Construct validity

Tuesday, November 26th, 2013

Reliability of a test as a questionnaire means that the questionnaire is measuring something in a reproducible way. Validity on the other hand means that the questionnaire is measuring what is intended to. In situations where other scales of the same or similar attributes are available, outcomes of an existing scale and the new one administered to a sample of people may be compared with their correlation indicating the validity of the new scale. However, if no other measure pre-exists, another approach is necessary called ´construct validity´.(1)

In a construct validation procedure the empirical data gathered with the new scale have to be compared with hypothetical expectations based on a theoretical construct. (2;3) If the expected relationship is found it may be appropriate to assume that the questionnaire is measuring what is intended to, i.e. is valid, but if not, the fault may lie with either the questionnaire or the hypotheses. An example of a construct validation procedure can be found in an article describing the validation of the gut feelings questionnaire by the Maastricht-Antwerp research group.(4)


(1) Streiner DL, Norman GR. Health Measurement Scales. A practical guide to their development and use. 3 ed. Oxford: Oxford University Press; 2006.

(2) Smith GT. On Construct Validity: Issues of Method and Measurement. Psychological Assessment 2005;17(4):396-408.

(3) De Vet HCW, Terwee CW, Mokkink LB, Knol DL. Measurement in Medicine. A practical guide. Cambridge: Cambridge University Press; 2011.

(4) Stolper CF, Van de Wiel WJ, De Vet HCW, Rutten ALB, Van Royen P, Van Bokhoven MA, et al. Family physicians’ diagnostic gut feelings are measurable: validation of a questionnaire. BMC Family Practice 2013;14(1).