How can and do empirical studies influence drug policies? Narratives and complexity in the use of evidence in policy making

Introduction

Some recent contributions to debates on drug policy and the use of evidence (e.g. Henderson, 2012; Nutt, 2012) have assumed that drug policy could be improved if politicians paid more attention to scientific evidence. While not disagreeing with the broad thrust of this argument, we would like to question some of the assumptions about how evidence can and does influence policy. This was the theme of the Sixth Annual Conference of the International Society for the Study of Drug Policy (ISSDP), which was hosted by the University of Kent in Canterbury in May 2012. Papers from this conference comprise the main body of this special issue. This editorial develops some theoretical ideas concerning the policy impact of empirical research, before introducing the articles which illustrate the variety of ways that drug policy analysis can be relevant to policy making. We are specifically interested in narratives – both of and in drug policy making – and the complexity of the policy process. We argue that these render some recommendations for improving drug policy somewhat naïve. Much more attention needs to be paid to issues of problem construction, politics, ideology, power and the messy complexity of the policy process.

The ISSDP conference opened with a plenary talk, inspired by the work of Colebatch (2005, 2006), from Ritter. She examined three narratives of what policy is. The narratives express multiple and overlapping accounts of policy and what might constitute ‘evidence’ or policy-useful knowledge in each account. The three narratives draw on varying accounts of governing. Each addresses different aspects of policy processes. We use these three narratives of policy process from Ritter's plenary plus the idea of narratives in policy making as a useful way to introduce this special issue and highlight aspects of the papers within each narrative.

Authoritative choice

The first is the narrative of policy as authoritative choice. In this narrative, policy is the technical process of solving official problems. The government is seen as the main agent in this policy process and it draws on research to help in this process. This has been referred to elsewhere (Nutley, Walter, & Davies, 2007; Stevens, 2007) as the rational-linear model of evidence based policy. As Ritter noted, this narrative most closely fits the description that many policy makers apply to their own use of evidence, as also demonstrated in interviews with policy makers in other research projects (Ritter, 2009; Stevens, 2011; Weiss, 1977). And it fits the assumptions that are shared by both Henderson (2012) and Nutt (2012) on how drug policy could be improved. It is strongly aligned with the ‘evidence-based’ policy (EBP) paradigm, and sees the role of evidence as instrumental in informing better drug policies. Much drug policy research is driven by the imperative to provide better and clearer research data to inform choices by decision makers. In their contribution to this special issue, Shanahan and Ritter exemplify the authoritative choice narrative in describing their work on cost benefit analyses of cannabis. In identifying the important benefits and harms of cannabis policy, and quantifying them in economic terms to assist rational decision-making, Shanahan and Ritter (2013) make apparent that such analysis itself involves choices about what to include and exclude. The apparent simplicity of judging costs against benefits of a policy may disguise a wide range of choices and judgements that have to be made.

Likewise, Santoro, Trioli, and Rossi (2013) use the tools of technical analysis (in their case, dynamic compartmental modelling) to show us how research can provide policy makers with a more accurate picture of developments in drug markets. Policy makers need to know how many people have used certain drugs. But they also need to know what patterns of use are prevalent within the market, and how and when people move between different patterns of use. The authors provide estimates of the proportions of Italian users of cannabis and other illicit drugs whom are considered to be occasional, regular and intensive users; and movements between various ‘states’ of use. These are all valuable points of information for policy makers, although Santoro, Triolo and Rossi note the range of uncertainties associated with these estimates.

More challenging, within the authoritative choice narrative, is gaining sufficient scientific knowledge about drug markets – as these are hidden, illicit phenomena not necessarily readily amenable to the type of quantitative analysis which is typically preferred for authoritative choice narratives. Three papers in this special issue highlight the challenges of drug market research. In their article, Barratt, Lenton, and Allen (2013) discusses the challenges associated with the growth in hidden internet services. A wide range of illicit materials are sold online, through services such as Silk Road, using the virtual currency Bitcoin. This means that drug traders are no longer confined by geography, making it even more difficult for law enforcement agencies to disrupt their activities. While this represents a threat to prohibition (i.e. that law enforcement can eliminate illicit drug use) it also represents an opportunity to move drug sales out of harmful open drug scenes, which are sometimes regulated by the fear of violence, to an online marketplace, regulated by the reputation of participants for honest dealing. Another drug market analysis, from Athey et al. (2013), examines cannabis cultivators across Denmark, Belgium and Finland and suggests that different characteristics are associated with the risk of apprehension in each country. In Finland, the proportion of growers who reported being arrested was much higher than in Denmark or Belgium, and there were significant differences in predictors of arrest between countries. In the final, third drug markets paper, Perrone, Helgesen, and Fischer (2013) examine the market for synthetic cannabis (often sold as K2, K3 or Spice in the USA) and mephedrone (often referred to as ‘bath salts’ in the USA). In addition to the common motivation for recreational drug use – curiosity – they show how techniques of prohibition that are common in the USA, such as drug testing of job applicants, can create incentives for people to seek out alternatives to the more traditional illicit substances.

