Self-affirmation theory suggests that some potentially stigmatised groups of learners, such as those from ethnic minority or poor families, face possible stereotype threats which could undermine their academic performance. Engaging in value affirmation writing activities at times when such threats are most salient can give individuals a positive sense of value, negating harmful feelings, and fostering academic learning. An important test of the useful productiveness of such a theory is the replicability of evidence concerning its predictions. This paper describes an independently evaluated randomised control trial of a self-affirmation intervention, replicating earlier studies which were mostly conducted in the US with ethnic minority students. The present study, involving 5,619 Year 10 and 5,188 Year 11 pupils (age 14 to 16), assesses whether the promising results can be replicated in England with pupils from low socio-economic backgrounds. The intervention involved pupils writing about self-affirming values, delivered at three crucial time points before a key school assessment. The results showed that pupils from lower socio-economic background in the intervention group made slightly more progress between their KS2 scores (end of primary education exam) and KS4 (national exam at the end of secondary education) results than similar pupils who did not receive the intervention. There was a small, but positive effect (+0.05) for the Year 11, and a sustained effect for the Year 10 pupils a year after the intervention (+0.04). Pupils who completed more exercises also performed better. Consistent with theory and previous studies, the replicated intervention had no effect for the majority of pupils who are not labelled as disadvantaged. While the effects are small, the intervention may still be worth considering, given that it does not cost anything, does no harm, and could help reduce the poverty attainment gap.