During the experiment’s second phase, Sherif placed these same individuals together in groups of two/three and asked them to publicly announce their estimates after each trial.
All participants tended to gradually change their estimates to be more similar to the others.