The main purpose of this study is to assess the factor structure and the measurement invariance of the Mini-International Personality Item Pool (Mini-IPIP; Donnellan, Oswald, Baird, & Lucas, 2006). The Mini-IPIP is a brief instrument evaluating personality traits according to the Big Five models. Two samples were collected comprising nearly 800 participants. Confirmatory factor analyzes revealed a five-factor solution consistent with the Big Five model. Measurement invariance analyses showed that the Mini-IPIP was reasonably invariant across samples, genders and age groups. Overall, results pointed to a satisfactory factorial structure and an adequate invariance of the measure.
The Big Five is the dominant model used to study normal personality across the life span in trait psychology. This model has fuelled a large body of research exploring the validity and relevance of its five-factors: extraversion, agreeableness, conscientiousness, neuroticism, and intellect/openness. It is generally accepted that personality traits are relatively stable, albeit changes are observed over the life course (Marsh et al., in press and Roberts et al., 2006). The cross-cultural stability of the FFM has also been the focus of many empirical investigations which generally replicated its factor structure across cultural groups (e.g., McCrae & Allik, 2002).
The Mini International Personality Item Pool (Mini-IPIP) was developed given the widespread interest in the Big Five taking into consideration critical assessment issues, such as questionnaire length (Donnellan et al., 2006). Although the Mini-IPIP has been found to possess promising psychometric properties, such as acceptable reliability and highly similar correlations with other Big Five measures and personality constructs than longer IPIP measures, further investigation of its psychometric properties seems warranted. Importantly, the factorial structure of the Mini-IPIP has not been optimal in previous studies, showing crossloadings and elevated correlations between factors that should theoretically be orthogonal (Cooper et al., 2010 and Donnellan et al., 2006). The issue of factor structure is central to the Big Five approach to personality. While the approach historically rested on factor analysis for the delineation of its main dimensions and the identification of their constituents, poor factorial structure and high correlations among factors have been seen as major shortcomings of the Big Five approach. In this respect, the findings regarding the Mini-IPIP are consistent with previous empirical investigations of this issue with other Big Five measures (e.g., Church and Burke, 1994, Marsh et al., in press and McCrae et al., 1996). Based on confirmatory factor analysis (CFA), these earlier studies usually lead to poor model fit and creative model respecification. However, a main argument leading to the development of the Mini-IPIP was that it would help to overcome these well-documented shortcomings of longer Big-Five personality measures (Donnellan et al., 2006), something that has yet to be empirically demonstrated.
In addition, the measurement invariance of the Mini-IPIP across meaningful subgroups of participants has yet to be investigated. In addition to representing a powerful test of the generalizability of a measurement model across samples and subpopulations, measurement invariance, also represents an important pre-requisite to meaningful and un-biased between-group comparisons. A measure is invariant when it measures the same latent trait across groups in the same manner and with the same precision (e.g. Millsap, 2011). More precisely, the invariance of factor loadings (i.e., weak invariance) tests whether the instrument measures the same construct across subgroups and is a prerequisite to comparisons of latent variances or relations among latent constructs. The invariance of the items’ thresholds (i.e. strong invariance) tests whether participants from different subgroups with similar levels on the construct present comparable scores on the items forming the construct and is a prerequisite to latent mean comparisons. Finally, the invariance of the items’ uniquenesses (i.e. strict invariance) tests whether the constructs are assessed with similar levels of measurement errors in the various subgroups and is a prerequisite to any group comparison based on manifest (no-latent) scores.
An interesting test of the construct validity of a scale that can easily be combined with tests of measurement invariance had to do with the investigation of potential latent means differences across subgroups in order to verify whether these differences replicate those from previous research investigating the same constructs. For instance, gender and age known to be associated with clear differences in mean-levels of FFM personality traits (Costa et al., 2001, Donnellan and Lucas, 2008, Feingold, 1994, Lucas and Donnellan, 2009, Roberts et al., 2006 and Terracciano et al., 2005). Investigating gender differences in 26 cultures, Costa and colleagues (2001) found that women scored higher on neuroticism, agreeableness, warmth (a facet of extraversion) and openness to feelings, while men were higher on assertiveness (closest to the Mini-IPIP extraversion factor) and openness to ideas (close to the Mini-IPIP intellect/imagination factor). In regard to age differences, results generally show that neuroticism tends to decline with age while conscientiousness increases. Interestingly, these potential latent means differences have yet to be investigated with the Mini-IPIP.
The objectives of the present study are to explore the factorial structure of the Mini-IPIP and to assess its measurement and latent mean invariance according to sample, gender, and age.
