Introduction
On assuming power, (New) Labour promised more accountability in the policy-making arena as part of their ‘modernising agenda’. Evidence-based policy-making, the development of policies devoid of dogma, was to be one of the vessels by which this could be realised (Blunkett, Reference Blunkett2000), and has become the benchmark by which policies are made and judged. The modernising agenda, for Solesbury (Reference Solesbury2001: 6), was part of the ‘opening up’ of government, initially characterised by the movement towards a ‘utilitarian turn’ in research. Here, researchers were required to produce work that was not only ‘useful’ but ‘useable’. This process was driven by the funders of social science who stressed the benefits of cumulating knowledge in the research enterprise. This model was also typical of the medical research process under the aegis of evidence-based medicine (Klein, Reference Klein2000).
Evidence-based policy is also synonymous with ensuring government policies provide value for money (National Audit Office, 2001). More recently, some stumbling blocks to its efficacy have become apparent, including its incompatibility with the principles of parliamentary democracy. Indeed, politicians as elected representatives of ‘the people’ have every right to ignore evidence and follow their instinct. There is, for example, a paucity of evidence that increasing the numbers of police-on-the-beat is an effective crime deterrent, but this runs contrary to the public's perception of the problem and is repeatedly ignored by politicians (Mulgan, Reference Mulgan2005: 225).
Criticisms that have arisen around the impact that politics has on the policy process are characterised by consummate sound bites, such as those which suggest that ‘policy-based evidence’ is actually more characteristic of the relationship between research and policy (e.g. Marmot, Reference Marmot2004). By the same token, accusations of ‘cherry-picking’ evidence lead to the frequent conclusion that much public policy is evidence-free (Finch, Reference Finch1986; Heinemann et al., Reference Heinemann, Buhm, Peterson and Kearney1990; Weiss, Reference Weiss1998).
With this in mind, Tonry (Reference Tonry2004: 146) has actually suggested that ‘on the small issues, evidence sometimes counts’; however, when it comes to the big issues, ‘politics is the order of the day’. ‘Big issues’ equate to heavily politicised, or ‘macro’ policy areas. These are not usually the remit of evidence-based policy-making, particularly in its guise as evaluation research. Such policies often contain no in-built evaluation and have no key performance indicators (KPI) through which to gauge impact. Heavily politicised policy areas are characterised by more ad hoc or muddled-through policy-making (Lindblom, Reference Lindblom1959). In such instances, there is intense media scrutiny of decision-making and prolonged conflict between competing interest groups and a permeating sense of crisis.
It must be stressed that if such views are staunchly held, then so is the retort that policy that seeks to use or engage with evidence on any level is (wholly) ‘evidence-based’. Both of these views of the evidence and policy relationship are problematic. They are premised on a zero-sum, linear platform which assumes a direct correlation between evidence production and policy formulation. Furthermore, they offer a static account of the evidence and policy connection, in that policies once labelled ‘evidence-based’ or ‘evidence-free’ must always be so. This neglects the dynamic nature of the policy process where new data constantly enters the public arena and may impact on policy.
Zero-sum conceptualisations are commonplace in politicised policy debates. However, the role of evidence in politicised policy areas is more complex than this. Models of research utilisation, depicted by Weiss (Reference Weiss and Bulmer1986) amongst others, have illustrated how there is often a more indirect way that research can impact on policy, as evidenced in the ‘enlightenment model’. Although existing models of research utilisation have provided a more nuanced understanding of the possible connection between research and policy, it is suggested here that these are also beset with difficulties, particularly in politicised areas. Consequently, a newer addition to this literature is required. Recent developments in UK drug classification policy are used to illustrate this.
