It was discovered that when participants were individually tested their estimates on how far the light moved varied considerably (e.g. from 20cm to 80cm).
The participants were then tested in groups of three. Sherif manipulated the composition of the group by putting together two people whose estimate of the light movement when alone was very similar, and one person whose estimate was very different. Each person in the group had to say aloud how far they thought the light had moved.
Results: Sherif found that over numerous estimates (trials) of the movement of light, the group converged to a common estimate. The person whose estimate of movement was greatly different to the other two in the group conformed to the view of the other two.
Sherif said that this showed that people would always tend to conform. Rather than make individual judgments they tend to come to a group agreement.
Conclusion: The results show that when in an ambiguous situation (such as the autokinetic effect), a person will look to others (who know more / better) for guidance (i.e. adopt the group norm). They want to do the right thing, but may lack the appropriate information. Observing others can provide this information. This is known as informational conformity.