Document Type


Degree Name

Master of Arts (MA)




Faculty of Science

First Advisor

Pamela Sadler

Advisor Role

Thesis Supervisor


This research extends prior knowledge of the statistical procedure of ipsatization, commonly utilized in interpersonal research to align data with theoretical expectations. The working hypotheses in prior studies have posited that a general factor, representing a response bias with no relevant substantive meaning, alters the data and interferes with analysis and interpretation unless removed by ipsatization. In the first of two studies, we initially investigated whether ipsatization removes important conceptual information from data when it removes a general factor. Three potential meanings of the general factor expected to occur in the Likert-scale version of the Social Behavior Inventory (SBI; Moskowitz, 1994) were modeled. When the resulting models did not adequately predict the data, the underlying structure of the data was analyzed with the discovery that a general factor does not exist for this version of the SBI. During study 2, this discovery was replicated in two larger datasets, leading to an investigation into whether ipsatization is still useful for a measure that does not possess a general factor. Despite a lack of a general factor to be removed, ipsatization did improve the structure and correlation patterns of SBI data with the resulting patterns matching those predicted by interpersonal theory. Thus, ipsatization can still be performed on this measure, which does not possess a general factor, thereby suggesting that the mechanism by which ipsatization improves data may not simply be the removal of a general factor as previously assumed. Several alternative mechanisms are discussed, with future research required to fully understand how ipsatization transforms the structure of data.

Convocation Year


Included in

Psychology Commons