Predictive validity is the extent to which a test or screening predicts scores on some standard measure such as job performance.
The validity of a test to predict job success is the correlation between test scores and some measure of job performance (for example, supervisor performance ratings). A test has predictive validity if the observed correlation is statistically significant.
Predictive validity shares similarities with concurrent validity in that both are generally measured as correlations between a test and some performance measure. But there are important differences between the two.
In a study of concurrent validity the test is administered at the same time as the performance assessment data is collected. This is a common method of developing validity evidence for employment tests: A test is administered to the current employees, then a rating of those employees' job performance is obtained (often in the form of a supervisor rating). Note the potential problem here: The incumbent employees are likely to be a more homogeneous and higher performing group than the applicant pool at large.
In a study of predictive validity, the test scores are collected first; then at some later time the performance assessment is collected. Tests are administered to job applicants, and then after those individuals work in the job for a year, their test scores are compared to their first year job performance scores. Thus predictive validity provides more useful data about the test's value because it is more like the real situation in which the test will be used. After all, most tests are administered to find out something about future behavior.
Predictive validity is considered to be a more powerful rating of a selection tool than concurrent validity. It is also much more expensive and time consuming to conduct this type of research. As a result, fewer selection tools available provide users with information on predictive validity.
Description of Research of Predictive Validity
A follow-up questionnaire was sent to companies that had been using our test for more than one year. In addition to some basic comparison data, the managers were asked whether the recommended candidates were still employed, whether any of them failed, whether they performed more effectively than salespeople hired using other methods, and whether the candidates were ranked in the top half or the bottom half of their sales organizations. Managers were also asked if they met the recommended conditions for hiring specified in the test report, and if they provided professional sales training to the candidates. In this way a distinction could be made between the performances of the new hires who were properly managed and trained and those who were not.
Results of the Research to Evaluate OMG's Candidate Screening
One year after testing, 95% of the candidates who were recommended for hiring were still employed by their respective companies. Only 5% of the candidates who were recommended failed and most of the companies that hired those salespeople failed to meet our recommended conditions for hiring. When we looked at the failure rate in companies that met the conditions for hiring, the failure rate dropped to less than 1%. In those companies where the conditions for hiring were met, 92.3% of the salespeople hired with this test outperformed those hired previously.
When hiring conditions were met and professional training was provided, the percentage of salespeople hired with this test who outperformed previous hires jumped to 99.9%. When the recommended conditions for hiring were not met, 44% still outperformed those hired previously and that number jumped to 78% when professional training was provided.
More recent follow-up, collected in October of 2006, substantiated earlier studies and showed that 92% of those recommended and hired were ranked in the top half of their respective sales forces after just one year. In contrast, 8% of those not recommended and hired were ranked in the top half, 16% of that group ranked in the bottom half, and 75% quit or were terminated.
These results demonstrate that this screening is clearly able to accurately distinguish between individuals who will sell successfully for a particular company and those who will not.