Predictive Validity
Predictive validity is the degree to which test scores forecast performance on a criterion measured at a later point in time. With concurrent validity it forms the criterion-related branch of validity evidence; the two differ only in the gap between test and criterion administration.
Logic
A proficiency test claims that a learner who scores well will function well in a target domain — university study, professional communication, immigration. Predictive validity asks whether that promise holds. The standard estimate is the correlation between scores at time 1 (the predictor) and a criterion measured at time 2: end-of-year GPA, supervisor ratings, course completion, written or spoken performance in the target setting.
The most studied example in language testing is the relationship between IELTS or TOEFL scores at admission and subsequent academic performance. Reported coefficients are typically modest — often between .20 and .40 — partly because GPA itself is multiply determined, partly because admission cuts restrict the score range, and partly because the criterion has its own measurement error.
Methodological cautions
Three issues recur. Range restriction lowers the observed coefficient: students admitted on the basis of a test rarely span the full ability spectrum, so corrections for restriction of range are often applied. Criterion contamination distorts results when raters know predictor scores. Criterion deficiency — the criterion failing to capture the construct of interest — sets a ceiling on any coefficient regardless of test quality.
Within Messick's unified framework, predictive evidence supports one strand of the validity argument; it neither proves nor disproves construct validity on its own. A test that predicts a flawed criterion may still misrepresent the construct.
References
- AERA, APA, & NCME (2014). Standards for Educational and Psychological Testing. American Educational Research Association.
- Bachman, L. F., & Palmer, A. S. (2010). Language Assessment in Practice. Oxford University Press.
- Messick, S. (1989). Validity. In R. L. Linn (Ed.), Educational Measurement (3rd ed., pp. 13–103). American Council on Education / Macmillan.