Font Size: a A A

Using aggregated micro-level data as measures of macro-level phenomena: The case of the NDNQI-RN satisfaction survey

Posted on:2007-10-31Degree:Ph.DType:Thesis
University:The University of KansasCandidate:Elliott, Carol Geanne ScottFull Text:PDF
GTID:2449390005476506Subject:Health Sciences
Abstract/Summary:
Job satisfaction is a major determinant of nurses' intent to stay in the profession. Measurement of nurses' overall perception of job satisfaction is crucial to the development of strategies, across all system levels , to enhance workforce stability, especially in times of critical nurse shortages. The study's conceptual framework posited that demographic homogeneity, professional socialization, and organizational context mediated nurses' development of shared perceptions regarding job satisfaction. The main objectives were to confirm that the Referent-shift consensus composition measurement model (Chan, 1998) supports the traditional and non-traditional psychometric indices when micro-level data are aggregated to infer macro-level measures; and demonstrate the isomorphic properties of job satisfaction across levels of analyses. Secondary data were used from the 2004 National Database of Nurse Quality Indicators RN Satisfaction Survey. Participants were 206 NDNQI hospitals with 55,516 RNs and 2900 units at the 50% response; and 66,160 RNs and 3,067 units at the 40% response rate. The typical respondent was baccalaureate prepared, white, females, 40 years old, licensed as a RN 5.73 years, and worked on the current unit for 4.16 years. Traditional modalities of aggregation revealed the 2004 NDNQI-RN Satisfaction Survey met criterion levels set for psychometric reliability and validity at 50% and 40% response rates. The study used Cronbach's alphas, interitem correlations, and an ANOVA-based ICC (2) to verify instrument reliability. Hypothesis testing verified construct validity using the ICC (1), t-test, and MANCOVA. These findings demonstrated that when data are aggregated, the NDNQI-RN-Satisfaction Survey is a reliable and valid measure of workgroup (unit, hospital) job satisfaction, and that the isomorphic properties of job satisfaction are maintained across levels. Shifting survey item referent from individual to group allowed data interpretation at multiple levels of analyses when using the Referent shift consensus model. The adapted structure of the NDNQI-RN Satisfaction Survey was found conceptually and psychometrically sound at the workgroup level.
Keywords/Search Tags:Satisfaction, Data, Using, Aggregated
Related items