The case study
The 1971 Misuse of Drugs Act (MDA) constitutes the main piece of legislation regulating drugs in the UK. It was enacted to combat the UK's perceived drug ‘problem’ at the time, although there is little agreement as to what this actually entailed. Under the MDA, drugs are separated into three classes − A, B and C − intended to reflect their relative harms and dangers, class A being the most harmful and class C the least. Sentences relating to drug offences are proportionate to the class of the drug. The Act has been the subject of much criticism since its operation (Young, Reference Young1971). It is widely held to be ineffective, in that since its implementation the levels of drug use have continued to rise; the opposite outcome of its intended goals (Rolles et al., Reference Rolles, Kushlick and Jay2006). As a result, critics claim that it was formulated for political reasons (so that the UK could fulfil its commitment to international drug control treaties), instead of with recourse to a robust evidence base. The Act also established Britain's first legal advisory body on illicit drugs – the Advisory Council on the Misuse of Drugs (ACMD) – which has come to assume a central role in reviewing British drug policy. There is a legal requirement for the government to consult this body if making any changes to the legislation, although it can act unilaterally if it wishes.
Since inception, the MDA has remained relatively stable, undergoing only minor alterations. One such episode was the 2004 cannabis reclassification from class B to C. Although a minor tweak to the legislation, it was of symbolic significance, representing the first instance in the UK whereby a widely used illicit substance had penal sanctions lightenedFootnote 1. Until this point, such occurrences had been politically unpalatable.
There were two initial triggers for this change. Firstly, a number of studies into drug policy and legislation (ACMD, 2002; Home Affairs Committee, 2002) suggested that in terms of its toxicity or harmfulness, cannabis was incomparable with other class B drugs. Secondly, the Police Foundation Report also recommended that reclassifying and making cannabis possession a non-arrestable offence, would reduce the number of ‘otherwise law abiding, mainly young people’ being criminalised, to the detriment of their futures (Police Foundation, 2000:7), thus removing a source of friction between the police and the wider community. This would also free up police time, enabling them to concentrate on the problems posed by class A drugs (Home Office, 2002). Juxtaposed to these events, was a pilot project of policing cannabis in Brixton, South London. The ‘Brixton Experiment’ effectively replaced the threat of arrest with informal disposal and a formal on-the-spot warning for those caught in possession of cannabis. The project was subsequently evaluated and judged a success by both the Metropolitan Police Authority (2002) and the local community (MORI, 2002).
Citing these findings, the then Home Secretary David Blunkett told the Home Affairs Committee in October 2001 that he was ‘minded’ to downgrade cannabis and would seek advice from the ACMD on the possibility of reclassification. Both parties reported back in early 2002 that cannabis should be reclassified. In July 2003, it was announced that cannabis would be reclassified to a class C drug, coming into force in January 2004. Although unique for a class C drug, the police maintained the power of arrest for cannabis-related offences (May et al., Reference May, Warburton, Turnbull and Hough2002).
In light of the change, early figures suggested that cannabis use had actually decreased (Roe, Reference Roe2005). Indeed, recent figures suggest that since reclassification, the proportion of people using cannabis has fallen from 25 per cent to 21 per cent (The Guardian, 2007). On this measure, it would appear that the change had ‘worked’. Such claims must be treated with caution, though, due to the aforementioned absence of meaningful KPIs.
Following reclassification, a significant issue to re-emerge was the long-running concern over cannabis use and the onset of mental illness (Mills, Reference Mills2003). A study by New Zealand scientists (Fergusson et al., Reference Fergusson, Horwood and Ridder2005) suggested that smoking cannabis – particularly virulent strains (Skunk) – virtually doubled the risk of developing mental illness, such as schizophrenia, for those with a family history of the illness and significantly increased the risk for those without. This was backed up by similar findings from the Netherlands (Henquet et al., Reference Henquet, Krabbendam, Spauwen, Kaplan, Lieb, Wittchen and Os2004) and Britain (Arseneault et al., Reference Arseneault, Cannon, Witton and Murray2004). Based on these findings and those from elsewhere, the then Home Secretary Charles Clarke ordered a review of the reclassification. The ACMD (2005) returned to the issue but found no grounds for a policy rethink.
The cannabis issue remained dormant for a while, but within weeks of being in the role, the then Home Secretary Jacqui Smith signalled her intention to review the status of cannabis. The ACMD (2008) reported once again that cannabis was not without risks, but maintained that class C was the correct position. Acting against this advice, however, Smith took the decision to reclassify cannabis back up to class B in July 2008.
