Dimensionality and construct validity of language tests

Abstract

At times 'unidimensional' tests have been characterized as being unduly constrained in what they can measure, and ultimately as lacking construct validity for communicative language assessment since communicative lan guage ability is held to be 'multidimensional' in nature. This simulation study considers the effects on statistical measures of test dimensionality that result from systematic sampling variation in both a single and a double-trait assessment model. Results are offered as evidence that there are both psychological and psychometric states of test dimensionality that are distinct, and that psychometric unidimensionality is not at all exclusive of psycho logical multidimensionality. Moreover, it is argued that psychometric uni dimensionality may often be rather more reflective than detractive of construct validity, even when complex multidimensional constructs underlie performance on the test.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 63 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.