Approval, disagreement, repeated discussion, and coordinated action transmit norms inside a network; similarity at one moment does not reveal direction.

Measure attitudes before and after entry into a network and record actual discussion or coordinated action. Change following contact supports influence more strongly than a one-time similarity among friends, which may reflect prior selection.

Friendship similarity combines influence and selection. People may adopt a group's norm after joining, but they also enter and leave networks partly because of shared values. Evidence becomes more persuasive when researchers observe attitudes before a network forms, measure the direction of discussion, and track whether change follows a peer's prior change rather than merely coinciding with it. Peer effects can concern behavior as well as opinion. A first-time voter may register because friends organize a group trip, even without changing party preference. That is mobilization through a network, not proof that the network persuaded the voter on policy. A longitudinal classroom study that measures the same students before and after they join new groups can test influence more credibly than a single friendship survey.

Video lesson: Political socialization | US government and civics | Khan Academy