BackgroundThe Agency for Healthcare Research and Quality (AHRQ) Hospital Survey on Patient Safety Culture was designed to assess staff views on patient safety culture in hospital settings. The purpose of this study was to examine the multilevel psychometric properties of the survey.MethodsSurvey data from 331 U.S. hospitals with 2,267 hospital units and 50,513 respondents were analyzed to examine the psychometric properties of the survey's items and composites. Item factor loadings, intraclass correlations (ICCs), design effects, internal consistency reliabilities, and multilevel confirmatory factor analyses (MCFA) were examined as well as intercorrelations among the survey's composites.ResultsPsychometric analyses confirmed the multilevel nature of the data at the individual, unit and hospital levels of analysis. Results provided overall evidence supporting the 12 dimensions and 42 items included in the AHRQ Hospital Survey on Patient Safety Culture as having acceptable psychometric properties at all levels of analysis, with a few exceptions. The Staffing composite fell slightly below cutoffs in a number of areas, but is conceptually important given its impact on patient safety. In addition, one hospital-level model fit indicator for the Supervisor/Manager Expectations & Actions Promoting Patient Safety composite was low (CFI = .82), but all other psychometrics for this scale were good. Average dimension intercorrelations were moderate at .42 at the individual level, .50 at the unit level, and .56 at the hospital level.ConclusionsPsychometric analyses conducted on a very large database of hospitals provided overall support for the patient safety culture dimensions and items included in the AHRQ Hospital Survey on Patient Safety Culture. The survey's items and dimensions overall are psychometrically sound at the individual, unit, and hospital levels of analysis and can be used by researchers and hospitals interested in assessing patient safety culture. Further research is needed to study the criterion-related validity of the survey by analysing the relationship between patient safety culture and patient outcomes and studying how to improve patient safety culture.

Statistical issues associated with multilevel data are becoming increasingly important to organizational researchers. This paper concentrates on the issue of assessing the factor structure of a construct at aggregate levels of analysis. Specifically, we describe a recently developed procedure for performing multilevel confirmatory factor analysis (MCFA) [Muthen, B.O. (1990). Mean and covariance structure analysis of hierarchical data. Paper presented at the Psychometric Society, Princeton, NJ; Muthen, B.O. (1994). Multilevel covariance structure analysis. Sociological Methods and Research, 22,, and provide an illustrative example of its application to leadership data reflecting both the organizational and societal level of analysis. Overall, the results of our illustrative analysis support the existence of a valid societal-level leadership construct, and show the potential of this multilevel confirmatory factor analysis procedure for leadership research and the field of I/O psychology in general.

Our findings support the idea that a more positive patient safety culture is associated with fewer adverse events in hospitals. Further research is needed to determine the generalizability of these results to larger sets of hospitals and to examine the causal relationship between patient safety culture and clinical patient outcomes.

This article integrates theory from the cognitive tradition in negotiation with theory on culture and examines cultural influences on cognitive representations of conflict. The authors predicted that although there may be universal (etic) dimensions of conflict construals, there also may be culture-specific (emic) representations of conflict in the United States and Japan. Results of multidimensional scaling analyses of U.S. and Japanese conflict episodes supported this view. Japanese and Americans construed conflicts through a compromise versus win frame (R. L. Pinkley, 1990), providing evidence of a universal dimension of conflict construal. As the authors predicted, Japanese perceived conflicts to be more compromise-focused, as compared with Americans. There were also unique dimensions of construal among Americans and Japanese (infringements to self and giri violations, respectively), suggesting that identical conflict episodes are perceived differently across cultures.