All 6 articles in this section
Mentoring evaluations find small average effects and large variation. Duration and the circumstances of ending predict outcomes better than anything about the matching.
Studies of parental influence find specific effects on subject choice, risk-taking and career aspiration. What is modelled at home predicts more than what is said.
The cognitive work of anticipating, tracking and organising is distributed unevenly in households, and the distribution begins well before adulthood.
Young carers are present in most schools, are substantially under-identified, and have measurably worse educational outcomes. Identification is the entire problem.
Peer influence during adolescence is substantial and operates through selection as much as through pressure. The distinction matters for what can be done about it.
Most community programmes close within two years. The failure modes are predictable enough to design around, and none of them concerns the quality of the activity.