2.21 The choice of evaluation approach will therefore depend on a range of issues such as:
• how complex is the relationship between the intervention and the outcome(s) of interest. How important will it be to control for other drivers of the outcome of interest? If control is important, this might point more towards an impact evaluation approach. Simple relationships can often be investigated just as robustly by process evaluations;
• the "significance" of the potential outcomes in terms of their contribution to overall policy objectives. More limited, intermediate outcomes might be more readily evaluated robustly, but might not give a close or direct measure of the benefits of the policy;
• how significant the intervention is, in terms of the identifiable change in practice or increase in resources it represents. This will affect the extent to which the intervention could be expected to generate a large enough effect to "show up" amidst the other potential drivers. The distinction between projects, policies and programmes, strategy and "best practice" initiatives is relevant here, since these can vary significantly in terms of how much they represent distinct and identifiable interventions3; and
• how the intervention is implemented, and whether this facilitates or hinders the estimation of the counterfactual. This is discussed further in the next chapter.
_______________________________________________________________________
3 Guidance for transport impact evaluations, Department for Transport, March 2010, provides a fuller discussion http://www.dft.gov.uk/