Accountability

Accountability is a state of, or a process for, holding someone to account to someone else for something —that is, being required to justify or explain what has been done. Although accountability is frequently given as a rationale for doing evaluation, there is con siderable variation in who is required to answer to whom, concerning what, through what means, and with what consequences. More important, within this range of options, the ways in which evaluation is used for accountability are frequently so poorly conceived and executed that they are likely to be dysfunctional for programs and organizations.

In its narrowest and most common form, account ability focuses on simple justification by requiring program managers to report back to funders (either separate organizations or the decision makers within their own organization) on their performance compared to agreed plans and targets.

In theory, this sounds attractive. It seems likely that such a system will contribute to good outcomes for programs and organizations through providing an incentive system that encourages managers and staff to focus on and achieve better performance and through providing information for decision makers that will enable them to reward and maintain good performance and intervene in cases of poor performance. In practice, as has been repeatedly found, many systems of accountability of this type are subject to several forms of corruption and hence are likely to reduce the sense of responsibility for and quality of performance.

The most common problem is a too-narrow focus on justification through meeting agreed targets for service delivery outputs. In organizations in which this is the case, and in which, additionally, rewards and sanctions for individuals and organizations are tightly tied to the achievement of pre-established targets, goal displacement is highly likely (in goal displacement, people seek to achieve the target even at the expense of no longer achieving the objective). The most notorious example comes from the Vietnam War, during which the emphasis on body counts, used as aproxy for success in battles, led to increased killing of civilians in one-sided and strategically unimportant battles. Public sector examples of this sort of problem abound, but there are also many private sector examples in which senior managers have been rewarded handsomely for achieving specific targets at the cost of the long-term viability of the company.

Another common effect is that what gets measured gets done, as intended, but what is not measured is no longer valued or encouraged. A program may turn to “creaming”: selecting easier clients so that targets of throughput or outcomes can be achieved, at the cost of reduced access for those who most need the service. Other important values for the organization, such as cooperation across different units of the organization, may no longer be encouraged because of the emphasis on achieving one's own targets. Finally, there are many reported cases in which such a system encourages data corruption: Reported outcomes are exaggerated or modified to match targets.

Disquiet about the effects of this sort of evaluationis at the heart of concerns about high-stakes testing of children in schools, in which case serious sanctions for children, teachers, and schools follow poor performance in standardized tests.

...