ERIC Number: EJ1275887
Record Type: Journal
Publication Date: 2020
Pages: 37
Abstractor: As Provided
ISBN: N/A
ISSN: EISSN-2295-3159
EISSN: N/A
Available Date: N/A
Ensuring Content Validity of Psychological and Educational Tests -- The Role of Experts
Beck, Klaus
Frontline Learning Research, v8 n6 p1-37 2020
Many test developers try to ensure the content validity of their tests by having external experts review the items, e.g. in terms of relevance, difficulty, or clarity. Although this approach is widely accepted, a closer look reveals several pitfalls need to be avoided if experts' advice is to be truly helpful. The purpose of this paper is to exemplarily describe and critically analyse widespread practices of involving experts to ensure the content validity of tests. First, I offer a classification of tasks that experts are given by test developers, as reported on in the respective literature. Second, taking an exploratory approach by means of a qualitative meta-analysis, I review a sample of reports on test development (N = 72) to identify the common current procedures for selecting and consulting experts. Results indicate that often the choice of experts seems to be somewhat arbitrary, the questions posed to experts lack precision, and the methods used to evaluate experts' feedback are questionable. Third, given these findings I explore in more depth what prerequisites are necessary for their contributions to be useful in ensuring the content validity of tests. Main results are (i) that test developers, contrary to some practice, should not ask for information that they can reliably ascertain themselves ("truth-apt statements"). (ii) Average values from the answers of the experts which are often calculated rarely provide reliable information about the quality of test items or a test. (iii) Making judgements about some aspects of the quality of test items (e.g. comprehensibility, plausibility of distractors) could lead experts to unreliable speculations. (iv)When several experts respond by giving incompatible or even contradictory answers, there is almost always no criterion enabling a decision between them. In conclusion, explicit guidelines on this matter need to be elaborated and standardised (above all, by the AERA, APA, and NCME "Standards").
Descriptors: Content Validity, Psychological Testing, Educational Testing, Student Evaluation, Test Construction, Expertise, Interrater Reliability, Test Items, Standards, Evaluation Criteria, Meta Analysis
European Association for Research on Learning and Instruction. Peterseliegang 1, Box 1, 3000 Leuven, Belgium. e-mail: info@frontlinelearningresearch.org; Web site: http://journals.sfu.ca/flr/index.php/journal/index
Publication Type: Journal Articles; Information Analyses
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A
Grant or Contract Numbers: N/A
Author Affiliations: N/A