These various important contributions to knowledge about drug practices, drug markets, harms and consequences all reveal the complexity of analysis in a domain where there are competing values and consequences. The authoritative choice narrative can appear to put insufficient attention on the complexity of decision-making in this context and to insufficiently acknowledge the gaps in our knowledge, especially when it comes to illegal products and behaviours.

There are other challenges for the authoritative choice narrative, including its privileging of only one policy actor (the government) and – usually – of only one form of knowledge (academic research). The EBP paradigm has been extensively critiqued for its narrow conceptualization of the relationships between science and policy (Hoppe, 2005; Marmot, 2004; Stone Maxwell, & Keating, 2001). A related critique, although based on a rather different epistemology, comes from Fraser and Moore (2011). They criticize the ‘circularity’ and ‘epistemological naïveté’ of EBP in the field of illicit drugs. It is viciously circular, they argue, because EBP accepts the neo-liberal logic of individual rational action, when it is rationality that is identified as lacking in drug users both by predominant accounts of drug use and policy responses to it. EBP is seen as epistemologically naïve because ‘it tends to take for granted that value-free, objective knowledge can be produced’ (Fraser & Moore, 2011, p. 2, italics in original). This narrative of authoritative choice neglects the complexity of the policy process, which is acknowledged by many observers, including Kingdon (1995), John (1998) and Hill (2009). And, with its implicit faith in technocracy, it fails to acknowledge the role of values in drug policy decisions (Humphreys & Piot, 2012).

Susanne MacGregor's paper (2013), shows how the narrative of policy as authoritative choice cannot capture the messy and value-laden process of British drug policy making. Through an analysis of primary interviews and secondary reports, she shows the influences of ‘cognitive bias, the shaping of attitudes, perceptions and decisions by reference to pre-existing sets of ideas’. Drug policy decisions are made on the basis of politicians’ own sets of values, and those values that they imagine to be held by the electorate, from whom they are increasingly socially distanced. Through these imaginings and interpretations, they construct the drug problem – and therefore the range of potential solutions to it – in particular, unevidenced ways. MacGregor's paper reflects a much more complex understanding of policy processes than that afforded by authoritative choice, and shows a number of the key actors, something which is highlighted in the structured interaction narrative, which we turn to next.

Structured interaction

The second narrative presented by Ritter was that of policy as structured interaction. This recognizes the interplay between organizations and stakeholders in policy fields. Decisions emerge, not from the choice of a sole authority, but from this interaction. This interaction is structured by the status of each of the participants in these discussions, and by the institutionalized processes of interaction. In their work on the policy influence of epidemiological monitoring systems, Ritter and Lancaster (2013) note the congruence of this narrative with Haas’ (1992) notion of the ‘epistemic community’. Under conditions of uncertainty, policy makers look to draw on the knowledge of groups that share ‘notions of validity’ and a ‘common policy enterprise’ (Haas, 1992, p. 3). A problem for the application of this narrative to drug policy making is precisely the absence of shared notions of validity, with different organizations and individuals drawing on different bodies of research, as was seen in the British debate over cannabis classification from 2004 to 2009 (Monaghan, 2008). Here, it may be more useful to use the ‘advocacy coalition framework’ of Sabatier and Jenkins-Smith (1993), with its emphasis on the formation of competing coalitions around different policy ‘core beliefs’. And, as Ritter noted in her presentation, the narrative of structured interaction does not explain the systematic exclusion of some policy stakeholders (e.g. illicit drug consumers) from drug policy debates (Lancaster, Ritter, & Stafford, 2013).

