A comparison of incident-based and scale measures of work stressors STEVE M. JEX and GARY A. ADAMS Department of Psychology, University of Wisconsin Oshkosh, 800 Algoma Boulevard, Oshkosh, WI 54901, USA TINA C. ELACQUA and DAVID J. LUX Central Michigan University, Michigan, USA Keywords: Incident-based measures; Qualitative measures; W o r k stressors; Self-report scales. Occupational stress researchers have often questioned whether scale measures of work-related stressors accurately reflect the day-to-day experiences of employees. As an alternative to such measures, some have suggested the use of qualitative measures. Unfortunately, however, there have been no direct comp'arisons of these two measurement methods. In the present study, female employees (n = 151) were asked to complete scale measures of three common work stressors, strains, and self-esteem. Stressors were also measured using a qualitative methodology based on critical incidents. Results indicated good convergence between stressors measured by these two methods, although there was some indication that order effects may have led to method bias in the qualitative procedure. Both sets of measures explained a significant amount of variance in strains, although the scale measures explained more. Only the stressors measured by scales were related to selfssteem, suggesting that these may be more influenced by dispositions than qualitative measures. Implications of these findings for occupational stress research are discussed. 1. Introduction Over the years, occupational stress researchers have expressed concern over the construct validity of self-report scales, since their use has been the most popular method of measuring both stressful job conditions (stressors) and employees' reactions to these conditions (strains; Brief et al. 1988, Jex and Beehr 1991, Jex et al. 1993, Spector 1995, Spector et al. 1988). Using self-report scales to measure psychological strains (e.g. anxiety, depression, job dissatisfaction) would appear to be reasonable, since these are essentially subjective psychological states. Furthermore, in the case of dissatisfaction and anxiety, there is evidence that self-report scales reflect employees' qualitative experiences (Jick 1979, Taber 1991). Using self-report scales to measure physical strain is somewhat more controversial, however, since the construct validity of such measures has been questioned (Costa and McCrae 1987, Frese 1985, Jex and Beehr 1991). Compared to strains, using self-report scales to measures stressors is considerably more problematic for three reasons. First, research has shown that self-reported assessments of the work environment tend to converge poorly with other data sources (Cortina el al. 1992, Spector and Jex 1991). Since stressors are essentially aversive characteristics of the work environment, this lack of convergence suggests that self-reports of stressors are at best approximate indicators of objective job conditions (Spector 1992, 1995). 02674373197 91200 @I 1997 Taylor & Francis Ltd