Field Notes

question

Does school behaviour grading cause later labour-force exit in men?

StatusWorkingUpdatedStands on1 sourceSupportsnotes

Cornwell, Mustard, and Van Parys (2013) showed that boys’ lower “Approaches to Learning” ratings — a composite of attentiveness, task persistence, eagerness to learn, organisational skills, and flexibility — account for essentially all of the gender gap in school grades. Boys receive worse grades not because they know less but because they behave in ways schools penalise. Male variability, education, and life outcomes sets out that finding and the grading literature around it.

Men without a college degree have exited the labour force in large numbers over recent decades, and the suicide literature identifies unemployment, low income, and low education as concentrated risk factors for men. LaCorte (2026) calls the classroom gap a “precursor” to adult crises including suicide.

But Lorant, Kapadia, Perelman, and colleagues (2021) found no causal evidence that education itself reduces suicide risk, suggesting the education–suicide association is confounded by shared underlying factors such as mental health vulnerability.

The chain from school behaviour grades through educational attainment to labour-force exit and suicide is correlationally plausible at every step but has not been causally tested as a single pathway while controlling for the underlying traits that may drive all of them.

Why it matters

If the chain is causal, interventions that reduce the behavioural penalty against boys (such as redshirting, de-emphasising sit-still demands, or grading academic content separately from classroom conduct) should produce downstream reductions in labour-force exit and suicide. If the chain is confounded — the same traits that cause poor behaviour grades also cause later difficulties independent of grades — then fixing the grading system would not change the downstream outcomes.

The question determines whether the boys’ education gap is a policy lever or merely an early signal of deeper problems that grading reform cannot reach. It also bears on how much weight Greater male variability can carry as an explanation: a variance ratio distributes boys into the bottom tail, but it does not say whether what happens to them there is set in the classroom or downstream of it.

What is already known

  • Non-cognitive skills (ATL) predict both grades and later life outcomes (Cornwell, Mustard, and Van Parys 2013).
  • The education–suicide association appears confounded rather than causal (Lorant et al. 2021).
  • Unemployment is a causal contributor to suicide risk (Milner, Page, and LaMontagne 2014), suggesting that whatever pushes men out of the labour force does increase suicide risk, even if education per se does not.
  • No study in this vault directly tests the full chain: school behaviour grades → educational attainment → labour-force participation → suicide risk, with controls for the shared underlying traits (impulse control, conscientiousness, mental health).

What a search in July 2026 established

Not the answer, but a sharper statement of why it is missing.

A scholarly search for longitudinal work linking childhood non-cognitive skills to adult labour-market outcomes returns a substantial literature — Carneiro, Crawford, and Goodman on British cohort data, and the Heckman-adjacent tradition more broadly — but that literature measures the trait and the outcome. The question here asks about the grade: whether being marked down for behaviour does damage over and above the behaviour itself, through tracking, credentials, and self-concept.

That is a mediation question, and it needs a design in which the grade varies independently of the trait that produced it. The obvious sources of such variation — grading-policy changes, teachers who mark behaviour differently, or the relative-age variation that ADHD relative-age effect exploits for diagnosis — are the places to look, and the search did not find them applied to this outcome.

What would settle it

  • Find longitudinal studies that follow the same individuals from early school grades through labour-force participation and measure both cognitive ability and non-cognitive skills. The key design requirement is panel data that allows controlling for the same traits at the point where grades are assigned and at the point of labour-force exit.
  • Check whether any sibling or twin study has isolated the effect of school grades on later labour-market outcomes while controlling for shared family environment and genetics.
  • Determine whether the “Approaches to Learning” construct has been linked to adult conscientiousness in any longitudinal study, since conscientiousness is one of the strongest predictors of both educational attainment and job stability.

Built on 1 source (1 external).

Working out connections…