The Heterogeneous Effect of Selection in Secondary Schools: Understanding the Changing Role of Ability. CEE DP 52

There has been a substantial rise in British education levels in recent decades, and new empirical evidence suggests that less able but wealthier children have benefited most from this expansion, at least during the 1970s and 1980s. There are many potential explanations for this trend. For example, during this period, the UK's highly selective "grammar school" system was largely dismantled. It is possible that reducing the extent of selection by ability in the UK education system may have altered the relationship between ability, family background and educational achievement, reducing the impact of ability on achievement and increasing the impact of family income and other family background factors. We test this hypothesis using data from the UK in the 1970s. Whilst the impact of the UK grammar system is of course of enormous historical interest, this issue also has significant policy relevance today. Certainly selection remains a topical issue. In the UK and indeed in the US, there have been recent policy initiatives that have attempted to explicitly or implicitly increase selection in the school system. Furthermore, in some parts of the UK, grammar schools remain an important feature of the education system. In this paper we explore and quantify the relationships between early cognitive ability, family background and school selection on the one hand, and educational achievement on the other, focusing particularly on the effects of selective schooling on different groups of students. For example, we assess the impact of selection on higher and lower ability pupils. Our results indicate that the most able pupils in the selective school system did do somewhat better than those of similar ability in mixed ability school systems. Thus the grammar system was advantageous for the most able pupils in the system, i.e. highly able students who managed to get into grammar schools. On the other hand, lower ability pupils did not do systematically better or worse in the selective school system. Many commentators have argued that the "comprehensive experiment" failed in England and Wales, reducing standards and educational achievement. To some extent our findings support this. The shift to mixed ability schooling did reduce the educational achievement of the most able. From an historical perspective, our results also suggest that the dismantling of the UK selective school system played some part in ensuring that the subsequent expansion of the education system disproportionately benefited less able (but wealthier) students. In terms of the current policy agenda, our evidence is potentially comforting to those who endorse increased selection in the education system, suggesting that the most able might benefit from a more selective system and that the negative impact of selection on the rest of the school population is likely to be small. However, the new forms of selection being introduced into the education system today are very different from the UK grammar school system of the 1960s. Furthermore, there are myriad other differences between the current UK education system and that of forty years ago, in terms of pupil characteristics, school standards, school enrolment patterns, curricula developments etc. All these other factors will also impact on pupil achievement, and interact with any selection that might be introduced into the system. Further research on current forms of selection is therefore urgently needed. Appended are: (1) Probability of attending a school in selective system conditional on comprehensive school density; and (2) Reduced form effects of local authority share in non-comprehensive schools. (Contains 11 tables and 24 footnotes.)