Types Of Validity: Criterion Validity And Construct Validity

The concept of validity has evolved over the years. However, experts were convinced that a test was valid for everything related to this test (2). As we have seen in other articles, there are four types of validity:

• content validity
• predictive validity
• criterion validity
• construct validity

In this article we will look at criterion validity and construct validity.

Criterion validity

Criterion validity considers whether the scores of a test actually evaluate the questions of the test. To estimate this type of validity, the makers of the test run the test and correlate it with the criteria. The criteria are measuring instruments that the makers have evaluated in advance.

This type of validity is similar to predictive validity. However, there are two major differences between these two validities (1):

• In criterion validity, the test makers obtain the test measurements and the criteria at the same time.
• The main goals of predictive validity and criterion validity are different. The main application of criterion validity is to find tests, which can replace other procedures that are less suitable for various reasons. For example, a collective intelligence test can be comparable to an individual intelligence test.

However, the main problem with this type of validity is that it is difficult to find tests that serve as valid and reliable criteria.

Construct validity

Construct is a hypothetical concept that is part of the theories that try to explain human behavior. These are, for example, intelligence and creativity.

This type of validity answers the following question: “How can the test score be explained psychologically?” The answer to this question can be seen as developing a ‘mini-theory’ about the psychological test.

Construct validity therefore consists of obtaining evidence to substantiate whether the observed behavior during a test are (a number of) indicators of the construct (1). The construct validation process includes the following (1):

• Formulation of hypotheses and relationships between construct elements, other construct theories and other external constructs.
• Selection of subjects or tests (indicators) representing concrete manifestations of the construct.

• Collecting data.
• Establishing a relationship between the data and the hypothesis. It is then examined to what extent the data can be explained by alternative hypotheses.

The procedure to establish construct validity

There are several procedures to establish construct validity (1 ), namely:

• Based on the theory at the time of the test , the psychologist derives certain hypotheses about the expected behavior of people who receive different test scores.
• They then collect data that confirm or deny these hypotheses.
• On the basis of the collected data, they decide whether the theory sufficiently explains the results. If not, they revise the theory and repeat the process until they get a more accurate explanation.

In this sense, the validation process is constantly being adapted and refined. In reality, the results of the studies don’t really “prove” or “validate” the whole theory. This is because you can never fully demonstrate a ‘construct’.

All you can do is just accept it as the best definition you can work with. There are three possible reasons why the results are negative (1, 3):

• First, the test may not actually measure the ‘construct’. In any case, it doesn’t measure what you want it to measure, but it does measure something.
• Secondly , the theoretical framework may not be correct. Therefore, any conclusions may also be incorrect.
• Finally, the design of the experiment did not allow the hypothesis to be tested appropriately. Design flaws are usually the easiest flaw to spot. However, pinpointing the error accurately is a trickier task. Of course, the ambiguous interpretation of the negative results is a drawback of the construct validity procedure.

The criterion validity and construct validity shed light on the validation of a test. Therefore, there are some aspects to consider during validation.

Practical implications of validating tests

Psychologists using tests should consider these implications when it comes to the four validation types:

• Before you can make decisions about individuals or groups, you need to collect all available information about the test.
• To predict or select, you need to validate the test in the specific situations in which you will use it.
• In any situation, the psychologist must keep in mind that new information constantly modifies theories about the nature of the traits and everything they measure.

Validity helps us analyze psychological tests. As you know a test is the more valid the better (without taking into account other variables). Unfortunately, this is not always the case when it comes to research, as other criteria come into play, such as economic factors and availability.