ABSTRACT
Despite the existence of multiple measurement instruments, measuring public service motivation still generates controversy. The most common cited measurement issues are the application of the measurement scale outside the United States, the dimensionality of public service motivation and the length of the original measurement scale. This article discusses these issues and evaluates a measurement instrument developed in a non-US environment, based upon empirical data. Although differences in items are noted, the factorial structure of the original measurement instrument remains in place. Nevertheless, there is some evidence that an additional dimension, “democratic governance” could supplement the other existing measurement instruments.
ACKNOWLEDGEMENTS
This research project was funded by the Research Foundation–Flanders. An earlier version of this article was presented at the 2006 European Group of Public Administration Annual Conference. The author wishes to thank Dirk Heerwegh, David Coursey, Annie Hondeghem, and James L. Perry, as well as two anonymous reviewers, for their constructive and helpful comments on previous versions of this article.
Notes
*N.K.: Not known.
(I) I scores are inverted.
a For PSM IV, the categories “Public Interest” and “Compassion” are collapsed into one category.
(I) Scores are reversed.
*t-values are statistically significant at p < .01.
Alternatively, one could resort to a two-step approach where the first step is an exploratory factor analysis (EFA) and the second is a CFA, validating the initial model. However, although in general an EFA approach is used in the early stages of scale development (Kelloway Citation1995) and CFA requires a clear hypothesis on the data structure, choosing between both is “like arguing about the best flavour of ice (Schriesheim in Hurley et al. Citation1997, 669).”
Two respondents were lost in the process of the random splitting of the sample. Therefore only 3504 respondents were retained.
In each output, LISREL suggest modifications, based on the scores of the modification indices. If these modification indices suggested to add paths (and thus loading an item on two factors), these items were deleted.