EMPLOYEE selection is one of the most critical functions for organisational success. With an array of methods to choose from, it is important to discern which are most effective in predicting future job performance and which should be avoided.
Research from Frank L. Schmidt and John E. Hunter in The Validity and Utility of Selection Methods in Personnel Psychology highlights key findings from over 85 years of studies, shedding light on what works in employee selection and what does not while also addressing additional variables such as gender, age, and other factors.
What works in employee selection
General Mental Ability (GMA) tests is the most validated and effective predictor of job performance. GMA tests assess cognitive capabilities such as problem-solving, learning speed, and adaptability, making them especially relevant across different job types. According to Schmidt and Hunter, GMA tests have an average validity of 0,51, and when combined with other methods such as work sample tests or structured interviews, their predictive validity rises even higher.
Work sample tests involve candidates performing job-related tasks, giving employers a direct view of their abilities. When combined with GMA, work sample tests achieve a validity of 0,63, a 24% improvement over GMA alone. These tests are particularly useful for jobs requiring specific technical skills but may not be feasible for every role due to their high cost and complexity.
Structured interviews outperform unstructured interviews due to their consistency. By asking all candidates the same predetermined questions, structured interviews reduce interviewer bias and provide a better assessment of job-related competencies. Structured interviews, especially when used alongside GMA, have a validity of 0,63, making them a reliable predictor of job performance.
Integrity and conscientiousness tests, which assess an individual's likelihood to engage in counterproductive workplace behaviors, have a validity of 0,41. Combined with GMA, integrity tests increase predictive validity to 0,65. Conscientiousness, a personality trait reflecting reliability and diligence, also predicts job performance well, with a validity of 0,31. Both tests are valuable supplements to GMA tests, especially in roles requiring high levels of trust and responsibility.
What does not work
Unstructured interviews, which lack a standardised format, are subjective and often lead to inconsistent evaluations. Their low validity (0,38) highlights how reliance on these interviews alone can lead to poor hiring decisions. Unstructured interviews are susceptible to interviewer bias and do not provide a reliable measure of a candidate’s potential job performance.
Years of education and job experience are poor predictors of future performance. Education only has a validity of 0,10, indicating that simply having more schooling does not equate to better job performance. Similarly, years of job experience have a low validity of 0,18, with research showing that after five years, the benefits of experience diminish significantly. Therefore, focusing solely on education or experience may overlook critical attributes like cognitive ability or integrity.
Graphology, or handwriting analysis, is still used in some countries as a selection tool, but research shows it is a completely invalid method for predicting job performance. Graphology’s validity is near zero, meaning that it has no meaningful relationship to how well an employee will perform in their role.
Interests though it seems intuitive that people whose interests align with the job might perform better, interest-based selection methods have a validity of only 0,10. While job satisfaction might increase when interests align, job performance is driven more by skills, abilities, and personality traits than by interests alone.
Personal variables
Gender is often discussed in the context of hiring and selection. However, research shows that gender alone does not predict job performance, making it an irrelevant factor for selection purposes. While gender diversity is crucial for building inclusive teams, it should not influence individual hiring decisions. Instead, the focus should be on selecting the best candidate based on validated predictors such as GMA, structured interviews, and job-related skills.
Like gender, age is often assumed to correlate with experience, maturity, or even job stability. However, age has no predictive validity for job performance. According to the research, age does not correlate with higher or lower job success, making it an ineffective criterion for employee selection. In fact, in the United States, selecting based on age could lead to legal issues due to age discrimination laws.
Years of education and job experience, while often emphasised in traditional hiring, have low predictive validity. Education has a predictive validity of 0,10, while job experience has a validity of only 0,18, as noted earlier. Although these factors may indicate that a candidate has acquired basic knowledge or familiarity with a field, they do not guarantee strong job performance. Skills, ability to learn, and personality traits are much stronger indicators.
Biographical data and personality tests, which include questions about a candidate's past life experiences, can sometimes predict performance with moderate validity (0,35).
However, the incremental validity gained from adding biodata to GMA is small (0,01).
While biographical data may provide additional insights into a candidate’s character, it should not be heavily relied upon in isolation.
Similarly, personality tests, such as those measuring conscientiousness or emotional stability, can add some value, but their validity is typically lower than that of cognitive ability tests. Conscientiousness tests, for example, have a validity of 0.31 but work best when combined with other methods, such as GMA or integrity tests.
Similarly, personality tests, such as those measuring conscientiousness or emotional stability, can add some value, but their validity is typically lower than that of cognitive ability tests.
Conscientiousness tests, for example, have a validity of 0,31 but work best when combined with other methods, such as GMA or integrity tests.
Conclusion
Selecting employees is a complex process that benefits from using scientifically validated methods. General mental ability tests, work sample tests, structured interviews, and integrity tests are among the most reliable tools for predicting job performance.
On the other hand, methods such as unstructured interviews, graphology, and selecting based on education or experience alone are less effective.
Moreover, factors like gender and age have no direct correlation with job performance and should not influence hiring decisions. By focusing on proven selection methods and avoiding invalidated ones, organisations can significantly improve their hiring success and, ultimately,their business performance
- Nguwi is an occupational psychologist, data scientist, speaker and managing consultant at Industrial Psychology Consultants (Pvt) Ltd, a management and human resources consulting firm. — https://www.thehumancapitalhub.com or e-mail: mnguwi@ipcconsultants.com.