This paper summarises the literature on Evaluability Assessments, and highlights the issues to consider in planning one. It notes that an EA should examine evaluability: (a) in principle, given the project design, and (b) in practice, given the available data and systems. It should also examine the probable usefulness of an evaluation. While there is limited systematic evidence on the effectiveness of EAs, their relatively low cost means they only need to make modest improvements to an evaluation for their costs to be recovered.
The paper uses the OECD-DAC’s definition of evaluability: “The extent to which an activity or project can be evaluated in a reliable and credible fashion”. Common steps in an Evaluability Assessment include: (i) identification of project boundaries and expected outputs of the assessment; (ii) identification of resources available for the assessment; (iii) review of the available documentation; (iv) engagement with stakeholders; (v) development of recommendations; and (vi) feedback findings to stakeholders.
The paper notes that Evaluability Assessments should cover three broad types of issues: programme design; the availability of information; and the institutional context. Other findings include the following:
- Locally commissioned Evaluability Assessments are likely to generate the most support and value. However, other complementary strategies may be useful, including centrally provided technical advice, screening of a random sample of projects in areas where little assessment work has been done, and mandatory assessments for projects with budgets above a designated size.
- The timing of an EA will depend on its expected outcomes: to improve the project design prior to approval; to inform the design of an M&E framework in the inception period; to decide if an evaluation should take place later; or to inform the design of an evaluation that has now been planned for.
- An EA’s recommendations should cover: (i) project logic and design, (ii) M&E systems and capacity, (iii) evaluation questions of concern to stakeholders, (iv) possible evaluation designs.
- EAs can be applied beyond specific projects, to portfolios of activities, legislation and other policy initiatives, country and sector strategies and partnerships.
- Ideally EAs would be carried out by independent third parties.
While there is limited systematic evidence on the effectiveness of Evaluability Assessments, their relatively low costs means they only need to make modest improvements to an evaluation before their costs can be recovered. The biggest risk of failure facing an Evaluability Assessment is likely to be excessive breadth of ambition: reaching into evaluation design or evaluation itself.