Impact Evaluation - Are We 'Off the Gold Standard'?
Impact Evaluation - Are We 'Off the Gold Standard'?
复制标题
影响评估——我们是否“偏离了黄金标准”？
DOI：
10.1057/ejdr.2013.42
复制
发表时间：
2014
期刊：
影响因子：
--
通讯作者：
Camfield L
中科院分区：
文献类型：
--
作者：
Camfield L
主题辩论部分的短文说明了评价和评价复制的重要性。它们可以改善发展政策和方案的质量，并解释有意和无意的后果。这些文章提出了这样的问题：影响评估是评估吗？随机对照试验（RCTs）能帮助决策者吗？对可观测的关注到底告诉了我们什么？在这篇引言中，我们总结了他们的论点，并提出了切实可行的建议，以提高评估研究的质量，无论是定量的，定性的还是混合的方法。本节以Lensink的文章开始，该文章清晰地总结了影响评估中使用的不同方法，以及它们可以为政策制定者和方案设计者提供的信息。他提出，进行影响评估的最佳时机实际上不是在项目结束的时候，而是在项目进行试点以找出“干预是否以及在什么条件下可能起作用，[并]能够测试不同理论的相关性”的时候。这可能需要研究人员和项目经理共同努力，通过定义项目目标和思考干预的因果效应来发展一种“变化理论”。然后，成功的试点评估可能是为项目的推出提供资金的先决条件。Lensink指出，由于许多原因（例如，缺乏比较组，或存在多个组成部分），评估现有项目是有问题的，因此总是需要额外的定性和定量技术来控制偏差并验证结果。然而，由于一个特定项目的结果可能在同一个国家的其他环境中不成立，更不用说在其他国家了，“没有在其他环境中得到复制研究验证的独立练习[rct][…]永远不能真正指导国家援助政策”。
The short pieces in the themed debate section speak to the importance of evaluation and replication of evaluations. They can improve the quality of development policy and programmes and explain intended and unintended consequences. The pieces raise questions such as: Is impact evaluation evaluation? Can randomised controlled trials (RCTs) help policy makers? and What does a focus on observables really tell us? In this Introduction we summarise their arguments and make practical proposals to improve the quality of evaluation research, whether this is quantitative, qualitative or mixed methods. 1The section begins with Lensink’s piece, which provides a clear summary of the different methods used within impact evaluation and the information they can provide policy makers and programme designers. He proposes that the best place for impact evaluations is not, in fact, at the end of the project, but while the programme is being piloted to find out ‘whether and in which conditions an intervention is likely to work [and] enable tests of the relevance of different theories’. This might involve researchers and project managers working together to develop a Theory of Change2 by defining project aims and thinking about the causal effects of interventions. A successful pilot evaluation could then be a precondition for financing the roll-out of a project. Lensink notes that evaluating existing projects is problematic for a number of reasons (for example, lack of a comparison group, or the presence of multiple components) and therefore additional qualitative and quantitative techniques are always needed to control for bias and validate outcomes. However, because results for a particular project may not hold in other settings in the same country, let alone in other countries,‘stand-alone exercises [RCTs], not validated by replication studies in other settings […] can never truly guide national aid policies’.