Despite the major increase in university student numbers in recent years there remains a strong negative socio-economic gradient in participation. The main reason for these differences has been shown to be educational factors, including prior attainment, family background and perceptions about costs and returns to education. In this paper, we examine the role of young people’s belief in their own academic ability (academic self-concept) as a way to explain differences in university participation rates. Using Next Steps data, we examine whether young people with higher academic self-concept are more likely to study A Levels, participate in further education and attend university. For those who do attend university, we examine whether young people with higher self-belief attend high status universities or study high status subjects. Results show that on average, controlling for prior attainment and other background characteristics, having high academic self-concept increases the odds in participating in A Level study, decreases the odds in taking part in further education, increases the odds in taking part in higher education (but the significance level disappears after taking A Levels is taken into account) and increases the odds of studying at a high-status university. While academic self-concept is an important predictor of later educational transitions, it does not entirely account for the social gradient in participation of university, further education or higher education. These findings have important policy implications for higher education participation and widening participation in particular.