Paying teachers by results

Don’t believe anyone who says women are less competitive than men in the workplace. New research with teachers shows just the opposite, says Victor Lavy.

Female teacher writing on whiteboardWomen are no less competitive than men in the workplace. This is the main result that I presented in a study published in June 2013 in The Economic Journal. Using experimental research with teachers in Israel, I analysed the outcomes of a real-life ‘tournament’, in which teachers of maths, English and languages could get large cash bonuses based on the test performance of their classes relative to others in the same school. The performance of teachers in this competitive environment was no different for men and women; nor did women’s performance vary with the gender mix of the teaching staff. What’s more, women teachers improved their performance in the competitive environment relative to the non-competitive one.

Recent evidence from laboratory and field experiments shows that males and females have different attitudes to competition and respond differently to it. These studies, which mainly involve children or university students, suggest that a sizeable part of the gender earnings gap could be explained by women being less effective than men in competitive environments, even if they perform similarly in non-competitive environments. Such findings are very important in the context of schools because of the recent expansion of performance-related pay schemes for teachers in the UK, the United States and elsewhere, and because women comprise a larger part of the teacher labour market than for other occupations with similar skills. If women are indeed less productive in a competitive environment, this finding may cast doubt on whether pay for performance, or merit pay, can improve school quality and students’ academic achievements, as many of these schemes involve some form of competition between teachers.

In contrast to this earlier work, I examined gender differences in competitiveness in a real workplace with adult participants. The tournaments, one for each subject, were part of an experiment with individual teachers’ incentives implemented in the 2001 academic year in 49 high schools in Israel. The research looked at gender differences in performance in this environment, and whether they vary by the gender mix of the participants. Teachers were awarded bonuses based on their tournament ranking, itself based on a value-added measure calculated by the difference between the actual mean performance of the teacher’s class and a value predicted on the basis of a statistical model that accounted for the characteristics of the students, the classes and the schools. Teachers were explicitly informed that they were competing against teachers of the same subject in their own school.

The results suggest that men and women’s performance under competition did not differ. I went on to examine some mechanisms that might explain that outcome. For example, there were no differences in teachers’ awareness of the programme, or in their effort or teaching methods. But there were gender differences in expectations of success in the competition, and in perceptions of the effectiveness of the incentive scheme. These were relatively large differences and they did not vary by the gender mix of the competition environment.

These results are different in some respects from the evidence from lab experiments. Several possible explanations can account for this difference. In the real world, most tasks are not completed instantaneously, so workers have time to plan, receive feedback, observe rivals and adjust their strategies and actions. In addition, competition in the new study is based on the regular activity of the participants for which they are educated and trained, not on an artificial laboratory task. Women teachers may therefore have more self-confidence and be less intimidated in competing against male rivals. In addition, the lab experiments are clearly set up as a zero-sum game, while the tournament experiment allowed the proportion of winners to vary across groups. Teachers were fully aware that they were competing against their colleagues, but if they helped each other anyway (for example, by helping other teachers’ students, or by helping other teachers become more effective), this would make the tournament very different from a lab experiment. Finally, it might be argued that since women dominate the teaching profession in terms of numbers, this is one of the few workplace environments in which we may expect not to find differences in competition by gender. But this argument cannot explain the vigorous response of women teachers to the incentive scheme, or the lack of difference between their response and the men’s.

A question might be raised about the applicability of these findings to other occupations. It might be easier for teachers to collude and behave strategically, as the group of participants is relatively small. So caution is called for in extrapolating the lessons of this study to other occupations. But the teaching profession is an important and large segment of the labour force, especially for women. So this evidence is important even if men employed in the teaching profession are not perfectly representative of men in other types of employment.

The evidence presented in this study provides additional support for the potential benefit of introducing incentives and competition in schools, and for further experimentation with these ideas in educational systems. However, it is important to note that the findings reported here reflect short-term adjustments by teachers. It is possible that a performance pay scheme may have a sorting effect on teaching staff composition in the longer term. Women may be more likely to leave the teaching profession under such a pay system.

Victor Lavy is professor of economics at the University of Warwick and at the Hebrew University of Jerusalem.

Share this page