Developing the Response
At an early stage in response development, it is likely that one or more responses will emerge as intervention favorites. Indeed, the most suitable response will often appear obvious to those planning implementation. Nevertheless, it is worth asking the following 10 questions about any planned intervention:
- What is the change mechanism? You may conceive an intervention as consisting of two components-a description of what is to be undertaken (e.g., target–hardening, high–visibility patrol), and an explanation of how the intervention is expected to result in the desired outcome. This latter component is known as the mechanism. 22In the problem–specific guides, each response’s mechanism is described in the “How It Works” column of the summary table of responses found in each guide’s appendix. For example, the mechanism associated with burglar alarms might be to increase the perceived risk of detection for the offender, while an overt property–marking scheme may reduce the rewards associated with a theft. In specifying the mechanism, there are some simple rules to follow:
- Keep it realistic. The mechanism should provide a plausible explanation of how you might expect an intervention’s implementation to reduce a problem. Implausible mechanisms are unlikely to work as expected.
- Keep it simple. Mechanisms that require a series of events to achieve an outcome are less likely to be successful than those that have a more direct result. For example, installing target–hardening in burglary–victim households has a direct result of making it more difficult to break into the households. By contrast, crime prevention publicity campaigns that extol the virtues of target– hardening require a series of events. These include the householder’s reading⁄hearing the advice, deciding to take action, buying the necessary equipment, and correctly installing it. This provides a series of opportunities for the intervention to fail.
- Deal with immediate causes first. It is far easier to generate positive results by focusing on the things that can be immediately changed, rather than focusing on problems’ long–term root causes. This is because there are likely to be many intervening variables that will influence the likelihood of long–term root causes’ being addressed sufficiently to reduce the problem at hand, and because this is likely to take a long time to show results. Of course, this assumes that we are sufficiently clear about how root causes relate to a problem, and this is seldom the case. For example, one could tackle mobile–phone theft by addressing the social inequality that motivates some people to steal from others. Alternatively, one could manufacture mobile phones so that they became inoperable once stolen, and therefore worthless. The former provides an example of a root cause about which little is known and about which it is uncertain how to address. The latter provides an example of a more immediate response.
- What evidence is there that the intervention has worked before? Innovative interventions, while likely to bring accolades if effective, are inherently risky. There is a high likelihood that they will not work, which means that the original problem will not be addressed. Therefore, in the first instance, you should consider applying “tried and tested” measures that have been shown to be effective. There are, potentially, three sources for identifying effective interventions–your own implementation experience, interventions implemented by others in the organization and by other agencies, and published information (e.g., in academic journals, official reports, and internet sites) on effective interventions. In considering whether to replicate interventions that others have shown to be effective, you should ask the following questions:
- What was the intervention’s context? In replicating a project, you can expect to receive only the same outcome as the original project if the context in which you have replicated it is similar to the original context. 23 By context, we mean the environment in which it is implemented, the type of intervention recipient, and the nature of the organizations undertaking the intervention. If these factors differ, you cannot necessarily expect to achieve a similar result. For example, a burglary prevention project that involves target– hardening one kind of household won't necessarily be as effective if it is replicated with other types of households, especially if the burglary rates differ or if the modus operandi for the burglaries differ.
- How precisely was the intervention implemented ? In understanding how an intervention worked, it is important to analyze how it was implemented. Here you must consider the resources devoted to an intervention, the sequence of events associated with the intervention, and any external factor that was thought to have been important in the intervention’s success (such as the involvement of particular people, or the fact that it followed on from another unrelated intervention that assisted in achieving the outcome). One particular issue to bear in mind is the role of demonstration projects, which tend to be high profile and well resourced. Replications of such projects often do not have either the profile or the resources devoted to them, which can mean the intervention dosage is much reduced and therefore unlikely to achieve the results observed in its original form.
- How reliable are the stated results? You should critically evaluate the stated intervention results before deciding on replication. There is a tendency to report positive results and ignore negative ones, especially when a report is prepared by the project implementers, who have little to gain from presenting the failures of their own work. It can also be necessary to read between the lines about what brought about a given outcome, as other factors that are not explicitly reported may be responsible for the results.
