The Foundations of Construct Validity
Since the mention of construct and construct validity in this course I have become interested in the origins of these concepts in the psychological literature. I think it is important to be aware and inform ourselves not only about the history of psychological testing in so far as it relates to such things as the origins of intelligence testing but also as it relates to the history and philosophical underpinnings of key concepts that today form a fundamental and seemingly indispensable component of psychological testing as a whole. I came across one particular article that outlines the foundations of construct validity and is deemed by some to be one of psychology’s classic papers. This paper is titled: Construct Validity in Psychological Tests (Cronbach and Meehl, 1955). A summary and or critique of this article is beyond the scope of this blog post. The aim of this post, rather, is to share some of the important points and ideas that I drew from this article in the hope that it will spark interest and curiosity among my peers and incite further inquiry into this subject matter. I will share these points and ideas below in a question and answer format. Note that the answers are not meant to be comprehensive but are meant to give a brief idea of some of the possible answers to such questions from the perspective of the authors.
Why must we rely on constructs?
The authors argue that without the use of a construct, a test can be used to predict a particular behavior when there is a known correlation between the test and the behavior. However, if we wish to predict behavior in diverse or novel situations, where the correlation between the test and the behavior is unknown, then we must rely on a construct. This construct allows us to accumulate information and knowledge around it, and if general enough, can help guide predictions in novel situations.
How does a test and its underlying criterion each relate to construct validity?
A test constructed on the basis of a specific criterion can nonetheless prove to have greater construct validity than the criterion itself. This reality has to do with the nature of scientific investigation: “We start with a vague concept which we associate with certain observations. We then discover empirically that these observations covary with some other observation which possesses greater reliability or is more intimately correlated with relevant experimental changes than is the original measure, or both” (Cronbach and Meehl, 1955). The authors use the example of temperature. (Note that Professor Stotland used a similar example early in the semester). The construct of temperature arose from the experience of objects that were hotter to the touch than others. However, the criterion of the human temperature sense is fallible and since the discovery of a statistical relation between expansion and sensed temperature along with other factors such as a theoretical structure describing the relationship of mercury expansion to heat have rendered the criterion of human temperature sense peripheral.
How does the use of a construct facilitate or hinder the convergence of findings from different studies?
In order to compare and contrast findings from different studies, it is essential that the different users of a particular construct (in this case the authors of the studies), accept the same nomological network. One must first accept the network or theory behind the construction of the test before asking questions of validity and comparing findings from different studies. Otherwise, questions of validity are rendered moot. Moreover, the authors of different studies supposedly using the same construct will be talking past each other and any comparisons between their findings will be invalid.
Concluding Remarks
Reading this article helped me gain further insight into the notions of construct and construct validity as they are used in psychological testing literature. Though not explicated in this post these concepts are founded on notions from the philosophy of science. Gaining a full understanding of these concepts requires a close reading of Cronbach and Meehl (1955) and further readings from philosophy of science. I hope that this blog post helps motivate further reading into these areas.
Reference
Cronbach, L. J., & Meehl, P. E. (1955). Construct validity in psychological tests. Psychological bulletin, 52(4), 281.
