Implications of item wording for hardiness structure, relation with neuroticism, and stress buffering.

Abstract Previous literature suggests three criticisms of hardiness research: (1) redundancy between hardiness and neuroticism, (2) inappropriate tests of hardiness hypotheses, and (3) a lack of support for hypothesized effects of hardiness. Psychometric concerns underlie these criticisms. Specifically, current hardiness measures rely extensively on negatively worded items, which may lead to redundancy with neuroticism, ambiguous factor structures, and attenuated criterion-related validity coefficients. Alternately, negative hardiness items may measure separate, but hardiness-related cognitive processes. In Study 1 ( N = 254), a series of confirmatory factor analyses suggested that positively worded hardiness items are distinct from neuroticism and hierarchically structured with three subdimensions nested under a second-order factor. However, the analyses were equivocal concerning whether negatively worded items measure hardiness or neuroticism. In Study 2 ( N =426), we examined the implications of the structural findings for criterion-related validity. A replication of Study 1 suggested support for a model in which negatively worded hardiness items are redundant with neuroticism. Regression analyses indicated that hardiness subdimensions predict health and academic performance criteria above and beyond the global dimension and that positive and negatively worded hardiness items predict different health outcomes, even after controlling for a measure of neuroticism. Weak support was obtained for hardiness as a buffer of the stress–outcome relationship. Funk's (1992) suggested agenda for hardiness research is used as a framework to discuss these findings.