Labor market outcomes of formal vocational education and training: a meta-analysis

39 studies (1990-2022) for vocational education and training versus general education; 27 studies for VET versus an unrestricted control group

International; formal vocational education and training compared against general education, and against mixed or no specific education.

Hedges g with robust variance estimation, restricted to studies using causal counterfactual methods — a real improvement on the usual VET literature. It is still a meta of mixed designs, and it was read at abstract depth, so the design-quality moderator, the age or time-since-graduation moderator (crucial for the decay question) and any publication-bias correction are unchecked. Its headline framing is the single most useful methodological point in this literature and it maps exactly onto the archive's `baseline` field: comparing VET against a GENERAL-EDUCATION counterfactual is a different quantity from comparing it against no specific education.

Against a GENERAL EDUCATION counterfactual, VET yields a small positive effect (g = 0.021), raising employment by 4.8 percentage points but showing NO impact on wages. Against an UNRESTRICTED control group — people with no specific education type — VET shows a medium effect (g = 0.071), boosting employment by 11.7 points and wages by 37.6%. The gap between the two comparisons is the whole policy question: VET beats nothing handily and barely beats general education on employment, and not at all on wages.

Medium. Restricted to causal counterfactual methods, but the underlying designs vary in how well they handle self-selection into track.

The first meta-analysis restricted to causal counterfactual methods across the international VET literature. No competing meta exists to check it against.