This article suggests a graphic technique that uses P-P plots to show the extent to which two groups differ on two variables. It can be used even if the variables are measured in completely different, noncomparable units. The comparison is symmetric with respect to the variables and the groups. It reflects the differences between the groups over…

Presents a model-based approach to rater reliability for essays read by multiple raters. The approach is motivated by generalizability theory, and variation of rater severity and rater inconsistency is considered in the presence of between-examinee variations. Illustrates methods with data from standardized educational tests. (Author/SLD)