Please enable JavaScript.
Coggle requires JavaScript to display documents.
Validity in Language Assessment Carol A. Chapelle - Coggle Diagram
Validity in Language Assessment
Carol A. Chapelle
Background
Open science increases sharing of instruments
higher need for careful validation
“types of evidence” does not equal to “types of validity”
Key concepts
Test content – materials that elicit performance
Performance – the answers or responses learners give to prompts
Reliability – how stable or consistent the scores are (about scores, not the tests themselves).
Construct – the meaningful way we explain or interpret the consistency of performance.
Target context – the area or domain where scores are used to make generalizations.
Test use – the actions or decisions made based on scores.
Consequences – the results or effects that come from test use.
Warrant – a statement that must be true in order to support an inference.
Key Issues for Validation
Generalization
Explanation
scores reflect the intended construct
Extrapolation
score meaning extends to target context
scores reflect performance consistency (reliability)
Domain Definition
appropriate observations of performance obtained
scores accurately summarize relevant performance
Evaluation
Recommendations for Practice
(Re)learn basics
avoid common assessment/reporting errors
Examine and build on prior validity arguments in the field