26 May 2017 | Should Public Management be an Experimental Discipline?, By Oliver JamesAndrew Osorio, · Categories: monthly_posts
There has been substantial growth in the use of experiments in public management over recent years. The substantive contributions are becoming broad and deep, and span a range of research questions about core topic areas. These observations were part of the motivation for Sebastian Jilke, Gregg Van Ryzin and I to put together an edited book about this experimental turn. We asked a set of leading experimental researchers of public management to contribute and were delighted when all agreed to participate. The result, Experiments in Public Management: Challenges and Contributions (Cambridge University Press) is about to be published.
The experimental turn is not a ‘mere’ application of a generic social science approach towards experimentation, or a passing fashion. Whilst the approach to causal inference and core methods of intervention, random allocation, and comparison of outcomes are shared with social science experimentation, their use in public management is distinctive. Experiments should take their place alongside other methods in public management. For a fruitful interchange between researchers using different methods, non-experimentalists will often need to know more about experimental methods, and experimentalists will need to understand and be open to methods and evidence gained from other approaches.
Contemporary public management experiments are particularly influenced by their use in health sciences, economics, political science and psychology. Experiments are sometimes criticised for being too narrow, missing out on ‘big’ questions, for example, about macro-structures or major issues of importance to public management. However, such a risk is mitigated by using a variety of methods and by breaking bigger issues up into smaller ones more amenable to empirical analysis. In any case, it is difficult to argue that the potential for experimentation is anywhere near being exhausted given their relative scarcity in public management to date. Field and survey experiments in particular have great potential for joint working with public organisations, and can help boost the external validity of findings from experiments in the laboratory to real world contexts.
Individual experiments sit within programmes of research using multiple experiments to tease out different causal mechanisms and explore mediation or moderation of treatment effects. Replication of experiments is an important part of verifying findings and thinking about contextual variation. In our own work we have sought particularly to develop such programmes. For example, experiments have shown that priming citizens to think about their need for a public service rather than encouraging a political mind-set can reduce party political based motivated reasoning in assessing performance information as evidence about service quality. Related experiments on public services into user choice and choice overload have shown that increasing provider choice sometimes reduces users’ likelihood of stating that they would switch away from a poorly performing provider. As part of a growing community of scholars using experiments we hope to contribute to advancing understanding of experimental methods appropriate to our discipline and to promote more and better experiments on an increasing range of research topics.
Oliver James, University of Exeter, UK.