The consistency of ratings by peers

Abstract

In the validation and improvement of rating scales used to study ratings by peers, how important are contextual factors? Although the data is for air force application, Ss were 144 paid college students. "… in each of the two replications of the design, six groups of six men were observed by 36 observers… . the Os and Ss reversed roles after two sets of four sessions each… . The approximately 100,000 ratings were punched on IBM cards and most of the computational work was carried out on IBM machines." A man's rating tended to vary from setting to setting and according to rating instructions and trait definitions. (PsycInfo Database Record (c) 2025 APA, all rights reserved)