Why Performance Information Use Requires a Managerial Identity: Evidence from the Field of Human Services

Abstract

Previous research suggests that civil servants can perceive their role differently and that these differing perceptions also influence their responses to managerial reform programs. Yet there is little research examining how different role perceptions influence the application of performance measurement. Using survey data from 742 human service organizations in Switzerland, the present study addresses this gap by investigating how professional and managerial role identities affect managers’ use of performance information. The results support the hypotheses that role identities indirectly influence the application of performance information through the effects on role conflict that may occur when managerial professionals measure the performance of their organizations. A lack of resources, stringent political control, and poor information quality are also found to be associated with role conflict and a limited use of performance information.

Keywords:

• performance measurement
• performance information use
• role conflict
• role identity

Notes

1 To test for nonresponse bias in our data, we compared the characteristics of respondents in the sample with the population parameters obtained from an administrative data set collected by the Swiss Federal Statistical Office (BFS, 2015). The three variables of age, gender, and job tenure were available for both the respondents and nonrespondents (though not for the work integration domain, which is why managers in this field were not considered in the nonresponse analysis). In our sample, 70% of the respondents were male, and 30% were female. This ratio is comparable to that in the general population (67% men and 33% women). The respondents’ average age was 54 years, which is only marginally higher than the national average (53 years). The respondents median job tenure fell between seven and nine years, which is identical with the population parameter. In sum, the comparison revealed no substantial differences between the respondents in our sample and the population, which makes us confident that our further analyses are not distorted by any nonresponse bias.

2 Before the hypothesized measurement model was tested, we conducted an explanatory factor analysis (EFA) in order to test whether managerial role identity and professional role identity are indeed two distinct constructs. We performed an oblique promax rotation because it seems plausible that the factors are correlated. The analysis showed two factors that correspond to PRI (eigenvalue =2.28, all item loadings above 0.59 and communalities above 0.40) and MRI (eigenvalue =1.29, all item loadings above 0.61 and communalities above 0.41). Overall, each item strongly loads on its expected construct and shows weak cross-loadings (the ladder is between −0.02 and 0.07). EFA thus demonstrates that the items used measure two distinct concepts.

3 In order to have an adequately large group size, we combined the facilities for the disabled and children and youth institutions in one group. Work integration institutions and facilities for drug addicts were also grouped together. We think that this approach is justified here because facilities for the disabled often provide services for minors, and substance abuse services sometimes include job placement services.

4 In contrast, some differences were found with regard to performance information use across service domains. Additional F-tests confirmed that the level of performance information use is significantly higher in nursing homes (M = 27.7) and significantly lower in children and youth institutions (M = 23.1) than in all other service domains. In the remaining domains, managers’ reliance on performance information (M = 25.3–26.4) does not significantly vary.

Additional information

Notes on contributors

Roger Pfiffner

Roger Pfiffner, Center of Competence for Public Management, University of Bern, Switzerland.