Organizational culture

1
Organizational culture

• Factor analysis of a Competing Values Framework
  instrument

Christian D. Helfrich, MPH, PhD Implementation
Research Coordinator Ischemic Heart Disease
Quality Improvement Research Initiative
2
Background

• Organizational culture
• Within organizations and groups, individuals
  share common beliefs, assumptions and values that
  create powerful behavioral norms (James et al.
  1990).
• These norms are taught to new members and guide
  members actions and interactions (Barker 1993).

3
Background

• Competing values framework (CVF)
• A dominant organizational culture model
• Linked to patient satisfaction (Meterko et al
  2004), physician and nurse turnover (Mohr et al
  2005) and adoption of QI activities (Shortell et
  al 1995)
• All VHA employees surveyed in 2004

4
Competing Values Framework
5
Specific Aims

• To validate an instrument based on the Competing
  Values Framework among non-managers
• Conduct item analysis to determine subscale
  reliability and assess the convergent / divergent
  qualities of the subscales
• Conduct exploratory factor analysis to determine
  if emergent factor solutions (i.e., emergent
  subscales) match conventional subscales
• Conduct confirmatory factor analysis to compare
  model fit between emergent and conventional
  subscales

6
Methods

• Design
• Cross-sectional, observational study
• Analysis
• Item analysis
• Cronbachs alpha
• Item-rest correlation
• Item-to-scale correlation
• Factor analysis
• Exploratory factor analysis
• Confirmatory factor analysis

7
Data

• 2004 All Employee Survey (AES)
• Employee-level survey
• 14 items based on CVF subscales
• 52 response rate among 200k employees (n
  102,118)
• Sample non-supervisory employees
• n 71,776

8
Questions so far?
9
Results - Respondents
10
Results Item analysis

• Item analysis see Table 1

11
Culture scatter plots, facility-level
12
Results - EFA

• Exploratory factor analysis see Table 2

13
Results - CFA

• Confirmatory factor analysis see Table 3

14
Discussion - External validity

• The CVF as a model, or the CVF instrument, may
  not generalize
• To VHA
• To non-managers
• Or to the combination of both
• Need for measurement equivalence / invariance
  analysis (ME/I)
• Among supervisory levels
• Over time

15
Discussion Internal Validity

• Measurement error in original instrument
• Modifications made to the survey used in VHA
• Wording of individual items was adapted
  primarily four items
• VHA instrument had two fewer items than the
  original 16-item scale
• Used normative (Likert) scales versus ipsative
  scales
• Terms such as bureaucratic and innovative
  likely carry normative connotations for lay
  readers
• Most original CVF items consist of two
  declarative statements
• Items within subscale were organized across four
  organizational domains institutional
  characteristics, institutional leader,
  institutional glue and institutional emphases

16
Discussion Construct Validity

• Possible poor construct validity for the four CVF
  culture types
• Three of four subscales may reflect single
  underlying factor
• One subscale may reflect multiple underlying
  factors

17
Conclusions

• CVF instrument did not perform as predicted in a
  population of non-managers from VHA
• May suggest caution in drawing inferences based
  on aggregated CVF scales when applied to
  populations where they have not been validated,
  such as non-managers
• Importance of validating organizational culture
  instruments in each new context they are used

18
Questions