A summary of Stasser, G., Titus, W. (2003). Hidden Profiles: A Brief History. Psychological Inquiry 14(3, 4), 304-313.
Jack and Jill want to get to the top of the hill safely. Should they not share all the info they have on how to do this? Stasser & Titus (1985) suggests not; they would tend to share only information they both know – some information remains hidden. In this way, the idea that ‘group decisions are better than those by individuals alone’ is eroded.
Stasser and Titus’ 1985 study: Four person teams of university students considered three candidates for student government: Best, Okay, and Ohum. When all information about the three was completely known beforehand to all four judges, 83% chose Best based on group discussion. When information was shared disparately (and where to any one member Okay seemed to have more desirable traits) 61% of individuals chose Okay, and 71% of groups chose Okay; group discussion increased preference towards Okay (rather than Best, which should be the case if all information was shared and considered without bias by all).
How did Stasser & Titus (1985) come upon the idea? This was associated with the idea of informational social influence and normative social influence in group processes. Stasser and David (1981) found preference certainty was affected by informational influence, preference itself was affected by normative influences. This is associated with Burnstein & Vinokur’s (1977) persuasive arguments theory (PAT): group polarization occurs because “members are sampling from a culturally shared pool…; individuals form their initial judgments by sampling from the pool, and these initial judgments will reflect the preponderance and credibility of the pro and con arguments in the pool” (Stasser & Titus, 1985, p. 306, top-left). Key propositions of PAT are that (a) group polarization is driven by persuasiveness of arguments, and (b) persuasiveness depends partly on novelty of information. But, this ignores confounding influence of pre-existing leanings of individuals: individuals, via group discussion, discover others also share their view, strengthening their opinions (c.f. spiral of silence in media, where majority opinion (held individually but not aired publically) is quelled by strong voice of media, making the actual majority less willing to speak up). And, “PAT assumes that novel arguments are…consistent with widely shared arguments,” (Stasser & Titus, 1985, p. 306) but this is not necessarily the case. What if novel arguments are unique to shared arguments? Thus, hidden profiles was born.
The idea was that shared information would be old-hat, and the new novel information would be like a lightning bolt to spur preference change. This was not the case. “What matters is how many support a position at the onset of discussion, not what is said during a discussion” (p. 306).
There is a problem with this reasoning though: sampling bias. Common information is more prevalent, therefore will be talked about more than unique information (collective information sampling, CIS). Dynamic collective information sampling (DCIS, Larson, Foster-Fishman and keys (1994)) develops this further, saying that early on in a discussion, the shared information pool will be depleted fast, with unique information being discussed later on, but by then, “opinions are often formed and consolidated early, diminishing the impact of information that emerges late in the interaction (p. 307). This finding was found not with university students but with medical teams discussing patient symptoms, meaning an ecological validity of the findings.
Social costs of unique information. Unique information is not repeated as much as common information during discussions; this is a non-intuitive phenomenon. Lower status individuals less likely to repeat unique information (Larson et al., 1996). This suggests that “the bearer of unique information…incurs some social costs” (p. 308). Lower status members are reluctant because there needs to be some vouching of the accurateness of the information. Labelling some members of groups as experts increased mentioning of unique information (Stewart and Stasser, 1995). This suggests that “social validation of information that emerges during discussion is important in gaining acceptance by the group” (p. 309). Also, “if groups are primarily concerned with finding a mutually acceptable decision, they may be predisposed to avoid unique information in their discussions and discount its relevance” (p. 309).
Procedural interventions. Assigning roles: if each member of a group is responsible for furnishing a particular sort of info, this should increase sharing and consideration of unique information. Role assignment increased amount of unique info discussed, but not by much (Stasser et al., 2000). Ranking candidates helped uncover unique info (Hollingshead, 1996). Not asking preference before discussion (Houlette et al., 2000) and framing the issue as a problem solving task (Stasser & Stewart, 1992) helped discover hidden profiles. Priming for consensus or critical thinking norms created 22% and 67% acceptance rate of the Best option, respectively. Counterfactual primes (scenarios read encouraging consideration of alternative possibilities) also helped uncover hidden profiles (Galinsky and Kray, 2002). Brauer et al. (2001) created dispersed and concentrated conditions; unique information shared equally between members, and unique information given mostly to only one member. The concentrated condition resulted in unique information being discussed more (the information was in relation to stereotypical/counterstereotypical information). This shows that “all hidden profiles are not created equally” (p. 311). This suggests that minorities armed with enough unique information can be strong changers (if we ignore the social costs of sharing unique information).
Mutual enhancement. In dyad discussions, sharing mutually shared common information increases impressions of competence and credibility, more than introducing unique information (Wittenbaum et al., 1999). Cognitive centrality: “people who know many of the same things that others know” have advantages; ability to validate others’ knowledge and their knowledge is validates by others (Kameda et al., 1997). This suggests strategic considerations; better to display behavior that ups one’s credibility.