Bifactor Model History: Unidimensionality Assumption and Group Factors. Does It Matter?

• In Event: AERA Roundtable Session Sunday 11:40 am Four Seasons Ballroom 2-3
  In Roundtable Session: Measurement Issues (Table 9)

Abstract

An increasing number of scientific papers are being published using confirmatory Bifactor model (Bm) to assess/confirm the factor structure. As mentioned by Reise (2012), some applications of Bm may be limited due to its restrictive assumptions. The objective of this proposal is twofold: (1) to recall the assumptions of the original Bm (Holzinger, 1935; Spearman, 1927) and (2) to verify how these assumptions (unidimensionality, group factors) are assessed for the confirmatory Bm. A total of 111 papers from six high-ranking scientific journals were analyzed (2010-2024). The results indicate that the Bm’s unidimensionality assumption is rarely assessed, and group factors identification method does not rely on the coefficient of belonging and the residual correlations, leading to misinterpretation of the factor structure.

Authors

• Eric Frenette, Universite Laval

• Nadine Talbot, Université du Québec à Trois-Rivières

• Sebastien Beland, University of Montréal

• Marie-Helene Hebert, TELUQ University