Tests to target dependability

Tests To Target Dependability

In 1985, when he became director of personnel for retail stores at the Morse Shoe Co. in Canton, Mass., Mark Rich faced two of the most serious problems in retailing: a high turnover rate and annual losses of millions of dollars through employee theft and paperwork errors.

"Most of our problems were with newly hired assistant managers," Rich says. "Too many were being fired for cause. I realized that if we had a better means of pre-employment screening, a more accurate way of predicting on-the-job success, we would cut our costs dramatically and improve our service."

Like most businesses, Morse Shoe had relied only upon traditional methods of employee selection: evaluations of applications, applicants' previous experience, interviews, and reference checks. "Although these techniques have a certain value, they must be used by highly trained, experienced executives to be good predictors of success," Rich says.

In taking various measures to solve the problems in the firm's retail Fayva division, Rich introduced a relatively new and promising psychological-testing method: a dependability test. It was similar to the first such test, which was developed for Macy's department stores. The dependability test, an adaptation of what industrial psychologists term "trait testing," was developed by Stanard & Associates Inc., a Chicago psychological-consulting firm specializing in test development. This variation on trait testing examines and evaluates an employee's or applicant's attitudes, practices, and values that are job-related, as opposed to a cognitive test, which deals with the ability to reason and learn.

"We call it a dependability test since that name seems to best describe what it does," says Steven Stanard, president of Stanard & Associates. "Although it is used primarily with applicants for nonmanagement positions, it is also used for entry-level management positions. The test predicts who will probably be successful by identifying those who are likely to be punctual, conscientious in their attendance, good producers, and honest. Those applicants who do not have these qualities are eliminated."

Each dependability test that is designed for a company is validated to show there is a clear relationship between a certain test result and job performance.

For the dependability tests developed for companies such as PhotMat, Yellow Freight, Macy's, Morse Shoe, and Carter Hawley Hale Stores, extensive validation studies involving thousands of employees were conducted by Stanard & Associates.

Rich says that dependability testing has proved valuable at Morse Shoe. "Our losses have been reduced by about one-third in the first year. Our other efforts have certainly helped, but the use of dependability testing in pre-employment screening has played a significant part."

In businesses where job failure and turnover can be costly, it is welcome news that there is a paper-and-pencil psychological test that can accurately predict whether an applicant will succeed in the job that he or she is seeking. According to estimates by the U.S. Department of Commerce, annual losses from employee theft might run as high as $40 billion, and the cost of turnover is staggering.

The use of psychological tests to screen job candidates is hardly new. Many major corporations use psychological tests routinely. According to Sanford Hotchkiss, an industrial psychologist in Youngstown, Ohio, "Pre-employment testing is the bread-and-butter business of many consulting firms. Millions of such tests are given each year."

Many other pre-employment tests are administered by employers both large and small, but problems may arise with them. Although many psychological tests used in pre-employment screening are designed to be administered and scored by nonprofessionals, errors can occur when such tests are interpreted by wholly untrained persons such as assistant managers, secretaries, or receptionists.

Examples of such tests that are often relied upon by nonprofessionals are "honesty" or "integrity" tests, which predate the dependability tests. These paper-and-pencil tests are easy to obtain and administer, and they are said to measure the probability that a person will be honest. But such tests have a poor track record, according to Dennis Sweeney, an industrial psychologist and vice president of Partners In Change, a Pittsburgh consulting firm. "There are a number of serious problems with such tests," he says. "Research on them is uniformly unencouraging, and there has been almost no independent confirmation of test validity.

"Many of the questions are rather transparent in their intent, and the tests tend to screen out naive applicants while allowing sociopathic personalities to pass undetected. They are the ones who can really hurt a business."

Furthermore, almost all of the "honesty and integrity" tests do little more than predict how people would do on a polygraph examination, not how they would do at a job. This is because the polygraph was used as the standard of accuracy by the designers of the tests. But the polygraph, long a favorite tool of business in a variety of circumstances, will soon play almost no role in private employment situations; as a result of a recently enacted federal law, polygraph use has been effectively prohibited except in special cases.

In certain ways, dependability testing seems very much like many of its predecessors in the field. "There are a number of significant differences between dependability testing and the `honesty and integrity' tests," Stanard says. "First, dependability testing tests for a broad range of negative behavioral characteristics--all the reasons why someone would be unsuitable for a job; `honesty' testing focuses on honesty/integrity and the likelihood of stealing.

