Meta-Analytic Evidence on Strategy Content and Public Service Performance

Strategic management practice abounds in public service organizations and it adoption is designed to boost public service performance. While strategic management includes content, planning and implementation, in this paper we focus on the venerable history of research on strategy content and its relationship with public service performance. Content is important because it is about the outlook of an organization. A range of strategy content models have been presented, however the strategic stances of prospector, defender and reactor developed by Miles and Snow (1978) and adapted to public service organizations by Boyne and Walker (2004) have received the most attention. This model is premised on the notion that prospectors (more innovative culture) outperform defenders (a culture based on core business), while reactors (whose culture is to await instruction) is worst. A number of integration studies have been published on strategy content (Boyne and Walker 2010; Walker 2013) that support the performance outcomes of different strategic stances and have implications for the practice of public service organizations.

However, these studies are dated and, importantly, used integration techniques that provided binary results on whether a particular strategic stance was effective. Prior work was not able to answer the question: How much does strategy content matter for public service performance, and under which conditions? In this paper we adopt rigorous meta-analytic and meta-regression techniques to identify size effects allowing us to answer this question. While our main focus is to understand if the impact of strategy content works better with particular dimensions of public service performance (e.g. effectiveness, responsiveness, outcomes), we also examine if (1) the goals of the public service organization, (2) governance or culture facets and (3) more technically if the way in which the research was designed influence the results.

Data for this meta-analysis were derived from a number of sources. First, Web of Science searches in the Public Administration category for English language journal articles used the following search terms, or derivatives thereof: strategy content, strategy stance, Miles and Snow, prospector, defender, analyser, and reactor produced a corpus of 542 articles. Second, we conducted citation searches in Web of Science on highly cited strategy content articles (Boyne and Walker 2004, 2010; Bryson et al. 2010; Poister et al. 2020; Walker 2013) and identified 539 citing articles of which 60 were potentially relevant. Third, we included the 25 articles analysed in Walker (2013). These three searches provided a list of 627 potentially relevant articles. Review of these articles resulted in 22 empirical articles (123 coefficients) suitable for meta-analysis and meta-regression on the independent variables of strategy content and the dependent variable public service performance.

When needed, these coefficients were all re-calculated into correlations and Fisher’s r-to-Z transformation was applied. Next, three meta-analyses were conducted, one for each strategic stance separately. Finally, three meta-regressions were constructed to see if effect sizes differ significantly based on conceptual, contextual and methodological moderators.