# What is Meant by Predictive Validity? Let's Figure it Out Together

Predictive validity is a term in psychometrics that calculates the future behavior of a person on the basis of his/her current cognitive scores with respect to a criterion measure. With some examples of predictive validity, let's understand it better.

PsycholoGenie Staff

Last Updated: Jul 14, 2017

Validity means the accuracy for which something is tested, it must correspond to the outcome or conclusion of that test.

You can calculate and evaluate the scores of a student writing a test. But how can you predict what that student will score in the future? As kids, it often so happened that our parents, specifically the one parent who took care of our studies, could predict our scores for the upcoming tests. And mostly they would get their predictions right. Darn, I always wondered how could they do that! Well, to put it simply, they could predict our scores combining the knowledge of our past test scores and how much we had studied.

They knew exactly how much of the subject we knew, since they taught us. They also knew our weak areas, where we were likely to commit mistakes, and what were our strong points. Based on all of these factors, they did their magical mind calculations and predicted our scores, which were always close to our real test results.

They knew exactly how much of the subject we knew, since they taught us. They also knew our weak areas, where we were likely to commit mistakes, and what were our strong points. Based on all of these factors, they did their magical mind calculations and predicted our scores, which were always close to our real test results.

Definition of Predictive Validity in Research

Predictive validity is the extent to which one test can be used to predict the outcome of another on some criterion measure. It is used in psychometrics (the science of measuring cognitive capabilities).

Predictive validity is similar to concurrent validity in the way it is measured, by correlating a test value and some criterion measure. The difference between the two is that in concurrent validity, the test and the criterion measure are both collected at the same time, whereas in predictive validity, the test is collected first and the criterion measure is selected later.

Example:

Concurrent validity is a common method for taking evidence tests for later use. Employment tests require the working employees of the company to give a test which is then correlated with the ratings of their job performances, which was taken independently.

In predictive validity, tests are conducted for the new applicants. After the newly joined applicants finish one year on the job, these test scores are correlated with the job performances of their first year. Any test that is taken is basically done to solve future issues. Hence, predictive validity is more useful in providing data related to test validity.

In predictive validity, tests are conducted for the new applicants. After the newly joined applicants finish one year on the job, these test scores are correlated with the job performances of their first year. Any test that is taken is basically done to solve future issues. Hence, predictive validity is more useful in providing data related to test validity.

Another example:

Many educational institutions while selecting students, assess them according to their high school scores. This helps them find students worthy for competing against other students of their college.

Predictive validity comes into use when the management of an educational institute wants to check out if their selection procedure was really optimal. They then check the selected students' scores of the first year of studies and correlate them to the scores of their high school.

A high correlation means students are consistent in their performance, and the selection procedure turned out to be successful. However, if the correlation is low, then the selection procedure needs to be changed. Even though predictive validity is not a perfect validity test, but it is quite useful.

To find out the predictive validity of anything, you will need to calculate the predictive validity coefficient.

Predictive validity comes into use when the management of an educational institute wants to check out if their selection procedure was really optimal. They then check the selected students' scores of the first year of studies and correlate them to the scores of their high school.

A high correlation means students are consistent in their performance, and the selection procedure turned out to be successful. However, if the correlation is low, then the selection procedure needs to be changed. Even though predictive validity is not a perfect validity test, but it is quite useful.

To find out the predictive validity of anything, you will need to calculate the predictive validity coefficient.

How to Calculate Predictive Validity Coefficient

Validity Coefficient:

Where, r = correlation coefficient 'r' between X and Y

X = predictor

Y = criterion

Maximum Value of the Validity Coefficient:

**r**_{XY}Where, r = correlation coefficient 'r' between X and Y

X = predictor

Y = criterion

Maximum Value of the Validity Coefficient:

**r**_{XY}**max**= SQRT (r_{XX}) (r_{YY})Predictive validity is that tool which helps you predict the future possibilities, even though not very foolproof, but it gives you a surety of your plans and their outcome. The weakness in this validation is that it fails to consider the complete data. Like if in the above university selection procedure, we apply predictive validity, then the students who failed to attend the tests with which the correlation is to be made, those students will not be counted, and hence, we get the result from an incomplete set of data.