Conventional wisdom holds that the standardized tests some employers require of job applicants serve as a barrier to equal employment. But a pioneering study shows just the opposite: Screening increases employers' precision in matching applicants to jobs and can raise productivity for workers of all races--without hindering minority hiring.
"Job testing has the potential to raise productivity by improving the quality of matches between workers and firms. But because of the near-universal finding that minorities fare relatively poorly on standardized tests, there is a pervasive concern that better candidate selection comes at a cost of reduced opportunity for groups with lower average test scores," says David Autor, associate professor of economics at MIT who conducted the study with David Scarborough of Black Hills State University.
Their study, "Does Job Testing Harm Minority Workers? Evidence from Retail Establishments," was recently published in the Quarterly Journal of Economics. The paper is available at http://econ-www.mit.edu/faculty/dautor/papers.html.
As part of their research, Autor and Scarborough studied hiring and job longevity among primarily high school-educated workers who were paid hourly wages for customer-service jobs in the private sector. The researchers relied on data from a national retail firm whose 1,363 stores switched from informal, paper-based screening to computer-supported, test-based screening over the course of one year.
"Access to this data gave us the unique opportunity to evaluate the effects of job testing on minorities in a competitive business environment," Autor says.
Both paper- and test-based hiring methods used interviews, but the latter relies significantly on a personality test administered and scored by computer.
The retailer's 100-item personality test ranked attributes such as agreeableness, conscientiousness and extroversion that are associated with success--or productivity--in customer service.
"These tests basically predict how many times you're willing to say, 'May I help you with that?' and 'Have a nice day!' before you run out of patience," Autor says.
An outside firm, Unicru, scored and analyzed the tests, highlighting problem areas and completing background checks and returned them to individual store managers. Qualified applicants were then interviewed.
Consistent with previous research, minority applicants performed significantly worse on the electronic employment test. But the researchers detected no change in the racial composition of hires once electronic screening was installed. Moreover, the authors found, productivity gains were equally large among minority and majority hires.
The findings are significant, according to Autor, because the outcomes do not support the accepted belief that minorities' relatively low scores on standardized tests mean that such tests harm the job prospects of minority workers.
"Initially, I was surprised. I expected the increase in productivity that followed job testing would surely come at the expense of minority hiring," Autor says.
But the test of insightful research may be that very surprise, the moment when accepted beliefs dissolve in the face of new facts. The paradox that job testing did not harm minority workers is resolved quite simply, Autor notes.
Before the computer test, the retailer informally screened for the personality traits that are measured by the test. Job testing made this screening process more systematic and precise, but research showed it did not tip the scales for or against any particular group of applicants. Consequently, the productivity gains from testing came from improved selection within applicant groups (e.g., minorities, nonminorities), not from hiring fewer minorities.
Autor is quick to note that discrimination in employment exists and that bias--arising from prior information, interviews or beliefs about a particular group--can affect equality in hiring and efficiency in the workplace even with electronic testing.
For example, if job tests exacerbate an existing bias, testing just increases hiring of groups favored by the test. Then productivity stalls and neither employer nor workers profit.
A version of this article appeared in MIT Tech Talk on May 7, 2008 (download PDF).
