A number of countries are running role model recruitment drives under the assumption that like is good for like: ethnic minority teachers should teach ethnic minority children, women should teach girls, and so on. The empirical basis for this would appear to be case study and personal reflection. This article will examine quantitative data to test the hypothesis that male teachers produce more positive attitudes amongst boys and female teachers amongst girls. Using data from the Performance Indicators in Primary Schools (PIPS) Project, information from 413 separate classes for 11 year‐olds (in England) was examined. One hundred and thirteen were taught by males and 300 by females. All the pupils completed questionnaires that were designed to measure attitude to school, reading, mathematics and science. In addition, background data on those pupils were collected, including cognitive measures, attainment scores, ability measures and home background measures. The data were examined to look at attitudes using multilevel models controlling for background factors. The analysis concentrated on interaction effects between the gender of the teacher and the gender of the pupil and the results gave little support for those who advocate recruitment drives with role models in mind.