"Secondly, the intent of the questions on `honesty/integrity' tests is transparent, but a job candidate cannot tell the purpose of the questions on a dependability test.

"A third difference is that dependability testing is not based upon polygraph results but upon real-life results. It is constructed so as to test for those qualities that have been shown to be job-related for any job being studied rather than attempt to meet any other theoretical standard."

The basic theory behind such a dependability test is that if a person has the values, the attitudes, and the habits of a successful salesperson, he or she will probably behave like a successful salesperson. Conversely, if someone is found to have the values, attitudes, and habits of employees who have been tardy, absent, or caught stealing, it is just as probable that he or she will eventually behave in the same way.

"No single area becomes too important in the evaluation of a job candidate with dependability testing," Stanard says. "Instead, it is a matter of finding a pattern, a mosaic of individual points that give direction and meaning to the results. The specific areas which the test predicts are job performance, tendency toward dishonesty, absenteeism, tardiness, and interpersonal relationships."

Each company must have a separately developed and validated dependability test because of the differences among jobs and settings in different companies. Even within a company, patterns of successful work behavior vary from one job to the next. The traits and skills needed to be a good store manager, for example, differ from those required to be a good salesclerk.

The typical cost of developing the test is $20,000 to $30,000, but each company owns the test developed for it, so the average cost of administering the test declines as it is used through the years, according to Stanard & Associates. For a firm with about 3,000 employees, for instance, the average cost of the test can be about 30 cents.

The questions used in each test are similar and deceptively simple. Some are designed to reveal behavioral patterns ("In school I sometimes had to go to the principal's office for acting up"). Some call for objective information ("What was your standing in your high-school class?"). Some are designed to reveal the person's self-image and level of confidence ("Most people are not as strict about right and wrong as I am"). And some reveal social attitudes ("It bothers me when a smart lawyer uses the law to get a criminal off free").

Although such questions may seem simplistic, few test takers seem able to bluff their way past the test's overall findings. "A candidate may lie on any one question, but a pattern emerges which allows generally accurate predictions when the test is complete," Stanard says.

Usually, each company scores its own tests. Some have a pass-fail cutoff, and others employ a system allowing for a range of scores. Subjective judgements and other facts are usually incorporated into the hiring process as well.

Ken Bernard, owner of Bernard & Associates, a consulting firm in Holmes Beach, Fla., is enthusiastic about the use of the dependability test in pre-employment screening. "There's no doubt that the test is able to make highly accurate predictions," he says. "That means reducing the number of new hires who do not work out and cutting employee theft significantly. And that allows for greater productivity and efficiency. Eventually, a business can save a lot of money."

Bernard employed the test experimentally early last year with a Florida restaurant that was experiencing severe losses from internal theft and employee turnover.

Bernard says that starting in January he administered the test to a total sample of 160 people--"140 were applicants [whose hiring was not influenced by the outcome of the test], and 20 were current employees. At the end of six months, we compared the predictions with the terminations that had been made. The correlation was very high."

Paul Gorsky, director of human-resource services for Carter Hawley Hale Stores Inc. in Los Angeles, agrees that dependability testing can be of significant value in selecting employees. He says: "We began using dependability testing in 1987 to reduce the number of employees who would have to be dismissed for cause--theft or a serious violation of our rules. This was the first time we had used psychological testing; previously, we had always relied upon the standard methods of employee selection."

As with all individually developed tests, it was necessary to validate the test by administering it to a sample of job applicants and comparing predictions with actual results six months later.

"We have been using the test in actual hiring decisions for only a relatively short time," Gorsky says, "but our preliminary figures show a 10 percent decrease in terminations for cause. I can't yet attach a dollar figure to that since other measures are involved, but our savings stand to be substantial."

The experience so far with dependability tests has shown they work best with businesses that have many employees and experience a high turnover rate of nonmanagement personnel, such as large retail establishments with numerous outlets or banks with many branches.

(If you want more information to help you decide if dependability testing might help your company's hiring procedures, you can call Stanard & Associates at 1-800-367-6919 or, in Illinois, (312) 337-1729, or you can write to the company at Suite 600, 180 North Michigan Ave., Chicago, Ill. 60601.)

No test is perfect, and dependability testing can make predictions only in terms of probabilities, but it seems clear that this is a major advance in pre-employment screening and will enable many companies to upgrade the quality of new hires, thereby reducing turnover, cutting losses, and improving customer service.

COPYRIGHT 1989 U.S. Chamber of Commerce
COPYRIGHT 2004 Gale Group