You`re on your best behavior, spiffed up and eager to please, applying for a sales job with a big chain store. You`re asked to take a test; it`s voluntary, you`re told.
You sit down and read it over, and it says that skipped, untrue or faked answers may result in your not getting the job. Wait a minute. Is that
”voluntary”?
You read on. The second question is, ”Do you always tell the truth?”
Now what do you do? If you lie and answer yes, will you be disqualified for faking? If you tell the truth and answer no, will you be disqualified for being a liar?
And how do you answer this one: ”How honest are you?”
Have you figured it out yet? You`re taking an honesty test, one of scores of paper-and-pencil tests employers across the nation use to try to determine if prospective employees are likely to steal or cheat on the job.
As the use of these tests has grown in recent years, so has an issue that has been troubling psychologists, congressional investigators and others:
How honest are honesty tests?
After grappling with that question in a two-year study, a task force of the American Psychological Association concluded in a March report that it`s impossible to say whether the tests now in use are generally valid and fair.
But the report also concluded that they are probably less harmful than alternatives they fear employers might otherwise use: privacy-invading security checks or vaguer, more intrusive and even less valid tests of
”dependability” and ”counterproductivity.”
A year ago, Congress` Office of Technology Assessment also concluded that there was no way to assess the tests` validity, because 95 percent of the research on their validity is done by the for-profit companies that sell them, and the independent research has been contradictory and incomplete.
In its study, the psychologists organization did try to evaluate some of the company-conducted research, as long as it met complex standards for psychological research. Based on that material and some independent research, the psychological association report concluded that people rated as honest by the tests are rarely found to be dishonest on the job later.
”Integrity and honesty tests do a pretty good job of predicting who`s going to be successful on the job,” says Wayne Camara, a psychologist who heads the association`s science directorate that conducted the study.
But do they unfairly weed out large numbers of people who would also be successful, that is, honest? And do they simply fail ”the stupid thieves,”
as Camara puts it, and pass smart thieves who are clever enough to give the
”right” answers?
”There`s no research on that. I mean there`s just nothing,” he said in a telephone interview from association headquarters in Washington.
Other critics have questioned the very premise of the tests: that honesty is a quality that remains constant in people regardless of circumstances.
Fraudulent promotion?
It`s hard to know how widely the tests are used. Developed in the 1920s, paper-and-pencil honesty tests remained in the shadow of electronic polygraph, or lie detector, tests for decades, Camara says. The tests, which typically sell for $7 to $30 a copy, began to find a mass market only in recent years, he says, as laws have restricted the use of polygraphs.
The most recent assessment of the tests` popularity that the American Psychological Association could find was done in 1985 by Paul R. Sackett, a University of Minnesota psychologist, who estimated that 5,000 companies were using them to test 5 million people a year. They are most often used by large retailers and other firms that hire large numbers of entry-level employees who have relatively easy opportunities to steal.
The psychological association`s March report found vast differences in the testing standards and marketing methods of companies that sell the tests. In fact, the task force wrote: ”Promotional claims for honesty tests, as perhaps for all other procedures used for pre-employment screening, vary from the circumspect to the fraudulent.
”We have seen a number of promotional brochures that are so clearly excessive and overblown as to make a test expert cringe in embarrassment. In the most flagrantly hucksterish of these, all problems associated with test use are unmentioned, and the purported reduction in actual theft that can be achieved is wildly exaggerated.”
The task force said honesty tests should be required to meet the association`s technical standards for psychological tests, something most companies couldn`t do now, Camara says.
Many companies that responded to requests for information, he says, had no idea what the task force was talking about when it requested technical information about how they scored their tests and explained their methods to their customers. Of the 45 testing companies the task force could identify, half did not respond to their queries, and most that did supplied seriously deficient documentation or none at all, Camara says.
Camara says the few reputable industry leaders, which sell 70 to 90 percent of the honesty tests given in the U.S. and employ credentialed psychologists, did supply ample documentation.
