# entrepreneurial culture measure equals .85. The market-oriented culture measure also

exhibited a high degree of reliability of .91. Finally, the hierarchy culture measure has a

reliability of .89. Therefore, all of the culture dimensions measures exhibited acceptable

reliabilities.

In addition, a multi-trait multi-method approach provides further evidence for the

convergent and discriminant validity of the dimensions of culture. Specifically, a

confirmatory factor analysis of a model in which both method effects (coworker vs.

focal) and culture effects (clan, entrepreneurial, market and hierarchical) suggests that its

fit to the data is high (Model 1: CFI=1.00; SRMR= .04). Comparing this model to one in

which the culture effects are removed shows significantly lower fit (Model 2: CFI=.68;

# SRMR=.12). The chi-square difference between model 1 and 2 is also highly significant

(^{2}= 89.39, 14df, p< .001). This result suggests that the culture dimensions exhibit

convergent validity. A comparison between model 1 and a model in which the traits are

perfectly correlated (model 3), on the other hand, suggests that the culture facets also

exhibit discriminant validity (^{2 }=53.05, 4 df, p< .001). Table 5 details the results.

2

df 5

^{2 c }-

df -

IFI 1.00

CFI 1.00

SRMR/ RMSEA .04/ .02

5.20

94.59

19

89.39^{*** }

14

.69

.68

.12/ .20

58.25

11

53.05^{*** }

4

.80

.81

.08/ .21

Table 5: Multi-trait Multi-method Approach to Culture

Structure Model 1: Freely correlated traits and freely correlated methods. Model 2: Freely correlated methods but no traits. Model 3: Perfectly correlated traits and freely correlated methods.

Note. The ^{2 }statistics for model 1 is not significant. The ^{2 }values for model 2 and 3 are significant at p< .001. IFI=incremental fit index; CFI=comparative fit index; RMSEA= root mean square of approximation. ^{c }The change in ^{2 }is based on comparisons between the Model 1 and the other two models. *** p< .001.

68