Normative Social Influence

What it is

In Asch's paradigm a participant identifies which of three comparison lines matches a standard, a task whose simplicity and lack of ambiguity is the point, seated alongside confederates who give unanimous incorrect answers on certain trials and answering last, and Axel Franzén and Sebastian Mäder ran it with five confederates to each naive subject and two hundred and ten participants, reporting an error rate of thirty-three per cent, which replicates the original.

Rod Bond's argument is about mechanism, challenging the assumption of a single fixed group-size function in formal models of social influence, arguing that normative influence should dominate with public face-to-face responses and informational influence with private responses and indirect communication, and finding that the size and conformity relationship differs accordingly and by whether a study used the Asch or the Crutchfield paradigm; the weighted average effect size of 0.89 usually attributed to him reaches this record through Carla Capuano and Peggy Chekroun's 2024 systematic review of his paper rather than from the paper itself, and is attributed that way. That review covers conformity research since 2004, identifying forty-eight relevant articles, and its verdict is stated in its own words, that recent replications and meta-analyses underscore the robustness of this effect, obviating the necessity to continually assess its existence as it persistently manifests.

In effect

Conformity is one of the few mid-century social psychology findings that came through the replication decade intact, and it came through in direct replications at larger samples, with the 2024 review reporting a conformity rate of 59.2 per cent in one 2019 replication against Asch's seventy-five, 82.67 per cent of participants conforming at least once in a 2020 study, and Franzén and Mäder's thirty-three per cent error rate. The incentive result is the most informative single number, because money for being right cut errors by eight percentage points and left a quarter of responses still wrong in the direction of the group, which is a better description of how a norm works than either a story about rational error or one about helpless compliance. Two further results from the same replication are worth keeping, since on political statements the conformity rate rose to thirty-eight per cent, above the perceptual task, and besides openness none of the investigated personality traits, including intelligence, self-esteem, need for social approval and the Big Five, were convincingly related to susceptibility to group pressure, which points the explanation at the situation rather than the person.

What it does not say

It does not say conformity has weakened over time, because Rod Bond and Peter B. Smith's 1996 meta-analysis of a hundred and thirty-three studies is widely reported in secondary sources to show a decline in effect sizes while its abstract could not be retrieved at any source for this record and every reachable source for the decline turned out to be a study aid rather than a paper, so that is unverified and is not asserted; what can be said at first hand is that the recent direct replications find rates close to Asch's. Nothing here is attributed to Asch's own 1956 monograph, which was not read, so the paradigm is described from the 2024 systematic review and the 2023 replication, which were. The 0.89 figure is not Bond's own stated result as read here, since it reaches this record through the 2024 review's report of his paper. And the laboratory evidence does not transfer directly to applied settings, because the lines task has a correct answer and a group physically present, which is the opposite of a person alone with a sign or a default, and the field-experiment evidence for norm messages is much smaller and partly unreplicated, so the two literatures support each other in direction and not in magnitude.

---