But even these companies, such as Reid Psychological Systems of Chicago and London House of Park Ridge, operate in a closed world, he says, constrained only by guidelines developed by a trade association that has no enforcement mechanisms or independent assessment of whether member companies are following them.
The psychologists association report said ”almost no evidence at all is available” regarding the validity of tests ”beyond assurances that evidence exists. For a few … extensive research has been completed. However, even in the latter cases, test publishers have relied on the cloak of proprietary interests to withhold information concerning the development and scoring of the tests.”
Massachusetts is the only state that prohibits the tests, though Rhode Island has enacted a partial ban and the attorney general of New York has recently proposed banning them.
Biggest mystery
Perhaps the biggest mystery about the tests is how they are scored. Because the tests are proprietary, test companies won`t tell anybody what the right and wrong answers are. The companies refused to give their scoring keys to either the Office of Technology Assessment or the American Psychological Association.
That means researchers can`t determine what concept of honesty the companies are using to design the tests.
David W. Arnold, a psychologist and lawyer who is general counsel and vice president for research of Reid Psychological Systems, says the tests are constructed empirically: Large numbers of people who are about to be hired answer a large battery of questions. The company follows up on the employees to see which ones are fired or become problems and how they answered the questions. It uses the information to refine the tests. The test-makers use no preconceived construct, he says.
Because employers risk violating discrimination laws if tests are found to be biased against minorities or women, test companies have done
considerable research to show that the tests have no adverse impact on them, and Camara says this research was clear and careful enough to convince the psychological association.
But the tests may unjustly screen out large numbers of honest people of all races and both sexes, he says. And this effect may be magnified if a test is scored with a single pass-fail cutoff point, instead of on a continuum of risk to be taken into account along with other information about the job applicant. The association found only one or two tests that use faking scales designed to detect and invalidate faked answers, Camara says.
Is honesty constant?
In its report a year ago, the Office of Technology Assessment concluded that because honesty and integrity are ”value-laden concepts,” honesty tests should be held to a higher standard of validity than other forms of psychological testing. But it also challenged the assumption that honesty is a predictable and constant quality, rather than one that varies with circumstances.
For instance, do employees who steal do so because they lack honesty, or do they steal in situations where they feel employers are exploiting them and remain honest in situations where they feel employers are fair?
No federal legislation has been proposed regarding honesty tests since the technology office`s report was issued. ”The (office`s) report did not seem to have a great impact, and I think that might be because it was so tentative,” says David Arnold at Reid. He calls it ”an inadequate report”
and says an industrial psychologist who reviewed it concluded that the technology office`s staff was influenced by biases.
But Nancy Carson, the technology office`s program manager for science, eduction and transportation who oversaw the study, rejects that criticism, saying, ”I think (the office`s) credentials are very strong.”
Despite the criticisms leveled by the psychological association report as well, Arnold says there was good news in it for the industry.
”They said these tests have validity, no invasion of privacy, don`t discriminate,” he says. ”There are some obviously below-standard publishers out there, (but) the good publishers account for the majority of the market.” Sampling your honesty
These questions, drawn from sample honesty tests prepared by Reid Psychological Systems and London House, are of the overt type:
Did you ever think about committing a burglary?
If you could get into a movie without paying, and be sure you would not get caught, would you do it?
A person regularly borrows money from his company without permission but always repays it at a later date. Do you think this person is honest?
The following questions, culled by the U.S. Office of Technology Assessment from honesty, dependability or ”counterproductivity” tests, are the veiled or covert type and are intended to reveal attitudes that correlate with honesty. Tests relying only on these kinds of questions are considered less accurate in predicting honest behavior than tests relying on overt questions:
How often do you blush?
How often to you make your bed (every day, never, etc.)?
True or false: Eating right is important to my health.
On the average, how often during the week do you go to parties?
