EDUCATIONAL RESEARCHER 18 School Choice Impacts: What Do We Know? by Stephen Gorard, John Fitz, and Chris Taylor students per capita, making this effectively a national “voucher” scheme. Since 1988 the number of families selecting schools other than the local catchment (neighborhood) school has grown greatly. Where families are denied access to their selected schools (due to over-subscription, for example), the number appealing against the decision has also risen. Academic researchers in the U.K. have responded to this leg- islation by considering the nature of the limited market (e.g., Le Grand & Bartlett, 1993), the reasons reported by families for choosing a school (e.g., David, West, & Ribbens, 1994), and the process of choice itself (Gewirtz, Ball, & Bowe, 1995). Until re- cently, less work had been done on the impact of choice and on considering how effective choice has been (see extended litera- ture review in Gorard, 1999). It is therefore somewhat surpris- ing that some work of the type described above has already been cited as showing something about the impact of the school choice legislation in the U.K. For example, work on the process of choice by Gewirtz et al. (1995) reported interviews with 120 or so families involved in making the choice of a new school in one year. Yet it is being cited, sometimes even by the authors themselves, as evidence of the long-term impact of school choice. This work has been challenged by some (e.g., Tooley, 1997) for its published errors in classifications, its lack of working-class respondents, and for its ignoring of the difference between choosing a school for the first child as opposed to for all subsequent children (see Gorard, 1997a). A larger sample, reported by Willms and Echols (1992), suggested that parents use their impression of the socio-economic status of existing pupils to help in choosing a new school. This is an important finding. However, from these data they, and others, have argued that schools are therefore be- coming more stratified in SES terms. When re-evaluating both of these studies and others in the light of the findings below, it is important to realize that neither study actually shows any- thing directly about the composition of schools. In addition, both studies involve samples rather than populations, are lim- ited geographically, and lack a direct comparator for their post- choice datasets. These studies cannot show the stratifying ef- fects of markets because neither has equivalent data from the prior non-market system to draw comparisons with (or indeed from any other year). In reaching their conclusions, both stud- ies assume that the status ante was an ideal situation, whereas it was, in fact, a school system already heavily stratified in terms of area of residence (known colloquially as “selection by mortgage”). Research that overcomes these common limitations of scale or time can often introduce further problems. Ambler (1997), for example, purports to test the impact of ERA88 by using cohorts We present here a summary of the findings of what so far is the largest study of school choice in publicly funded schools, and the first analysis of changes over time in the characteristics and performance of students in an entire national school system (that of England and Wales). Our finding, in contradiction to some smaller studies re- ported previously, is that the socio-economic stratification of school students declined after the introduction of choice policies. We also show that standards in publicly funded schools rose relative to those of private schools over the same period. The extent to which these changes can be attributed to the impact of market forces in educa- tion is the subject of our discussion. This article is in part an answer to the question posed in a past issue of Educational Researcher by Goldhaber (2000): “School Choice: Do We Know Enough?” It is also a summary of the find- ings of what so far is the largest scale study of school choice and the first analysis of changes over time in the characteristics of stu- dents in an entire national school system (see Gorard & Fitz, 2000a). As such the paper refers to a great deal of empirical evi- dence from the U.K. concerning the long-term impact of mar- kets in education. Some commentators have rightly complained, and some are still complaining, that such evidence does not exist (Archbald, 1996; Fuller, Elmore, & Orfield, 1996; Goldring & Hausman, 1999; Jeynes, 2000). We hope therefore to bring this evidence to the attention of a wider audience. There is insuffi- cient space here to summarize all the varied data sources, meth- ods of analysis, and findings employed or produced thus far, and so throughout this article the interested reader is directed to other publications In the U.K., unlike in the U.S., a national program of parental choice of schools has been implemented. Thus all publicly funded schools in England and Wales are “choice schools,” and all publish raw-score outcome figures termed “performance” ta- bles. We have described the various policy instruments and sub- sequent court rulings in more detail elsewhere (Fitz, Halpin, & Power, 1993; White, Gorard, Fitz, & Taylor, 2001). The key legislation was the Education Reform Act of 1988 (henceforth ERA88), which gave all families the right to express a preference for any school (even one outside their local education authority), and denied schools the right to refuse anyone entry until a stan- dard or planned admission number was reached. Previously, local authorities assigned children to schools almost entirely on the basis of where they lived. Parents now have the right to ap- peal against their assigned schools. Funding to schools follows Educational Researcher, Vol. 30, No. 7, pp. 18–23