ITC Guidelines for Translating and Adapting Tests (Second Edition)
Confirmation guidelines – ITC Guidelines for Translating and Adapting Tests (Second Edition)

ITC Guidelines for Translating and Adapting Tests (Second Edition) C-1: Sample size and representativeness for validation

What does C-1 require?

Select a sample whose characteristics are relevant to the intended use of the adapted test, and large enough and relevant enough for the planned empirical analyses. Stability of the statistics should drive the choice, because factor analytic and item bias techniques only give meaningful results with adequate numbers, and the size required grows with model complexity and weaker data quality. The sample should also represent the population the test is meant for; where that group differs in level or spread from the source group, consider a matched source subsample or compare standard errors of measurement instead of reliability. Practice notes put the floor at about 200 respondents in each language version for item bias work, 500 for IRT and model fit work, and around 300 or more for studies of factorial structure.

Maintained by Gerard Blokdyk

Other controls in Confirmation guidelines – ITC Guidelines for Translating and Adapting Tests (Second Edition)

Query this from an agent

The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.