Several areas in the UK allocate children to secondary schools based on exam results at age 11. While many studies have investigated how attending academically selective schools affects pupils' subsequent educational attainment, we know very little about how grammar attendance affects other outcomes, such as pupils' self-confidence, academic self-esteem and aspirations. We investigate this by applying propensity score matching techniques to rich data from the Millennium Cohort Study. Results show that attending a grammar school has very little impact upon pupils' socio-emotional outcomes. Expanding grammar schools is therefore unlikely to benefit pupils in this respect.