Multiple streams theory in Sweden: an error III

One can only applaud researchers embarking on the task of applying political science theories to policy research in health promotion. This is a much-needed effort, and not just for theories of the policy process. In the concluding chapter of their health promotion textbook, Keleher et al. observe that public health and health promotion continue to be a largely a-theoretical enterprise (Keleher et al., 2007). And as we found in the health promotion policy development and analysis literature, even when theories are applied they are often at odds with the more recent advances in the social and political science or being applied the wrong way (Breton and de Leeuw, forthcoming). We have profound concerns of a similar nature over the Gulbrandsson and Fossum article recently published in this journal (Gulbrandsson and Fossum, 2009). Praise is due to the attempt to apply a solid theory of the policy process to public health environments in Sweden. But there are two main problems with the study undertaken: (1) the lack of clear integration of the set of theories called upon by the authors to, it appears, supplement John Kingdon's multiple streams theory (Kingdon, 1984) and (2) the applicability of Kingdon's theory to policy-making at the local public health government level in Sweden. We will elaborate on the latter. The authors set out to evaluate whether the concepts of policy window and policy entrepreneur are applicable to the local public health arena. Although an interesting question, there is a more pressing one that needs to be answered first: Is Kingdon's theory relevant to the cases analysed in this paper? Kingdon built his theory upon extensive field work at the US federal government level. All cases investigated were about contentious public problems in the realms of transportation and health (i.e. health care services). The road to policy change in these policy domains was all but a smooth one. The number of people with a stake in the outcome of the policy debates was at times overwhelming, as were their diverging views on what the actual problems were and on the types of policy instruments (i.e. solutions) to address them. In most, if not all cases, there were stakeholders resisting every plan to significantly change the policies, seeing their interests or the one of their constituencies better served by the status quo. In fact, and beyond Kingdon's theoretical notions alone, the descriptions and analyses of policies in the political science literature entail conflicts, opposition, resistance and struggles (Sabatier, 2007). The study of policy development, not surprisingly, has some roots in studies of war and warfare (Lasswell, 1936). None of the cases analysed suggest that controversy existed, or significant debates took place prior to policy adoption. Were the ‘policies’ analysed merely functional components of broader institutional choices driving the Swedish policy ‘world view’ (Surel, 2000)? In this case, the ‘policies’ would only stand as mere non-contentious adjustments to already adopted policies; policies that both define the kind of problems amenable to public scrutiny and the realm of the possible/acceptable measures to tackle them. Hence, the apparent ease with which participants in the study succeeded in identifying cases for the research team to analyse. Being vehement proponents of the Theory-Based Evaluation approach (Birckmayer and Weiss, 2000), we believe that theories (or a validated conceptual logic) must be matched to real problems in evaluation efforts. This is to enable evaluators not just to show whether something has worked, but in particular why. We question whether Gulbrandsson and Fossum's contribution is relevant, in this perspective. Perhaps, Swedish traditions in policy-making do not centre on the resolution of conflict, but focus on processes of consensus development (Heclo, 1974). If this is the case, the application of the multiple streams theory to the nine cases studies is a clear demonstration of an ‘Error of the Third Kind’ (‘solving the wrong problem’, cf. Mitroff and Featheringham, 1974) and should be scrutinized as much as other studies that would make a type I (‘false-negative’) or type II (‘false-positive’) error. It is critical that health promotion policy researchers boost their efforts to disentangle the policy change process leading to significant policy changes. Many health, social and environmental policies fail to promote population health or, if they do, leave significant segments of the population behind (CSDH, 2008). If we fail to go beyond analysing what appear to be mere incremental policy adjustments, we are likely to be irrelevant to the health promotion professionals and policy activists striving for real policy shifts. In conclusion, it is obvious that much is yet to be learned about the development of healthy public policy and health-in-all policies. The application of trialled and tested theories from social and political science is the only way to build a credible body of knowledge and sound evidence base for further development of such policies. But this will have to be done rigorously and conscientiously, matching problems to appropriate theories. Our understanding of the efficacy of health promotion advocacy for better policies would gain from the critical lessons delivered by such theories, for instance in the realm of coalition building (Breton et al., 2008).