Management scholars propose that every organization has an identifiable "culture" that influences performance in domains such as quality improvement implementation (Shortell, O'Brien et al. 1995) and patient-care quality and efficiency (Rondeau and Wagar 1998). Culture is, nonetheless, abstract, and in order to effectively address it, we must first accurately and reliably measure it. A survey instrument based on the Competing Values Framework (CVF) is widely used to assess culture and is presented as well-validated, reliable and generalizable (Shortell, O'Brien et al. 1995; Rondeau and Wagar 1998; Meterko, Mohr et al. 2004). However, recent research in the Department of Veterans Affairs (VA) finds problems with reliability and construct validity of some subscales among non-supervisor employees (Helfrich, Li et al. 2007). This response difference, or measurement non-equivalence, can introduce significant measurement error (Vandenberg 2002).
Evaluate the measurement equivalence and invariance (ME/I) of the CVF scale among VA employees. Specifically to (1) Determine if scale response is primarily a function of facility or individual respondent characteristics; and (2) Assess measurement equivalence / invariance among two key subgroups: among supervisory levels and between physicians and nurses.
We analyzed cross-sectional survey data from the 2006 VA All Employee Survey (AES). We conducted multivariate regression to assess culture scales as a function of facility and respondent characteristics, and structural equation modeling to test ME/I of the organizational culture scale across supervisory levels and between physicians and nurses.
There are significant differences in culture scores among employee subgroups. Three subscales, Entrepreneurial, Team and Rational Cultures, are significantly, positively correlated with supervisory responsibility. Gender, race, age, and tenure with the VA were also significant correlates of culture subscales in multivariate regression, with female, non-White race, older in age (>30 years) and with VA for a shorter time (<3 years) more likely to perceive their facility as having an emphasis on flexibility and creativity (Entrepreneurial Culture), teamwork and empowerment (Team Culture), and efficiency and measurable outcomes (Rational Culture). The sub-analysis on physicians and nurses found physicians more likely to consider their facility as having Teamwork, Rational and Hierarchical cultures.
Measurement equivalence / invariance findings were mixed. Goodness of fit indices suggest the 4 subscales of the CVF instrument are good representations of the data across subgroups (i.e., configural invariance). However, the assessment of factor loading invariance across supervisory levels was not supported.
Overall, we find statistically significant, and possibly operationally important, differences among subgroups in the CVF organizational culture scale.
Culture as measured by the Competing Value Framework varies significantly among employees at different supervisory levels, and other subgroups including by gender, race and clinical profession, and may not be an organizational level construct. Assuming that a culture exists at an organization level, assessing culture exclusively among senior leaders likely provides a distorted view of the overall culture, and pooling organizational culture data across supervisory levels and other subgroups may obscure important differences in the perceived culture among employees.
None at this time.
Organizational issues, Quality assurance, improvement