Evaluating is the gerund of the noun ‘evaluation’. The Oxford English Dictionary gives three meanings for that word. The second with first usage dated to the late 18th century is to do with estimating or establishing the magnitude of any general quantity, but particularly scientific and engineering quantities. The third with a first usage in the 1960s is precisely about the subject matter of this chapter: the study of methods of evaluating the impact of social interventions. The first, and now largely obsolete, meaning makes ‘evaluation’ an exact synonym of ‘valuation’ and has to do with saying what something is worth. Actually that original meaning seems very useful in grounding the discussion in this chapter. Evaluation can be understood as the process of looking at what has resulted from some action and saying whether or not that action (or set of actions in normal interventions) was worth doing – if it had any positive value. In an ideal world we would want to attach a quantity to that value, and as we shall see, efforts are made in that direction, but in general we want to know if things have simply got better than they were before. Of course we also have to do some basic economic thinking here. We have to consider the amount of resources devoted to a purpose, consider how those resources might otherwise have been utilised – the opportunity cost – and again, ideally, have some mechanism of establishing a rate of return on resources deployed – the ability to conduct some sort of quantitative cost-benefit analysis. Can we do any or all of these things? How might/should we do these things? Why are social scientists engaged in doing them at all? These are the questions we will attempt to answer here. Inevitably there is a considerable overlap between the themes of this chapter and those of Chapter Two which dealt with constructing evidence. Here we want to focus in on the specific sort of evidence which tells us whether or not something has worked, and of course the whole purpose of evaluation as a process for telling us what works is so that we can establish transferable knowledge – in other words we can do again somewhere and somewhen else what has worked here and now.