Running concurrently to the cannabis episode was a debate between various constituencies concerning the efficacy of the MDA legislation. This debate has specifically focused on the ‘evidence’ for its tri-partite structure (Nutt et al., Reference Nutt, King, Saulsbury and Blakemore2007; Reuter and Stevens, Reference Reuter and Stevens2007; Royal Society of Arts, 2007; Science and Technology Committee, 2006) and whether this was the most effective means to regulate drug use. Central to this was the publication of the Science and Technology Committee Report. Independent research commissioned for this inquiry (Levitt et al., Reference Levitt, Nason and Hallsworth2006) suggested that increasingly evidence was playing a key role in policy formulation; in effect, that the policy was evidence based. In response, MacDonald and Das (Reference MacDonald and Das2006) claimed that in light of the Science and Technology Committee's findings, the UK drug classification system was actually an ‘un-evidence-based mess’. Neither standpoint accurately depicts the relationship between evidence and policy-making in this scenario.
The complexity and nuance of evidence in a contested policy area
A more complex and nuanced explanation is required to highlight the nature of evidence in politicised policy areas than the zero-sum varieties allow. Drawing on data from the author's own research, including in-depth interviews with a range of key actors in this policy subsystem, a closer look at the issue of drug classification reveals certain complexities in this area. Elsewhere (Monaghan, Reference Monaghan2008a, Reference Monaghan2008b), it has been suggested that in contrast to the Government view, a range of critical perspectives and appreciations (Vickers, Reference Vickers1965) relating to the efficacy and direction of UK drug policy, each with a specific understanding of evidence, can be ascertained. These illustrate how evidence is embedded in the decision-making process, but its selection does not take place on a level playing field.
As indicated, the MDA has remained stable and is still viewed favourably by the government. According to Vernon Coaker MP, one-time Parliamentary Under-Secretary at the Home Office:
The Government maintain that the classification system has withstood the test of time and that it continues to discharge its function fully and effectively . . . Our tri-partite classification system allows clear and meaningful distinctions to be made between drugs. Its familiarity and brand recognition should not be dismissed. (Hansard, 2007, June 14)
This position – that the system is fit-for-purpose – is understandable in that countenancing any radical change to drug legislation is tantamount to political suicide. In terms of the evidence-base, in oral evidence to the Science and Technology Committee, Coaker highlighted the centrality of one piece of research, here referred to as the ‘Nutt et al. matrix’ (Nutt et al., Reference Nutt, King, Saulsbury and Blakemore2007), to the Government decision-making process:
We have a matrix which we use. That matrix is part of the way we determine which drug should be in which category. Of course, we always look at the evidence that people give us; we always look at the opinions that they give to us . . . we have a scientific basis for determining harm. The ACMD refer to that when they classify drugs. (Science and Technology Committee, 2006: 105)
Published in the Lancet in 2007, but with its origins in the Police Foundation Report (2000), this research ranks the main illicit substances alongside alcohol and tobacco in terms of their harmfulness. This is based on a risk assessment over nine domains, three each relating to: (a) physical harm – damage to organs or systems; (b) dependence – the propensity of the drug to produce dependent behaviour and; (c) social harm – the myriad of ways drugs harm society, for example through the various effects of intoxication and anti-social behaviour. Each drug was scored on a 0–3 scale with a value for each drug derived from which a rank order of harm was produced (Nutt et al., Reference Nutt, King, Saulsbury and Blakemore2007). The ranking of the matrix did not correspond exactly with the current A, B, C classification employed by the MDA. The most noticeable discrepancy being ecstasy, a class A drug, deemed by the matrix to be less harmful than all other class A substances but also ketamine, benzodiazepines and cannabis all class C drugs at the time of publication.
The matrix supports the philosophy of the MDA, but questions its validity in its current form. It has also come to assume a dominant position in the deliberations of the ACMD. Although the authors of the matrix stop short of suggesting that this is the solution to the drug classification problem, supporters suggest that it is a step towards a better evidence-base for the development of drug policy (Hall, Reference Hall2007: 927). It is suggested here that the matrix is central to one of at least three recognisable appreciations that contradict the government line that the current system is fit for purpose. These three perspectives start from the premise that the relationship between evidence and policy in this area is turbulent. This is due to the fact that alcohol and tobacco, the most widely used substances (along with caffeine), and generally regarded as the two substances which cause the most harm, are absent from the legislative framework of the MDA. The three strands of critique are matched with tailored policy solutions. These include:
(a) The ‘rational’ view that certain drugs are located in the wrong class. The relative harms and dangers of certain substances do not match drug user's experiences, thus discrediting the framework; the solution is to devise a framework which more accurately reflects the harms and dangers of substances, incorporating drug treatment into a regulatory system within a criminal-justice framework.