- How difficult will it be to implement the intervention? Interventions that appear complex on paper will nearly always prove even more difficult to implement in practice. You should therefore give preference to interventions that involve
- simple, well–tested processes;
- a one-time activity, rather than a recurring process or program;
- actions by one person or a small team, rather than by multiple units or organizations; and
- limited line management decision-making or senior management authorization.
- Does the intervention rely on external partners’ actions? The more organizations involved in a project, the greater the problems you can expect. 24 This is partly due to the fact that you can exert limited pressure on an external partner to act as you would like, and partly due to different perspectives on the problem’s priority.
- Are regulatory or high-level policy changes required to implement the intervention? Interventions that require changes to laws or changes to official policies can be difficult to implement. They will require significant effort to be expended in lobbying for changes by those in authority, and even if successful, will take considerable time to come to fruition. If you deem such changes necessary, they are best treated as part of a package of measures, with shorter–term interventions that provide “quick fixes” balanced with longer–term policy changes.
- How will the intervention interact with other interventions being implemented in the same area/with the same group? Interventions are seldom undertaken in isolation, and are usually presented as a package of measures to tackle a given problem. Even if you undertake our intervention as a stand–alone measure, there may be other measures being undertaken with the same target group⁄area that may affect the outcomes. It is therefore important to take into consideration the interaction that your intervention may have with others. There are three ways in which interventions could interact: 25
- Interactive measures comprise interventions designed to work in complementary and cumulative ways. Often there will be an important sequencing of interventions, in which an intervention’s success is based on the successful implementation of a previous intervention, and in which there is mutual benefit in the two interventions’ being implemented together.
- Combined measures comprise measures that work independently of one another. Each measure may be tackling the same problem, but they have no influence on each other, either positively or negatively. This is the most common form of approach, with measures presented as a kind of shopping list and an assumption that the more interventions that can be applied to a problem, the more likely it is to be reduced.
- Contradictory measures consist of interventions that work against each other, with the successful implementation of one’s having a negative impact on the others’ effectiveness. For example, a covert alarm (in which one might plan to catch an offender in the process of committing a crime) would be counteracted if the premises also received target–hardening to prevent the offender from gaining entry (e.g., setting off the alarm).
- What will be the stakeholders' reactions to the intervention? You must take stakeholders’ views into account regarding an intervention’s success and how others view it. Stakeholders include the project team, others within your organization, external partners, and intervention recipients. There may be political reasons why stakeholders may not necessarily view success in a favorable light, especially if you have addressed a problem that they themselves should have addressed. You also should consider the intervention cost in terms of its effects on relationships with stakeholders. For example, doggedly pursuing an intervention may damage relationships with an external partner, which may have longer–term ramifications for others working with that partner. Alternatively, success in reducing a problem may be achieved at the cost of good community relations. For example, heavy–handed police enforcement could cause public unrest in the community experiencing the problem being addressed.
- Will any negative consequences accrue from the intervention? In addition to the damage to relationships noted above, you should carefully consider whether the intervention will have any other negative consequences. Examples of these might include
- the impact on other measures (such as causing contradictory interventions),
- the opportunity cost of being unable to address other problems while implementing the current intervention,
- media reactions to the measures,
- the displacement of the problem to other areas/times, and
- changes in modus operandi to more serious forms of crime.
- How long will it take for the intervention to show results? You might expect some interventions to have an almost instantaneous effect. For example, increasing high– visibility police patrolling is intended to have an immediate (if short–lived) impact on a problem. At the other end of the spectrum, preschool education programs can have an impact on reducing the likelihood of offending, but can take many years to produce results. 26
It is also important to consider how long it takes to implement the intervention. For example, high–visibility police patrolling takes as long as the patrolling itself. By contrast, the apparently simple intervention of blocking an alley to the rear of houses with a set of gates has been shown to take, on average, a year to complete in the United Kingdom. 27
- Can the impact be measured? Ideally, it should be possible to measure whether one has had an impact on the problem to know whether the problem remains and how well the responses worked. However, this is not always the case. For example, in some cases, where multiple interventions have tackled the same problem, it may not be possible to identify which interventions were responsible for the reduction. In other cases, interventions may be undertaken with no clear idea of the scale of the problem to start with and no way of measuring whether the responses have been effective. This was the case in an organized vehicle–crime reduction program undertaken by the U.K. National Criminal Intelligence Service, in which the scale of vehicles stolen for export could not be measured, either before or after intervention. 28
