What Do Tests in School Actually Measure?

The Importance of Having a Clearly Defined Psychological Construct

Psychological tests are used to measure psychological constructs. This statement seems fairly obvious. After all, isn’t that the whole point of creating the test? They are intended to tell us something about an aspect of ourselves that cannot otherwise be quantified, and our score on a test is interpreted as our level of that construct, either compared to others, or compared to some absolute standard. But what happens when the only thing a test measures is the ability to do well on that test?

Let’s start by thinking about the type of test that we’ve all had the most experience with — evaluations in school. These are used to determine our grades in classes, our CGPAs, whether or not we are accepted into university programs and possibly whether or not we are hired for a job. With so many of our opportunities riding on these tests, they must be measuring a fairly important construct, one that determines our abilities as well as our future successes. Have we ever stopped to wonder what this construct is?

For some, the most obvious answer might be that an academic test is measuring some inherent level of intelligence. But if this were the case, why would we need an exam for every class? Why wouldn’t we just take one intelligence test and have that determine our grades and our admissions into programs? Maybe a test is measuring our ability to learn the material taught in a course. But we could just as easily argue that it is measuring the teacher’s ability to teach it, or simply measuring the compatibility of the teacher’s teaching style with the student’s learning style. Alternatively, we could argue that test results reflect our ability to block out distractions while studying, our ability to manage stress, the difficulty of the test questions, the number of stressful events that occurred in our lives the week prior to the test, and probably many more. With all of these factors playing a role, how do we (the people taking the tests) and others interpret the results? And what effects do these possible interpretations have?

There are several different ways in which a person’s performance can be interpreted. Successes and failures can be attributed to either internal or external factors, and those factors may be either stable or unstable. For example, if a student fails a test and the teacher reasons that this is because the student was going through a lot of stress in the days leading up to the test (an unstable external factor), the consequences may not be so bad for the student (the teacher may let them rewrite the test). On the other hand, if the failure is attributed to a stable internal factor (the student is simply not smart), the teacher may encourage them to transfer into an easier class, or simply have lower expectations for them the next time. Similarly, the way the student interprets having done poorly will influence the way they feel about their abilities. If they think that the test was simply very difficult, then they may study harder for the next one, and they will likely be able to maintain a positive image of their academic abilities. On the other hand, if they attribute it to something internal (not being good at the subject), they may experience a drop in self-esteem, and they may avoid other courses and possible career paths related to that subject.

These very different outcomes demonstrate the importance of construct validity — having a clear idea of what it is we are measuring, and what observable behaviours are related to that construct. Test grades may be very useful in determining the extent to which a class has, on average, learned the material of a course, and they may be useful, to an extent, in predicting job performance. However, different jobs likely have different definitions of success — some of these may be predicted by doing well on academic tests, while others may not. If we want to know whether a test will be predictive of certain behaviours, it is important to know what construct those behaviours are reflecting, and make sure the test is an accurate measurement of that construct. Despite all this, people still believe that test grades are important, and interpret the results as meaningful. In other words, even though this construct is somewhat unclear, it can still have very real effects on the way people perceive one another, and the way people perceive themselves.

References

Taylor, D. M., & Moghaddam, F. M. (1994). Stereotypes, Attributions, and Discrimination. In Theories of Intergroup Relations: International Social Psychological Perspectives (pp. 159–177). Westport, CT: Praeger Publishers.