(b) The ‘radical’ view that prohibition is failed philosophy. There is little evidence that this most restrictive form of regulation actually has any deterrent effect; the solution is to legalise and regulate the use of all substances.
(c) The ‘conservative’ view that the MDA makes an unnecessary distinction between drugs in terms of their harms and dangers. In reality, all drugs are harmful and no hierarchical distinction is necessary between illegal drugs in terms of risks; the solution is rigorous control of all currently illegal drugs, allied with more emphasis on drug prevention education (Monaghan, Reference Monaghan2008b).
For the radical and conservative perspectives, the MDA is meaningless and of limited value. For the rationalists, it is anomalous, but with modification it can be rectified.
The rational position is closest to that of the government in terms of philosophy. Indeed, since the 2004 reclassification of cannabis, the matrix of risks and harms of drugs has assumed a dominant status as the evidence-base on which drug policy should be developed. It is here that we witness an overlap of medical and social appreciations of evidence. According to one senior civil servant from the Home Office, in relation to the 2004 reclassification and 2005 review:
I think David Nutt's work was probably the dominant evidence-base and the ACMD did a sort of review of the literature with colleagues from Department of Health and that fed into it.
Supporters of the radical perspective argue that the prevailing philosophy of the MDA prevents it from being evidence-based as it operates in the paradigm of prohibition, an ‘evidence-free zone’ (Rolles et al., Reference Rolles, Kushlick and Jay2006: 6). For these thinkers, the Nutt et al. matrix is also located in this paradigm, sharing a general philosophy which articulates a commitment to the regulation of drugs with criminal penalties based on a hierarchy of harms. Consequently, this does not factor into the analysis the potential harms caused by the illegal status of drugs. These include the harms caused by unregulated strengths and purity of substances. This perspective is succinctly summarised by a journalist, writer and member of a leading NGO advocating drug law reform:
I think there is a lack of clarity about harm because people don't distinguish between the harm caused by prohibition and the harm caused by drugs and I think this is skewed rather by alcohol . . . In terms of drug harm, what most people see in real life is alcohol, leery people fighting in the street, but the assumption is that alcohol is safe because it is legal and all other drugs are worse than alcohol . . . In public health terms, alcohol is really quite conspicuously harmful and the harm caused by cocaine, for example, is really hard to identify . . . So I think that drug harms are radically misperceived because of the illegal activities associated with them. But I appreciate that Colin Blakemore and David Nutt have tried to put this on a footing and get it right but still I find that odd . . .
In essence, there is recognition from this perspective that their concept of ‘evidence’ does not sit comfortably with that of the government. This stems from differing opinions on the nature and causes of drug harm, which for this group stem from the (il)legal status of drugs.
The key feature of the conservative perspective is the concern with the deleterious properties of the drugs themselves. There is also a sense that the initial implementation of the MDA has done little to address the drugs problem. In this sense, there is some overlap with the radical position. For the conservative perspective, however, it is not the law itself that produces the problem per se, but that the current legislation does not target the correct areas. There is not enough emphasis on initial prevention and a three-tiered structure sends out the messages that some substances are less dangerous and more socially acceptable than others. Consequently, one former UK government advisor and current advisor to the Scottish Parliament stated that:
I don't think that the [MDA] is enormously shaped by evidence. I think that in certain senses it is a bit of a mystery as to why certain drugs are placed in some categories and others are in a different one. I think that one can envisage that the classification system is something that should be much more evidence-based . . . devoting an enormous amount of time to placing drugs in their right position [like the Nutt et al matrix] . . . But my anxiety is that is rather fiddling while Rome burns and that actually the classification system is neither here nor there . . . It is something of an irrelevance in terms of the choices and decisions that people make over whether to use or not use illegal drugs.
