This blog post also appears as part of the Bristol Social Marketing Centre's 'Spotlight on Behaviour Change' e-newsletter series available here. To sign up for these newsletters, written by a range of BSMC team members, please email email@example.com.
Apparently the rates of violent crime and murder have been known to jump when ice cream sales do, and the global increase in Facebook use correlates with the Greek debt crisis. These are obvious examples of the fallibility of ‘evidence’, but the problem of ‘evidence’ for behaviour change activity is a little more subtle.
Dr Harry Rutter from Public Health England explained the problem in terms of his rather Edward Monkton-esque ‘Dangerous Olive of Evidence’ at a recent ESRC behaviour change seminar. The olive flesh is the vast pool of ‘evidence’ necessary for policy makers and analysts to understand the vast context of the complex wicked problems they are trying to tackle. The tiny hard pip is the evidence that is currently being collected. The Dangerous Olive metaphor suggests that we are so often not fully in the picture when it comes to the causes and potential solutions to complex problems like obesity or climate change. (Dr Rutter worked on the Foresight Obesity report and has a genuine insight into the complexity of context).
A more traditional way of conceptualising the problem is with the elephant metaphor. If blindfolded people examine a single discrete part of an elephant, they are bound to disagree on what really constitutes the animal’s elephantness.
Moving (temporarily) away from olives and elephants, we can examine the problem with the evidence for behaviour change; one that has two interconnected parts. The first is that we are simply not collecting the full range of evidence we need on which to base our behaviour change activity. In fact the very phrase ‘behaviour change’, which has become so engrained in our lexicon, is tantamount to this. The implication from this phrase is that changing individual’s behaviour is the route to success in tackling wicked problems. This does not take into account the huge influence of structural forces - such as the activities of big business - and this is in part because these forces are not those about which evidence is routinely collected. For example we have ample evidence that eating more vegetables decreases obesity, but very little about the huge impact that reconfiguring the policies for fast food marketing to children might have (e.g. McDonald’s sponsorship of school resources), so it appears that encouraging more vegetable consumption is the better option.
Partly (and carefully avoiding discussions about corruption, commercial lobbying and political bias), the evidence problem described above is the result of what research is possible to do and what is less easy. Indeed, a huge and highly valuable body of evidence has been built up on the efficacy of various individual-level intervention types (see the BCT taxonomy work). This evidence is of unequivocal quality, but its abundance naturally implies that the individualist behaviour change strategies with which it is concerned enjoy a level of significance which may not reflect the reality of a complex context. The Dangerous Olive’s fleshy area remains largely unexplored.
The second part to the problem relates to the discriminations of the leading consumers of evidence; policy makers. There exists an undeniable bent towards evidence that feels ‘truly scientific’; preferably quantitative and with some kind of randomised and controlled element. The insistence on quality and rigour in research is commendable, but such narrow requirements have contributed to the problem discussed above – the bias towards collecting evidence which is possible to collect under such conditions. Crucially important evidence lying within the Dangerous Olive’s fleshy areas, for example relating to the influence of prodigious and unscrupulous commercial activity on cultural discourses around food, eating or active travel, may not be researchable using quantitative, positivistic approaches reminiscent of a comforting biomedical paradigm. To switch metaphors, ethnography, discourse analysis or case study research may not ‘feel’ like real science to some policy makers, but might be required alongside traditional approaches if we are ever to get a sense of what the whole elephant is really like.
Of course there are responsibilities here for scientists using non-positivist methods to achieve acceptable rigour as much as there are responsibilities for policy makers to embrace a much needed paradigm shift and encourage the submission of a range of types of evidence for the underpinning of policy decisions. As such, the need for improved communication and transparency between academic researchers and policy makers is a prerequisite for this evolution.