Teaching vocational and practical skills

The program (what a school should actually do)

This subject has a single organising rule, and every finding under it is a variation: the outcome that matters is not the outcome that is measured. Vocational and practical teaching are routinely judged on test scores, which are flat, while the outcomes they actually move — completion, employment, earnings, conduct — go unmeasured or unreported. Two of the three meta-analyses in this subject could not compute an earnings effect at all.

If you build a career and technical programme, build a whole school, not a shelf of electives.

• The replicated positive is high-school completion: +7 to +10 points (Massachusetts), +10 points (Connecticut, on an 80% base), +8 points (Wake County lottery). Three states, three designs, three teams, one number.
• The large positive is male earnings, and it rests on two studies: the nine-site randomised trial gives men +17% a month over eight years (about $30,000, growing from +$260 to +$361 a month), and the Connecticut discontinuity gives +32.6% quarterly, still +34.5% at ages 23 and over.
• Depth, not breadth. Ordinary elective CTE course-taking returns about 2% per year of upper-level coursework and zero for introductory courses — and that is an observational estimate on students who positively selected in. (evidence: career & technical education)

Teach life skills late, short, and attached to a decision.

• The knowledge always moves: school financial education gives +0.25 SD on knowledge and +0.05 SD on behaviour — a five-to-one ratio. Across 76 randomised trials at all ages the figures are 0.204 and 0.100, and the school-age cell is the smallest behavioural effect in the paper (0.064 for under-14s).
• The behaviour moves only when the decision is near. State personal-finance graduation requirements shift college financing from private to federal loans and are associated with higher credit scores and sharply lower delinquency at 18-22; the same kind of mandate studied on 1957-82 cohorts observed decades later gives nothing on saving, investment income, credit score, delinquency, bankruptcy or foreclosure.
• Dose and cost: about 7-12 hours, $23-60 per student. (evidence: life-skills instruction)

Fix the enrolment mix before you open the doors.

• Every credible CTE design — the randomised trial, two discontinuities, a lottery, and Norway's national reform — finds approximately zero for female students. The mechanism is not a different response to the treatment; it is which programmes girls enrol in: 73% of male Connecticut technical students are in trades, 52% of female students are in cosmetology, hospitality and childcare. Norway's vocational expansion narrowed the class earnings gap about 20% and widened the gender gap by 8%.

What to refuse to spend money or time on

• Driver education. Three randomised trials, 17,965 young people, three continents, official licensing and police crash records: licensure relative risk 1.04, time-to-licence 111 days against 300, and crash relative risks of 1.01, 1.03 and 1.10 — every point estimate above one. This is the archive's clearest case of a school course producing net harm, and the mechanism is well identified: it gets teenagers driving sooner without making them better at it.
• Any life-skills course sold on a knowledge post-test. The post-test will pass. It passed for abstinence education, whose students got better at naming STDs and behaved identically to controls for six years: abstinent 49% versus 49%, age at first intercourse 14.9 versus 14.9, unprotected sex identical.
• A personal-finance requirement that displaces mathematics. The one study that estimated both from the same policy variation found personal-finance mandates did nothing and an extra mathematics course raised home equity $1,500-3,000 and cut foreclosure 0.4 points.
• Any forecast of vocational earnings past the mid-twenties. No lottery or discontinuity design anywhere has measured earnings past 33.
• The claim that vocational earnings advantages decay after 45, as though it were established. It rests on a design with no exogenous variation in education type; its mechanism failed an independent test; and every discontinuity design contradicts it.

The selection problem, which is this subject's defining feature

The single most instructive result in the subject is inside one Finnish paper. In the same population, the same cohorts, the same outcome and the same years:

Same data, opposite sign. The only thing that changed is identification. Romania and Croatia reach the same conclusion from national reforms — the observed vocational-general differential is mainly driven by self-selection — and the US transcript data says the sorting runs the opposite way from the usual assumption: students positively select into vocational coursework, so the funnelling premise behind most detracking argument is not true where anyone has checked it.

The same structure runs through the life-skills half. Interventions to improve financial literacy explain 0.1% of the variance in financial behaviour; measured financial literacy correlates with the same behaviours at r ≈ 0.13 (1.8% of variance) — and the partial effect collapses when numeracy, propensity to plan, risk tolerance and confidence are controlled, or when literacy is instrumented. The correlation between financially literate adults and financially successful adults is largely a fact about people.

The hereditarian bottom line for a founder

Vocational interests are about 50% heritable — roughly two-thirds of their stable variance — in a twin design that includes 92 pairs separated in infancy and reared apart, and heritability rises with level of aggregation exactly as it does for cognitive measures. The authors' own model is that heritable precursor traits of aptitude and personality guide the development of interests through gene-environment correlation: children steer themselves toward the work they are already suited to.

Three consequences a school builder should internalise:

1. "Vocational students earn less at 45" is not a fact about vocational schooling. Track choice is an expression of a substantially heritable disposition, and every design that breaks the selection — Finland, Norway, Romania, Croatia — kills or reverses the observed differential. The whole lifecycle-decay literature is genetic_confound_risk: high and should be read as such.
2. Where the design is clean, the effects are real. Lotteries, admissions cutoffs and birth-date discontinuities cannot sort on genotype, and it is precisely those designs that produce the graduation and earnings results this subject rests on.
3. Heritable interests are an argument for offering the pathway, not for assigning children to it. Heritability describes variance in a population and licenses no prediction about an individual child — the same rule the archive applies to athletic talent screening. The CTE result that works is admission by application to an oversubscribed school, which is students sorting themselves. The result that harms is institutions sorting ten-year-olds.

Confidence

Open questions a founder should watch

• Nobody has measured CTE earnings past age 33 with a credible design, anywhere — and the decay claim is located at 43-50. Linking existing cutoff designs to older administrative earnings would settle a forty-year argument.
• Would the gender null close if girls enrolled in the trade pathways? Nobody has tested it, and the enrolment mix varies enough across systems that somebody could.
• What is the effect of ordinary CTE course-taking inside a comprehensive high school — which is what most American CTE actually is? There is no credible causal design on it.
• Nobody has randomised a full-year high-school personal-finance graduation requirement. The randomised literature averages seven hours; the mandate literature has no randomisation. The intervention schools actually run has never been tested with the design that would settle it.
• No modern randomised trial of driver education exists, and the Cochrane reviewers' argument that today's lighter programmes would be worse rather than better has never been tested.
• First aid and CPR instruction were not reached in this pass. They are the one life skill whose target behaviour is both rare and immediately consequential, which is a different problem shape from everything graded here.