Boyd's paper (2013) exemplifies the structured interaction narrative, using the case example of Vancouver's Insite supervised injection facility. He documents the struggle to open – and keep open – the Insite supervised injection facility in Vancouver's downtown Eastside. The policy window for the establishment of Insite was opened by a change in Mayoral administration, with the election of Larry Campbell in 2002. The persuasive narrative was that drug injecting had become a public health emergency, rather than a criminal justice problem. But even with the Mayor's support, it was only possible to open the facility in the context of a rigorous evaluation, in order to gain exemption from the Controlled Drugs and Substances Act (CDSA). A side-effect of this is that there are now a wealth of high quality studies which demonstrate the effectiveness of the supervised injection site (e.g. Marshall, Milloy, Wood, Montaner, & Kerr, 2011; Small, Van Borek, Fairbairn, Wood, & Kerr, 2009). As Kerr, Macpherson, and Wood (2008), have also shown, such evidence did not speak for itself to assure the continuation of Insite. Boyd describes the three court judgements that have kept Insite open, but only under a continuing exemption from the CDSA which is specific to this facility; other cities that wish to open a supervised injection site will have to fight their own battles. The multiple players and formal institutions where structured interactions are played out highlight the interplay between organizations and stakeholders.

A second example of the structured interactions narrative of policy processes comes from Tieberghien and Decorte (2013) in their analysis of developments in Belgian drug policy. This article focuses specifically on the use of scientific evidence by policy makers. It finds examples of use of scientific evidence and experts to enlighten drug policy discussions, but more frequently observes ‘symbolic/tactical’ use of evidence to bolster pre-existing policy positions. This account is consistent with the narrative of policy as structured interaction and with the idea that a principle use of evidence is to support prevalent stories in the making of policy.

Social construction

The third narrative Ritter presented was that of policy as social construction. This suggests that there are no phenomena which are inherently problematic, but that the social process of problem construction makes them so. This perspective obviously has a long philosophical and sociological pedigree, encompassing contributions as diverse as those of Merleau-Ponty (Allen, 2007) and Berger and Luckmann (1966). It sees policy making as a process which constructs problems, rather than simply responding to them. In this third narrative, policy-making creates an understanding of what the problem is, whose problem it is and how that problem gets framed. As described in Lancaster, Hughes, Chalmers, and Ritter (2012) the death of a young man in a well-known nightclub district in Sydney provided the opportunity for a number of different stakeholders (police, community leaders, licensees and researchers) to frame the problem differently. The problem framing then leads ineluctably to a series of solutions. Bacchi (2009) describes a systematic analytic approach to study policy within this social construction narrative.

The paper by Hall and Carter (2013), while not explicitly a social constructionist perspective, does provide an example of the relationship between the definition of the problem (in this case brain disease) and the potential policy solutions. In their article, Hall and Carter discuss the idea that addiction is a ‘chronic relapsing brain disease’ (they argue that it would be more accurate to state that ‘severe forms of addiction can become a chronic relapsing brain disease’). This idea has been recruited to the support of the narrative that, as addiction is a disease, then medical – rather than criminal justice – responses are appropriate. Hall and Carter discuss some of the dangers of this narrative, including its use in supporting unevidenced and unethical initiatives, such as: compulsory treatment; invasive neurosurgery (e.g. deep brain stimulation); and high risk, high-cost policies of genetic screening and targeted drug vaccination. These could all be prioritized over more broadly based social policies which enhance psycho-social support and recovery for a wider range of people who have problems with drugs. Again, the narrative of what the problem with drugs is will affect the solutions that we decide to provide.

The social construction narrative sees government as neither the principal actor (as in the authoritative choice narrative) nor as a theatre of structured interaction, but rather as a less structured forum through which different discourses arise, compete, merge and emerge. The challenge for the social construction narrative of policy is its flirtation with epistemological relativism. It recognizes no extra-discursive anchor point (e.g. discoverable reality) on which to secure the policy debate. It lends itself readily to analysis of how policy is made but is less informative for academics and policy makers who are faced with the need to recommend actions that can reduce identified harms (e.g. drug related deaths and infections).

