The systematic appraisal of a specific language programme or course (its design, delivery, and effects) to inform decisions about its continuation, revision, or expansion. Where curriculum evaluation often targets the framework, programme evaluation targets a bounded instance: a particular EAP course at a particular institution, a national initiative, an in-company training scheme.
Lynch (1996, Language Program Evaluation: Theory and Practice, Cambridge University Press) is the first book-length applied-linguistics treatment. Lynch proposes a context-adaptive model: evaluation design is not chosen off the shelf but built from the programme's audience, purpose, and constraints. He argues for shifting beyond pre/post test gains toward investigation of programme process through qualitative, naturalistic inquiry alongside quantitative measures.
Kiely and Rea-Dickins (2005, Program Evaluation in Language Education, Palgrave Macmillan) consolidates the field a decade later. Drawing on case studies across school, tertiary, and workplace contexts, they map the evaluator's stance (insider, outsider, commissioned), the political economy of evaluation, and the methodological repertoire available: questionnaires, observation, interviews, document analysis, test data, classroom recordings.
Norris (2009, "Language program evaluation", in Doughty & Long eds., The Handbook of Language Teaching, Wiley-Blackwell) and the special issue of Language Teaching Research 13(1) he edited the same year argue that programme evaluation in ELT remained underdeveloped relative to assessment research, and call for evaluation embedded in routine programme operation rather than commissioned only at moments of crisis or funding review.
Two purposes drive evaluation design and have been distinguished in the wider evaluation literature since Scriven (1967):
The same instrument can serve either purpose; what differs is timing and audience. A mid-course questionnaire treated as formative becomes summative if its results are filed for the funder rather than used to revise teaching.
A defensible evaluation specifies, in advance:
Lynch and Norris both emphasise that utilisation — whether findings actually feed decisions — is the strongest test of an evaluation, not its methodological elegance.
Sponsors, administrators, teachers, learners, and external regulators bring incompatible criteria. A programme rated successful on completion rates may be judged costly by funders, narrow by employers, or stressful by teachers. Kiely and Rea-Dickins treat this incompatibility as the substance of evaluation rather than noise to be averaged out: surfacing whose criteria count, and why, is part of the work.