half the classroom, one study

Problem statement

Federally funded adult education under Title II of the Workforce Innovation and Opportunity Act (WIOA) serves adults in three streams — adult basic education, adult secondary education, and English as a second language — and, per the snapshot, "learners seeking to improve their English proficiency make up about half of all those in federally funded adult education programs." When IES commissioned the first systematic review to apply What Works Clearinghouse standards across adult education strategies, the team screened more than 2,000 references, found only 54 that tested a strategy's effectiveness, of which 22 met WWC standards — and "only one study examined English language acquisition instruction strategies," which found no positive effects. The unsolved problem is not merely that the studies are missing; it is that no evaluation approach has been fitted to how these programs actually operate (open enrollment, part-time and interrupted attendance, adults juggling work and family, providers that bundle strategies), so the evidence gap persists three decades into a federally mandated "evidence-based" regime.

Why this matters

Nearly one in ten U.S. adults is limited English proficient, and while most are in the labor force they earn less than English-proficient peers; the snapshot's framing population is the "nearly 43 million U.S. adults [who] lack the basic English literacy skills required to succeed in the workforce." WIOA "encourages adult education programs to use evidence-based strategies," and providers must choose instructional approaches every day — but the review concludes that "research has yet to identify specific literacy instruction, numeracy instruction, and English language acquisition instruction strategies that lead to positive effects on learner outcomes." Money and learners' scarce hours are being allocated on tradition rather than evidence, and the field cannot even tell whether its dominant service line works.

What’s been tried and why it hasn’t worked

The rigorous studies that exist cluster on literacy instruction (13 of 22 studies), and those "found no positive effects of the tested strategies on any learner outcomes"; career-pathway approaches (bridge classes, integrated education and training) show promise on educational progress but "not consistently positive effects on employment and earnings." Two structural features of the studies undercut usefulness: interventions are usually bundles ("studies tested combinations of various strategies, making it difficult or impossible to isolate the effect of a particular strategy"), and outcomes are split — studies of instruction measured short-term academic skills but not credentials, employment, or earnings, while studies of career strategies measured the long-term outcomes but not skills, so nobody has traced the causal chain. The IES congressional briefing (2023) adds that "much of the existing research on adult education is outdated or sparse," that knowledge is imported from K–12 and higher education rather than adult-education settings, and that "the adult education system needs rigorous and reliable data and research." The population's own constraints — adults "balance jobs and family responsibilities," enter and leave programs continuously, and pursue varied goals — make conventional cohort RCTs with fixed pre/post testing expensive and attrition-prone, which is a plausible reason ESL instruction in particular has gone unstudied; that mechanism is inferred here rather than stated in the source and should be verified.

What would unlock progress

An evaluation architecture built for open-entry, open-exit, part-time adult programs: rolling-enrollment randomized designs, embedded low-burden outcome measures administered inside instruction (so attrition does not equal missing data), and administrative linkage to the National Reporting System and wage records so that skills, credential, and earnings outcomes are observed on the same learners. Adjacent precedents: pragmatic and stepped-wedge trial designs in health services research were invented for exactly this kind of continuous-intake service delivery; the digital learning platform experiments now emerging in K–12 (see `education-learning-platform-experiments-missing-demographics`) show how to embed randomization in routine instruction. IES's CREATE Adult Skills Network and the Adult Skills research program are the natural homes for such infrastructure.

Entry points for student teams

A research-design team could work with one local ESL provider to design and pilot a rolling-enrollment comparison of two instructional strategies with a five-minute embedded oral/written proficiency probe, documenting feasibility, attrition, and statistical power — the design itself is a contribution the field lacks. A data team could map, using public NRS tables, how enrollment, persistence, and measurable-skill-gain reporting vary across states, identifying where administrative data could support quasi-experimental studies. A team with linguistics or learning-science background could conduct a structured comparison of adult-ESL curricula in use against what the K–12 English-learner and second-language-acquisition literature would predict, producing testable hypotheses. Skills: program evaluation, applied statistics, second-language acquisition, adult-education practice.