Evaluating the Invariance of the Factor Structure of the EPQ–R–N Among Adolescents

Abstract

We examined the factor structure of the Neuroticism scale of the Eysenck Personality Questionnaire (EPQ–R–N; S. B. G. Eysenck, Eysenck & Barrett, 1985) and its factor invariance across sex and racial/ethnic groups in a sample of 1,979 adolescents. Using confirmatory factor analyses, we compared a hierarchical model to previous models of the EPQ–R–N and to single-factor and 3-factor structures. The hierarchical factor structure in which a general factor coexists with 3 group factors (depression, social concerns, and worry) was superior to alternative models. The general factor accounted for more than 60% of the variance in EPQ–R–N total scores and was invariant across sex and ethnicity. The 3 group factors varied across ethnicity and sex. We discuss the implications of these findings for conceptualization and assessment of neuroticism using the EPQ–R–N.

Acknowledgment

The research reported in and the preparation of this article were facilitated by National Institute of Mental Health Grants R01 MH65651 and R01 MH65652 and by the Patricia M. Nielsen Research Chair of the Family Institute at Northwestern University.

Notes

The hierarchical model (McDonald, 1999) assumes that observed scores may be decomposed into four parts:

where g is a general factor (that is, a factor that is common to all k scale indicators), c is a k × 1 vector of unstandardized general factor loadings, f is a r × 1 vector of group factors (that is, factors that are common to some but not all k indicators) with r < k, A is the k × r matrix of unstandardized group factor loadings, s is the k × 1 vector of specific factors that are unique to each item, d is the k × k diagonal matrix of unstandardized factor loadings on the item specific factors, and e is the k × 1 vector of random error scores. The model expressed in (1) assumes also that all factors (g, f, and s) are uncorrelated with each other and with e, all the errors (e) are uncorrelated with each other and without loss of generality, and that the variance of each of the common factors (g and those in f) equals 1. In addition, each of the group factors (in f) are required to have nonzero loadings on at least three indicators to identify the model in the absence of additional constraints.

That is, in this broader hierarchy, there are three levels of breadth: narrow group factors such as the three anxiety sensitivity group factors that are common to a small number of items, broad group factors such as the general anxiety sensitivity factor that are common to the items comprising some but not all of the narrow group factors, and a negative emotionality general factor that is common to all items.

An advantage of using the hierarchical factor model in (1) is that it permits the derivation of a coefficient giving the proportion of variance in the scale scores accounted for by a general factor—omega_hierarchical (ω _h ). More specifically, ω _h is defined as ω _h = 1′cc′1/1′S1, where S is the observed variance/covariance matrix among the k indicators comprising the scale (Zinbarg, Revelle, Yovel, & Li, 2005; Zinbarg, Yovel, Revelle, & McDonald, 2007; for an algebraically equivalent expression, see McDonald, 1999, Equation 6.2.1). In contrast, a higher order model must first be transformed into a hierarchical model via a Schmid and Leiman (1957) transformation before it can be used to estimate ω _h .

^aThe hierarchical model contained three group factors and a general factor in all comparisons.

^bThese modified models were constructed to allow the comparison of a hierarchical and an orthogonal model.

* p ≤ .01.

These analyses may be obtained from N. Mor.

^aThe hierarchical model contained three group factors and a general factor in all comparisons.

^bThese modified models were constructed to allow the comparison of a hierarchical and an orthogonal model.

* p ≤ .01.

^aMales, N = 445; females, N = 596.

^bAfrican Americans, N = 183; Whites, N = 619; Hispanics, N = 240.