Greenhalgh, Harvey & Walsh 2005). Many prominent social researchers attribute such adverse side effects to the interventionist approach that dominates mainstream politics (Kazi & Spurling 2000; Sanderson 2002). In order to fully understand the reason behind this approach, the preliminary stages of interventionist methodology must first be examined. These stages are recognised as follows: problem identification, which involves measuring and ranking a specific aspect of society, such as crime; public disclosure, in which the information gathered in the previous stage is made publicly available; sanction instigation, which typically involves the broader community taking action by staging a boycott or lobbying the government to take action; and miscreant response, in which the failing parties are publicly shamed and criticised, thereby theoretically influencing an improvement in future performance (Pawson, Greenhalgh, Harvey & Walsh 2005).

Each stage within this series of events is often easily justifiable to both the public and policy makers alike, particularly in instances in which the call for policy reform is emotionally driven or instigated by a dramatic event. Pawson (IBID) argues, however, that the theories underlining each stage of this process are fallible and capable of failing at any point. In stage one, for instance, a problem may be falsely attributed to something that correlates with incidences of the problem. For example, the anti-vaccination movement insists that vaccinations are responsible for the increase in the rate of autism among children. As a result, advocates of this movement passionately campaign in favour of government intervention in order to prevent parents and doctors from vaccinating children (Calandrillo 2005). Further investigation, however, has revealed that this rise in autism is far more likely attributable to a combination of improvements in diagnostic methods, as well as to the fact that a range of social disorders have been added to the official autism classification (Tanne 2005). The reason that vaccinations were initially identified as being the cause of a rise in autism is because they are typically administered around the same time that children will begin to

exhibit symptoms of a disorder (Offit 2011). Nevertheless, the medical profession has been forced to endure every stage of the process outlined above as a result of the incorrect ‘problem identification.’

Assuming, in another circumstance, each stage of the interventionist process is valid, issues still arise regarding the universal applicability of the resulting policy. This question arises because the interventionist process is extremely fragile, as it is embedded within various social systems (Pawson, Greenhalgh, Harvey & Walsh 2005). In other words, a policy that successfully addresses homelessness in the urban parts of Sydney, for example, will not necessarily have the same relevance in rural areas of NSW. Moreover, the logic behind the two fundamental strategies of current evidence-based policy, ‘meta-analysis’ and ‘narrative review’, is argued to be insufficient (Pawson, Greenhalgh, Harvey & Walsh 2004). Pawson (2006) is particularly critical of the use of meta-analysis, referring to it as ‘systematic obfuscation’ (p. 38). While narrative review is certainly more sympathetic with the realist approach, both strategies have been criticised for sharing common limitations with regards to providing a template for important policy decisions (Pawson 2001). Both perspectives are only capable of making inferences from evidence that is already available (IBID). To return to the earlier example of the anti-vaccination movement, it would not matter which of these two strategies was used to evaluate the issue. The limited information available, namely that ‘children are being diagnosed with autism soon after being vaccinated’, only allows for the conclusion that vaccines may or may not cause autism. Note that the assertion ‘vaccines cause autism’ cannot be ruled out, nor can a more likely alternative be offered. This point is extremely significant, as the interventionist process relies heavily upon both meta-analysis and narrative review (IBID). Thus, the risk of public policy being subject to weak inductive reasoning, thereby leading to bad policy decisions, is considerable (Pawson 2002). This risk appears to increase alongside public outrage, and its consequences can be

dire. Fortunately, the anti-vaccination movement did not gain enough momentum to have a discernable effect on legislation. However, there are other examples, in which government intervention fuelled by public outrage has been prematurely imposed. One case study related to the US sex offender notification and registration programme (commonly known as Megan’s Law) reveals that the initiative was rushed into law following the murder of Megan Kanka by a convicted sex offender, who was living anonymously in her community. Understandably, the horrendous nature of the crime was at the forefront of the policy process, and evidence of its effectiveness only trickled in as an afterthought (Pawson 2002). Not surprisingly, subsequent research has shown that Megan’s law has not effectively addressed any of the potential issues in relation to released sex offenders and has instead become an increasingly expensive initiative due to the associated logistic difficulties (NCJRS 2008) In order to formulate the kind of EBP to which Pawson is referring, the hypothesis must be tested against new evidence, different variables must be added and subtracted in order to measure and compare the outcome, before ruling out all other alternate possibilities prior to reaching a final decision. This intricate process differs fundamentally between the type of EBP to which Pawson refers to and the interventionist approach. While a strong argument is made, suggesting that the interventionist process is inherently flawed, its key advantage over the realist alternative is its expediency. As previously mentioned, the realist approach demands a far deeper understanding of issues surrounding policy development. Like the scientific method upon which it is modelled, the realist approach is time-consuming and resource intensive (Solesbury 2001). Moreover, the public demand required in order to push through reform is often short-lived and is likely to focus on other issues by the time detailed empirical analysis is able to yield useful results. Pawson (2001) also acknowledges this shortcoming, stating, ‘Evaluation research is tortured by time constraints. The policy cycle revolves quicker than the research cycle.’