Adjusting for Group Size Effects in Peer Nomination Data

Abstract

AbstractAdjusting nomination‐based sociometric and peer assessment scores for biases due to variations in group size has been a long‐standing concern for peer relations researchers. The techniques that have been typically used to make these adjustments (e.g., proportion and standardized scores) are known to have fundamental problems that limit their utility. This study introduces a regression‐based procedure that adjusts nomination‐based scores for variations in group size and compares it with the standardization and proportion procedures. Analyses were conducted on sociometric and peer assessment scores of 1594 fourth, fifth, and sixth graders from 63 classrooms. The advantages of the regression‐based procedure over standardization and proportion transformations are outlined. Implications for the accuracy and validity of nomination‐based measures and the research findings based on them are discussed.