Psychological tests are the tools used by psychologists to measure the variables they are interested in knowing. Just like a scale is used to measure weight, in psychology we have instruments that allow us to know certain parameters concerning anxiety, emotions and personality, among others. However, these tests are only useful if they meet quality standards.
So, each measuring instrument is based on the formula X = V + E. Where X represents the measurement given by the test, which is the sum of V, the actual score, plus E, the measurement error. This simple equation shows us a way to investigate the quality of psychological tests. And through it, we can build instruments where X and V are as similar as possible in all subjects. On the other hand, it is important to mention that the study of how to measure psychological constructs is called psychometrics.
What indicators should we consider in determining whether psychological tests are good or bad? To answer these questions, let's talk about aspects to consider when making a good instrument. In addition, we will also explain the concepts of validity and reliability, which will show us the quality of the test. Go further.
How to build psychological tests?
Building a test is a laborious process that requires many hours of work and research. As a preliminary step in the construction process, three essential questions must be answered:
- What do we want to measure with him?
- Who do we want to measure?
- What are we going to use it for?
The first question allows us to know the variable under study. It seems trivial to define what you want to measure, but if it is not done correctly, it can lead to confusion. A historical example of this has occurred with intelligence: different instruments have been developed to measure it but no one has defined it. We still suffer the consequences, having many different definitions of construction intelligence and different tests that measure different things.
A key aspect when it comes to defining a construction to measure is knowing what we are measuring. Psychological constructs can not be observed directly (see, for example, anxiety), but they can be measured through the behaviors they generate. It is therefore necessary to explain the behaviors by which the studied variable manifests itself.
The second question is useful for adapting the test to the population we want to measure. Obviously, we can not do psychological tests that are valid for all ages and all conditions. It is therefore essential to know the target population and to adapt the instrument to the needs and qualities of the subjects.
All tests are built with one goal in mind and can even be multiple, such as diagnosis of a disorder, choice of subjects, research, etc. The third question is to focus the instrument on its usefulness. Even if two tests are trying to measure the same thing, they can be very different if your goal is different. For example, if we are trying to measure intelligence, we will not develop the same tests if we want to detect gifted children or if we want to detect deficiencies.
Thus, the answers to these three questions are the foundation of any psychological test. It will be essential to study them in depth if we want to build a valid and reliable instrument.
The quality of psychological tests
In psychometrics, when it comes to measuring whether a test performs its function, there are two basic indicators. These indicators are validity and reliability. Throughout history, a multitude of statistical formulas have been designed to calculate and test the quality of the tests. But what is validity and reliability?
The validity of a test
When we talk about the validity of a test, we are talking about the ability of this test to measure what it wants to measure. In other words, if we want to measure a person's degree of anxiety, the test will be valid if it measures anxiety and only anxiety. It may seem somewhat absurd that a test measures something that we do not want to measure, but when the construction is poorly defined, it is possible that the construction of the instrument results in very low validity.
To measure the validity of a test, there are several statistical resources. The most common are the correlation of the test that we want to measure with another, whose validity has already been verified previously, and evaluate it through various expert judges and see how well these judges agree.
The reliability of a test
The degree of reliability is the scale that indicates the accuracy of a test when measuring. That is, a test will be reliable if we pass the same instrument twice to the same person and get the same result. If each time we pass the test, the result is different, it would mean that the measurement error would be so high that it would distort the results. It's as if every time we put the same object on a scale, it gives a different weight.
There are also different statistical strategies for measuring reliability, but the best known is to pass the same test twice to a group of people and see the correlation between the first and the second measure. A high correlation would show us a great reliability and that the test fulfills its function.
Finally, psychological tests are used in all branches of psychology, most applied to research. It is therefore essential that its construction be fully controlled so that the results are valid and reliable.