Construct — the underlying ability or trait that a test is designed to measure — e.g., ‘communicative competence’ or ‘reading comprehension’ — central to test validity arguments.
Definition
The underlying ability or trait that a test is designed to measure — e.g., ‘communicative competence’ or ‘reading comprehension’ — central to test validity arguments.
In Depth
The underlying ability or trait that a test is designed to measure — e.g., ‘communicative competence’ or ‘reading comprehension’ — central to test validity arguments.
In-Depth Explanation
A construct is a theoretical, unobservable psychological attribute that a test attempts to measure. You cannot directly observe “reading comprehension” or “communicative competence” — only performances on tasks that a test designer claims reflect those abilities.
| Construct concern | Definition | Example |
|---|---|---|
| Construct definition | Theoretically grounded specification of the target ability | “Grammatical competence” vs. vague “English proficiency” |
| Construct validity | Extent to which a test measures what it claims | Does the JLPT measure Japanese ability, or test-taking ability? |
| Construct-irrelevant variance | Score variance from factors other than the construct | Test anxiety, typing speed, cultural familiarity affecting scores |
| Construct underrepresentation | Failure to sample the full construct domain | A “Japanese proficiency” test measuring only reading, neglecting speaking |
For language testing: The JLPT’s construct is implicitly reading and listening proficiency — it does not measure speaking or writing. IELTS and TOEFL attempt broader constructs of academic English proficiency. The question of whether communicative competence is even measurable by standardised tests (Hymes 1972) remains an active debate.
History
Classical test theory assumed tests had fixed reliability without addressing what they measured. Cronbach and Meehl (1955) introduced construct validity as a formal concept, distinguishing it from content and criterion validity. Messick (1989) proposed a unified validity framework placing construct validity at the centre of all validity arguments — subsuming all prior validity types. Kane (2006) developed the argument-based validity framework, specifying the inference chain from test performance to construct interpretation.
Common Misconceptions
- “High reliability means the test measures the right construct.” Reliability is a precondition for validity but does not guarantee construct validity. A test can measure something very consistently while measuring the wrong thing.
- “Construct and skill are synonymous.” Skills (reading, writing) are constructs, but constructs can also include more abstract attributes (language aptitude, metacognitive strategy use, pragmatic competence) that cross-cut the four skills.
- “Low scores mean the learner lacks the target ability.” Low scores may reflect construct-irrelevant factors — test anxiety, unfamiliar format, cultural background knowledge gaps — rather than deficit in the target ability.
- “All language tests measure the same construct.” Different tests operationalise different constructs of “language ability,” making direct score comparisons across tests problematic.
Social Media Sentiment
Test validity discussions appear in language teaching communities on Reddit and Twitter when assessing JLPT, IELTS, and teaching certification exams. Learners rarely use the term “construct” but intuitively complain about construct-irrelevant variance (“the JLPT doesn’t test speaking, so it doesn’t really measure my Japanese”) and construct underrepresentation. Language testing researchers on academic Twitter debate construct definitions for new test formats (game-based, computer adaptive).
Last updated: 2026-04
Practical Application
- Interpreting test results: Ask what construct a test actually measures. A high JLPT N2 score indicates strong reading and listening proficiency in Japanese — but says nothing about speaking.
- Designing assessments: Specify the construct before designing any classroom test. What exactly are you measuring with this assignment? Clarity reduces construct underrepresentation.
- Recognising construct-irrelevant variance: If test anxiety or time pressure significantly affects your performance, your score may underrepresent your actual ability on the target construct.
- Reading SLA research: When studies claim to measure “proficiency,” examine how that construct is operationalised.
Related Terms
See Also
Sources
- Messick, S. (1989). Validity. In R. L. Linn (Ed.), Educational Measurement (3rd ed., pp. 13–103). American Council on Education. Foundational unified validity framework.
- Kane, M. T. (2006). Validation. In R. L. Brennan (Ed.), Educational Measurement (4th ed., pp. 17–64). American Council on Education. Argument-based validity framework.
- Cronbach, L. J., & Meehl, P. E. (1955). Construct validity in psychological tests. Psychological Bulletin, 52(4), 281–302. Original formulation of construct validity.