Peer ratings. Students in each classroom rated each other on likeability and popularity in a complete
round-robin fashion. Students rated how much they liked each classmate (except themselves) on a 7-
point Likert scale (1 = dislike very much, 7 = like very much). They also rated how popular they thought
each classmate was (except themselves) on a similar scale (1 = very unpopular, 7 = very popular). Data
were coded as missing if participants did not rate a peer or if they gave an impossible rating (e.g.,
marked two scale points or in between two scale points). Missing ratings (5.74%) were replaced by
the average group rating in the classroom (Kenny, 2007).