Editing Asch conformity experiments (section)

=== Partner dissent and accuracy ===
:Experiments were also designed to determine if the partner effect on subject conformity was due to the partner's dissent from the majority or their accuracy in answering questions.<ref name="Asch1955" /><ref name="Arsch1956" /> In one experiment, Asch identified two classes of dissenter: "extremist" (under this condition, dissenters always chose the worst of the comparison lines and the majority chose the line closest to the standard in length) and "compromising" (dissenter: closest to standard; majority: worst comparison line). In compromising dissenter trials, subject conformity decreased overall and when they did conform, they conformed to the dissenter, not the majority. Compromising dissenters were seen to control the "choice of errors". In trials with an extremist dissenter, subject conformity decreased dramatically with only 9% of respondents continuing to answer with the majority. Therefore, partner dissent was found to increase independence, moderating errors (conformity). 
;Withdrawal of a partner
:Asch also examined whether the removal of a true partner partway through the experiment influenced participants' level of conformity.<ref name="Asch1951" /><ref name="Asch1955" /> He found low levels of conformity during the first half of the experiment. However, halfway through the experiment the partner rejoined the majority, answering in lockstep with the group. When their partner switched, the subject conformity rose to levels consistent to if they had never had a partner at all. Asch classified this finding as a "desertion" effect. In a variant of this study, the partner left the experiment halfway-through altogether (an excuse was provided for their departure). Under these conditions, the partner's influence lingered through the second half of the experiment; the subject's conformity to the group increased after the partner's departure, but not as drastically if the partner was perceived as having switched sides.<ref name="Asch1955" />
;Majority size
:Asch also examined whether decreasing or increasing the majority size had an influence on participants' level of conformity.<ref name="Asch1951" /><ref name="Asch1952b" /><ref name="Asch1955" /> When paired with a single individual who opposed their answers, the subject retained high levels of independence in their answers. Increasing the opposing group to two or three persons increased conformity substantially. Increases beyond three persons (e.g., four, five, six, etc.) did not further-increase conformity.
;Written responses
:Asch also varied the method of participants' responding in studies where actors verbalized their responses aloud but the "real" participant responded in writing at the end of each trial. Conformity significantly decreased when shifting from public to written responses.<ref name="Arsch1956" />