It is apparent from the above discussion that there are a range of views over the extent to which the government draws on evidence in making classification decisions. It is suggested here that the government, in spite of not following the wider claims of the Nutt et al. initiative, has incorporated the methodology of the latter into its decision-making process; a point recognised and accepted by all perspectives. In contrast to the zero-sum positions outlined above, there is, then, widespread recognition that some semblance of evidence is used in the decision-making process. This does not necessarily mean that classification decisions are evidence-based, because various other manifestations of evidence are left unused, particularly the radical and conservative kinds.
The question that remains is how to explain this imbroglio. In essence, what is required is an explanation of how particular kinds of evidence can set the perimeters, or are utilised in the process of policy decision making, without necessarily directly shaping the outcomes of the policy. The emergence of new data must also be factored in as the dynamic nature of drug classification policy shows. This subtlety is often missed in the existing models of research utilisation.
Models of the evidence and policy relationship
Stevens (Reference Stevens2007), by drawing on the work of Weiss (Reference Weiss1977, Reference Weiss and Bulmer1986) amongst others, has produced a four-tier typology of the use of research in the policy-making process, which includes: the ‘linear model’, the ‘enlightenment model’, the ‘political/tactical model’ and the ‘evolutionary model’. The first three models are established in the literature, with the final one introduced in this journal – and for which he claims supremacy – a vital addition to the literature, but also beset with problems.
The linear model portrays a direct link between evidence production and policy decision-making. Young et al. (Reference Young, Ashby, Boaz and Grayson2002: 216) distinguish between two manifestations of the linear model, suggesting that in the ‘knowledge-driven’ variety, research ‘leads’ policy with the role for the expert ‘on top’. By contrast, in the ‘problem-solving’ model, research follows policy and the expert is ‘on tap’. Although considered to be the ‘common-sense’ understanding of the research and policy relationship (Lindblom, Reference Lindblom1980), the rationalist assumptions of this model have been the subject of much criticism. The model is premised on a static view of the policy process where research is directly linked to policy. Young et al. (Reference Young, Ashby, Boaz and Grayson2002:218) suggest that the policy-making process is ‘rarely characterised by rational decisions made on the basis of the best information’. Social research, in most cases, takes a long time to produce and is often typified by lengthy, verbose reports which are often of little use for policy-makers who require clarity and concision.
For Stevens (Reference Stevens2007), the enlightenment model represents a more ‘sophisticated’ attempt at explaining the relationship between evidence and policy. It refers to the arbitrary way that research can enter the policy arena. In this scenario, social science evidence ‘percolates’ into the ‘informed publics’ consciousness and comes to shape the way they see the world. In effect, this is premised on a dynamic view of the policy process and broadens the platform on which the evidence and policy relationship can operate. There is, however, the lack of a filtration mechanism by which evidence is siphoned out, and as a result there is an in-built assumption that all types of evidence have an equal chance of being used in decision-making.
The political/tactical model is characterised by the selective use of evidence to satisfy the ‘short-term’ interests of policy-makers. For Weiss (Reference Weiss and Bulmer1986) these two are also separate. In the ‘political model’, various interests and ideologies coalesce to pre-determine the position of policy-makers on particular issues or problems. New research or evidence has a legitimising or justificatory function. In the ‘tactical model’, the actual findings of research are unimportant, but the process of research is of utmost significance. When faced with calls for a policy change, seen to be unpalatable, the government is able to delay the process by playing what Duke (Reference Duke2003: 17) calls the ‘research card’. In effect, they procrastinate by claiming that policy change will be based on the outcome of ‘ongoing’ research. The political model is, however, premised on a similar logic to the linear model, assuming a direct association between research production and policy formulation.
