SIOP Members Create Tool for Removing Biases in Personality-Based Job Analysis
by Stephany Schings, Communications Specialist
An ever-increasing number of companies today are turning to personality testing to determine whether an applicant is right for a job.
However, what happens when the personality traits determined as necessary for success at a particular job are inaccurate?
In their paper entitled “Using Web-Based Frame-of-Reference Training to Decrease Biases in Personality-Based Job Analysis,” which is currently awaiting its publication in the forthcoming journal Personnel Psychology, SIOP members Herman Aguinis, Mark Mazurkiewicz, and Eric Heggestad explain how natural biases toward a person’s own personality characteristics can cause him/her to rate those characteristics as necessary for a job even when they are not.
“We hypothesized that people tend to say ‘you need what I have’ when determining what characteristics are needed for a job,” explained Aguinis, who was the lead author on the paper. “Results showed exactly that: for a very large number of traits, peoples’ personalities were very highly correlated to what characteristics they felt were required for the job.”
For example, Aguinis explained, if a person is extroverted, he/she may believe that extroversion is necessary for a job even if this is not necessarily true. This may lead to problems in employee selections, he continued.
“We argue that some of the personality-based tools, because people have these biases, will lead to hiring more people just like the person who analyzed the job but not necessarily better workers,” Aguinis said.
To help solve this problem, the three developed a study using a frame-of-reference training program to help raters understand and correct their biases, which they hypothesized would then lead to more accurate job analyses.
Aguinis, Mazurkiewicz, and Heggestad identified a sample of administrative assistants and supervisors in a large city government. They randomly assigned half of the participants to their Web-based frame-of-reference training program, which taught participants about their biases and asked them to complete the job analyses without considering their own personality characteristics and basing the answers on what the average person would need to complete the job successfully. The others completed the job analyses as usual.
The online training session program lasting only 15 minutes decreased bias by 75% for administrative assistants and 70% for supervisors.
Job analysis is already a fundamental tool in human resources management. This information is used for staffing, training, performance management, and various other HR activities.
The problem arises, Aguinis said, when this job analysis is skewed or biased, providing inaccurate data about a job. Aguinis said the accuracy of the job analyses is a determining factor in providing effective human resources management.
“Hundreds of companies sell off-the-shelf tests that measure job applicants’ personalities. There’s such a frequency of use of personality tests everywhere,” Aguinis said. “We’re saying that if we use personality tests, we need to first develop good methods for determining what are traits needed for a job.”
Mazurkiewicz said this relevance to practitioners makes the group’s research even more important.
“I think our paper is interesting not only for research but also for practitioners, even if they don’t use the same personality-based job analysis tool we did,” Mazurkiewicz said. “It really comes back to the fact that before you do any job analysis you should train the raters on how to use the tool.”
“Our research was conducted using real employees on a real job in a field experiment,” he said. “It’s research, yet this is something practitioners can use because it will help them make choices in the selection of their employees.”
That practicality lies in the method of the research, he added: “Our approach follows the scientist–practitioner model: It’s research that is rigorous and relevant. You do practice that’s based on solid research and research focused on practice.”
Mazurkiewicz added that the paper is only the beginning of what he hopes to be a growing field of research.
“It really was a good learning experience,” he said. “I truly believe this is an area that needs more research. There’s a lot of research that still needs to be done, but this is a good first step.”
In addition to the pending publication in Personnel Psychology, a previous version of the group’s paper was also presented in 2006 at SIOP’s conference in Dallas.
Herman Aguinis, a Fellow of SIOP, is a professor of Management and the holder of the Mehalchin Term Professorship of Management at the University of Colorado Denver Business School. He received his PhD in industrial and organizational psychology from the University at Albany, State University of New York, in 1993.
Eric Heggestad is an associate professor in the Psychology Department at the University of North Carolina-Charlotte. He received his PhD in personality research in 1997 from the University of Minnesota.
Mark Mazurkiewicz is a graduate student in I-O psychology at Colorado State University.