of group polarization. As Serano correctly points out, lesbians who have no problem with trans* men and who might even “lust after trannybois” find trans* women “creepy” and “effeminate” and don’t hesitate to openly state these viewpoints in their absence. Probably what initially made festival-goers and supporters move toward these extreme viewpoints was they seemed to fit the (misperceived) group norms, and as a result the misperceived group norms then became the implicitly accepted group norms.
Sanders, G. S. & R. S. Baron. (1977). Is social comparison irrelevant for producing choice shift? Journal of Experimental Social Psychology, 13(4) , 303 - 314
Serano, J. (2007). The Whipping Girl: A Transsexual woman on sexism and the scapegoating of femininity, Berkeley: Seal Press.