CHAMPAIGN, Ill. — New research co-written by two University of Illinois Urbana-Champaign experts who study personnel psychology points to a better way of helping researchers and employers identify the differences in how people respond to personality tests.
The tool — a hybrid of two existing measurement models called the “Mixture Dominance-Unfolding Model,” or “MixDUM” — could help human resources practitioners make better selection decisions by extracting more accurate personality information about job candidates, thus potentially paving the way for managers to create a more productive, diverse and equal workplace, said Bo Zhang, a professor of labor and employment relations and of psychology at Illinois and the lead author of the paper.
“The core contribution of this research is that we developed a tool to help researchers identify the differences in how people respond to personality questionnaires,” he said. “Previous research assumes everyone is using the same response process, but we’ve proposed that there are two ways in which people might respond to personality items, and different people may adopt a different way. We need to correctly identify the way each person adopts to respond to personality items. Thus the creation of our mixture model.”
The first response method was dubbed a “dominance process” and the second an “unfolding process,” Zhang said.
Using extraversion as an example, in the dominance approach, “it’s assumed that the more extraverted you are, the more likely you will agree with an extraversion item,” he said. “Basically, this dominance approach assumes that there is a progressively increasing relationship between your extraversion level and the probability of agreeing with an extraversion item.”
In the unfolding approach, “you will only say that you strongly agree when there is a close match between the contents of this item and your true extraversion level,” Zhang said. ”If you are more extraverted or more introverted than the degree of extraversion implied by the item, you will disagree with it. The unfolding approach assumes an inverse U-shaped relationship between your extraversion level and the probability of agreement.”
Those are vast differences in how people respond, Zhang noted.
“Some might follow the dominance approach, while others might follow the unfolding approach,” he said. “If we don’t have a tool to help us differentiate between these two types of responses, we just keep applying the same methods to score everyone’s response, which will lead to inaccurate estimates of people’s true personality.”
But the mixture model solves this problem by integrating the two response processes into one model, Zhang said.
“The tool that we developed can help us differentiate between these two types of responses in the population instead of assuming that everyone is using the same response process,” he said. “We can also do the scoring based on which group you belong to. So that’s the major contribution of this model. We were able to discover the population heterogeneity regarding how people respond to a personality questionnaire.”
The implications of the research point to making it easier for managers to screen employees more accurately to determine if they would be the right fit for a particular work environment, Zhang said.
“If we want to create a more diverse, a happy workplace, that means we need to place the right person in the right position,” he said. “And how do we place the right person in the right place? The first step is to assess their personality accurately. And our model is a more granular, nuanced way of assessing personalities.
“It’s not a way of measuring personality, it’s just a better way of extracting the relevant information from the data we collected from the questionnaire, which means we can obtain a more accurate estimate of their personality.”
Zhang’s co-authors are R. Philip Chalmers of York University; Lingyue Li of the University of Illinois Urbana-Champaign; Tianjun Sun of Rice University; and Louis Tay of Purdue University.
The paper was published by the journal Organizational Research Methods.