Ultimately, the established models of evidence-based policy operate at a high level of abstraction and pay scant attention to what Bulmer et al. (Reference Bulmer, Banting, Blume, Carley and Weiss1986: 10) has termed the ‘content of the negotiations’. These are the mechanisms through which research is selected for use in policy. The evolutionary model, by contrast, is able to explain the pattern of selection of evidence used in policy-making. This does not rely on a deterministic explanation equating the phenomenon with the manipulation of evidence by policy-makers. It suggests that the pattern of selection is premised on ‘classic’ evolutionary social theory of ‘survival of the fittest’ (Spencer, Reference Spencer1891). Research comes to influence policy when it is consistent with the ideas of those in positions of power and thus comes to form the basis of government decision-making:
[I]deas come from evidence and compete for attention in policy, as genes arise and compete for survival. The ideas may be facts, findings or recommendations that have been produced by academics, journalists, think tanks, pressure groups or others. Some of these ideas fit the interests of powerful groups and some do not. Ideas that do fit will find powerful supporters. Others will not. Those ideas that fit will therefore have groups and individuals that can carry them into policy . . . The ideas that do not fit will tend not to be picked up by people who have the power to translate them into policy. This evolutionary advantage leads to the survival of the ideas that fit. (Stevens, Reference Stevens2007: 28)
Giddens’ (Reference Giddens1990) notion of structuration is, therefore, central to this. Issues of power are also predominant. In essence, it is not the power of the idea alone that ensures its survival in policy, but there is a reciprocal relationship between the idea and the carrier (the powerful supporter). These conditions, under which an idea either flourishes or founders, are shaped in terms of what Stevens calls the ‘mechanisms of selection’, of which there are four main kernels.
The first mechanism is trawling. Policy-makers may fish for evidence, select the beneficial bits and throw back those that are un-required. Second, and similarly, policy-makers can ‘farm’ for evidence. Here research is specifically commissioned to provide evidence for the proposed policy, although only that which actually supports the policy is published. A common tactic employed in this regard is the use of ‘repetition’, whereby powerful actors are able to focus the attention of potential critics towards evidence that supports a said policy. This often involves evidence being ‘ripped out of context’. In doing so, this evidence becomes a key part of the knowledge-base justifying the ‘validity’ of the policy to which it equates.
Drawing on the ideas of Chomsky and Herman (Reference Chomsky and Herman1988), a third mechanism is the tactic of creating ‘flak’. In this scenario, a significant amount of disquiet is created over evidence that enters the public arena that is detrimental to the chosen policy direction. The media play a powerful role here, discrediting potentially troublesome evidence. Stevens’ final mechanism involves the imposition of ‘strain’ (Chambliss, Reference Chambliss, Chambliss and Mankoff1976) on organisations or individuals who produce and advocate ‘unhelpful’ evidence to those in power. In academia, this could occur through the actual, or threat of, funding withdrawal for projects seen as anathema to powerful groups.
Explaining evidence use in a politicised policy area
Overall, the established models of evidence-based policy can, at best, describe what the relationship between evidence and policy formulation is in any given policy domain. They are less capable of explaining why and how this connection occurs. It is suggested here that the evolutionary model is capable on both accounts. It can offer a plausible account of evidence utilisation in UK drug classification policy, providing an explanation of the centrality of the Nutt et al. matrix (and rational appreciations of evidence) to government decision-making. This involves the mechanism of flak.
For the rational perspective, the matrix provided a welcome dose of science into this area of drug policy and was thus viewed as a long overdue development. Additionally, the rational appreciation works within the current framework and does not require any radical change of philosophy. It suggests that any defects in the current MDA system can be overcome by making it more scientifically valid. The rationale is that individuals can make informed choices on the use of drugs if supplied with accurate information, which could reduce drug use. This requires minor tweaks to the legislation. The conceptualisation of evidence cited by this perspective is, therefore, contemporaneous (of the same moment), but also consanguineous (of the same kin) with the legislation to which it refers. It also has its origins in medical discourse, a discipline noted for its evidence-based policy-making credentials.
The radical perspective advocates the repeal of the current prohibitionist legislation of the MDA and its replacement with a system of legalised regulation. In effect, the main problem with the current system stems from its enactment in the first instance. This is because, although drug harms are innate to the substances themselves, they are magnified by their illegal status. Control of the market is surrendered to illegal producers and distributors, and purity, for example, goes unregulated. The main danger for the drug user occurs because of the legislation. In causal terms, this is after its implementation.
