Sign up for our Monthly Highlights newsletter
Don’t miss the roundup of our newest and most distinctive insights
September 8, 2022 | Klein Consultants
The most common application of predictive validity in a real-world setting is in the process of selecting students for university. Most universities accept students based on their high school grade point averages in an effort to find the brightest and most dedicated students.
The basic assumption in this process is that a high-school student with a high grade point average will achieve high grades at university. Hundreds of studies have been conducted to test the predictive validity of this approach. To accomplish this, a researcher compares the grades earned after the first year of study to the high school grade point averages.
A high correlation indicates that the selection procedure was flawless, while a low correlation indicates that something was wrong with the approach. Most studies show a strong correlation between the two, and the method’s predictive validity is high, though not perfect.
Intuitively, this makes sense; previously outstanding students may experience homesickness or decide to spend the first year drinking beer. Underachieving college students, on the other hand, frequently become dedicated, hardworking students in the relative freedom of the university environment.
In the example of university selection, this approach does not test students who did not attend university due to low grades, personal preference, or financial concerns. This creates a gap in the data, and because the method is dependent on an incomplete data set, the researchers must always make some assumptions.
If the students with the highest grade point averages perform better after their first year of university, and the students who barely make it do the worst, researchers predict that non-attendees will perform even worse. This downward extrapolation may be incorrect, but predictive validity must take such assumptions into account.
Furthermore, scholars argue that using predictive tools like standardized test scores can exacerbate racial and class disparities in postsecondary educational access. Scholars have also observed that standardized admissions test scores measure cognitive abilities, whereas college success necessitates a variety of cognitive and non-cognitive characteristics.
Despite this shortcoming, predictive validity is still regarded as a highly effective measure of statistical accuracy. It is regarded as the most important measure of quality in many fields of research, and researchers are constantly looking for ways to maintain high predictive validity.