The deployment of narratives in policy making

Within each of the above narratives of the policy process we can also recognize the use of narratives in policy making. If policy, as Hajer (1993) has suggested, is made up of the structuration (i.e. the widespread acceptance) and institutionalization (i.e. the integration into practices of governance) of certain discourses, then we need to pay attention to how policy actors deploy narratives (in the sense of stories which make sense of a particular selection of events and observations) in persuading politicians and the public of the validity of their own discourse (Hajer, 1995). The creation and deployment of narratives in policy making was observed by Stevens (2011) in his ethnography of drug and crime policy making in the UK. Civil servants repeatedly revised documents to ‘improve the narrative’ and ‘sell the policy’. In these documents, they used both ‘killer charts’ (Stevens, 2011) and ‘killer facts’ (Bowen, Zwi, Sainsbury, & Whitehead, 2009) to make their narratives more persuasive (see also Weiss, 1977 on ‘tactical research utilization’). They knew that policies would be more likely to be successful (in the sense that they would be agreed and acted on – not that they would necessarily have positive impacts on the problem in question) if they fitted with the currently dominant narrative that structured existing policies. In the UK, for example, the repeatedly stated Home Office line on drug legislation is that ‘drugs are illegal because they are harmful’. Evidence that some drugs are less harmful than they have been stated to be, or that drug legislation is not reducing their harms, does not fit this narrative. It is therefore not likely to be included in policy documents by people inside government who wish to see their policies accepted and their careers enhanced.

An important aspect of the creation and maintenance of narratives by politicians is the role that ideology plays. Matthew-Simmons, Sunderland, and Ritter (2013) present research examining the extent to which a prevailing ‘ideology’ can be identified in the general public. While debate in the press and political rhetoric often repeats a two-sided discussion of ‘zero tolerance’ versus ‘legalization’, Matthew-Simmons et al. find a more complex underlying structure of attitudes in Australia. In this latent class analysis, there are not two but six underlying attitudes to drugs. There is differentiation between ‘detached’ and ‘committed’ prohibitionists, as well as between ‘harm reductionists’ and ‘legalizers’. It is interesting that none of the six groups form a majority. Given the complexity of the underlying structure of public attitudes towards drug use, it would be difficult for Australian politicians to deploy narratives that could command majority support.

A number of papers already discussed in this special issue contribute to our understanding of how narratives are deployed in the active creation of a persuasive or suitable ‘story’. For example, Boyd highlights the persuasive narrative for Vancouver – drug injecting had become a public health emergency, rather than a criminal justice problem. In Perrone et al., the interventions against new psychoactive substances conform to stories told to justify the restrictive practices (such as drug testing in workplaces), but the practices adopted to implement this narrative may actually have the opposite effect. The Athey et al. paper suggests that different stories about drug policies are told in different countries. Tieberghien demonstrates that the use of narratives in policy does not always lead to more restrictive drug policy. Finally MacGregor focuses our attention on politicians, and how symbolism, tribalism and taboos are central aspects to generating acceptable narratives.

Conclusion

Smith and Joyce (2012) show how our desire for understandings of the policy process that can be easily grasped mirror the politicians’ need for ‘data and concepts that help simplify (rather than capture) messy realities’ (Smith & Joyce, 2012, p. 73). They argue that the process by which evidence informs policy is a kind of complex system, in that it is characterized by: the need to study interactions in the system as a whole, rather than isolating its constituent parts; the presence of non-linearity and feedback loops which means that small actions can have big effects (and vice versa); periods of inertia, punctuated by sudden change (as in the punctuated equilibrium model of Baumgartner & Jones, 1993); sensitivity to initial conditions, implying path dependence (see, e.g. Schelling, 1978); and the need for interdisciplinary analysis to comprehend the process of knowledge translation.

This editorial has suggested that there are different conceptualizations (narratives) of the policy process, that evidence is used in the creation of persuasive narratives (stories) within the policy process and that our understandings of the use of evidence in policy must take account of the complexity of the policy process. The articles in this special issue provide examples of these various aspects. The variety in the articles displays some of the breadth of both topic and method that is covered by members of the ISSDP. This also suggests the complexity of the issue that has to be grasped by participants in drug policy debates. Narratives can be seen as tools for reducing the complexity of social reality – excluding some features and emphasizing others – in order to achieve a manageable level of understanding. Whether we think about policy as authoritative choice, as structured interaction or as social construction will influence our choices about how we engage in policy debates. And the findings which researchers insert into policy discussions will inevitably be shaped into narratives for use in policy arguments. By increasing both the range and quality of these findings, we can hope to improve the quality of these arguments in the policy process. But we should not think that this process is any less complex than the other phenomena that we study.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this article.