The conservative position suggests that significantly altering and simplifying the current legislation, to more accurately depict the true dangers of substances, should be the direction that UK drug classification policy takes. In effect, this involves switching to a system that places more emphasis on drug prevention and drug enforcement. For this perspective, the current system fails to invest in the former aspect to a satisfactory degree, thus the main danger for drug users occurs before the MDA can have any impact.
In sum, the rational understanding of the drug problem and its associated evidence-base does not rely on speculation as to what a policy would look like if all the ‘evidence’ in the debate was considered. Conversely, the radical and conservative perspectives’ solutions to the drug problem involve much speculation as to how their chosen direction for policy would represent an improvement on the current deadlock. It would not be possible, for example, to base their policy solutions on ‘scientific’ evidence as no data exist about the efficacy of their policy goals in the UK context. Speculation is contrary to the principles of evidence-based policy and, as a result, the rational perspective can direct flak at the other perspectives on the grounds that their appreciations of the issue are actually devoid of scientific credentials.
It appears, then, that the evolutionary model can explain how certain kinds of evidence obtain a privileged role in policy debates. Ultimately, however, the evolutionary model is also beset with difficulties. At its most simple level, the evolutionary model assumes that the ideas that survive because of their utility to powerful groups will be used in policy outcomes. In effect, this assumes a direct causal connection between evidence production and policy formulation, not dissimilar to that of the linear model. It is premised on linear logic and offers a static and narrow view of the evidence and policy relationship on the grounds that once evidence is adopted by powerful groups, the policy becomes evidence-based and will remain so.
As we have seen from the above discussion, the role of evidence in politicised policy-making occurs more in the decision-making process than in the outcomes. Allied to this, heavily politicised policy areas are dynamic and often in flux. The cannabis issue is testament to this, highlighted by the changing status of the drug in this decade and the associated data called on to illustrate this. With its static outlook, the evolutionary model struggles to explain this development.
Overall, it is argued here that to understand the role and nature of evidence in politicised policy areas, a newer model of the evidence and policy process is required that: (a) focuses on the mechanisms of selection of evidence, but (b) does not restrict the role of evidence to outcomes and (c) recognises that the policy process is dynamic. This model has yet to be fully developed, but the above criteria represent its pre-requisites. In this sense, it can be seen as a synthesis of the evolutionary model with its emphasis on the mechanisms of evidence selection, and the enlightenment model with its dynamic view of the wider policy process. It is suggested that this could be useful for explaining the nuanced relationship between evidence and policy and does not rely on zero-sum statements that are, at best, inaccurate and, at worst, unhelpful. This synthesised appreciation of the evidence and policy relationship is tentatively introduced as the ‘processual model’. It is tentative as one must be wary of generalising from the findings of one case study and so the familiar caveat that more research is needed is expressed.
Conclusion
Politicised policy areas are not usually the foci of studies of evidence-based policy-making. Although they are ad hoc, evidence is still embedded in the decision-making process. This complexity is often missed and zero-sum appreciations of the evidence and policy relationship abound as a result.
Overall, it is suggested that the processual model offers a subtler explanation of the evidence and policy relationship, particularly in the context of heavily politicised areas. Lessons are learnt from the established models of research utilisation. The enlightenment model, for example, offers a more flexible understanding of the role that evidence can play in policy. Ultimately, however, the established models operate at too high a level of abstraction. The actual minutia of the way evidence is selected in such policies is often missing from these models. The evolutionary model overcomes this issue, but its logic is too similar to the linear model. In this sense, the evolutionary model is synthesised with the enlightenment version into the processual model. As the name suggests, this views the nature of policy-making as a process rather than a single event. In this sense, the role that evidence can play in policy decisions is widened. Evidence that survives filtration is often utilised in policy deliberations, but as recent developments in UK drug policy testify, these may not necessarily be translated into policy outcomes but they do constitute a significant aspect of the decision-making process.
Acknowledgement
The paper is, in part, based on doctoral research funded by the Economic and Social Research Council. Particular thanks go to those individuals and organisations that willingly gave their time to participate in the study. In addition, thanks also to the editors and two anonymous referees from Social Policy and Society.
