Technology

Employee Testing: Send in the Clones

What should employers be looking for in prospective IT workers? The answer's all around.
Karen BannanSeptember 15, 2001

If there’s one thing they’ve got plenty of in California, it’s shortages. Indeed, over the past few years the Golden State has been struck by an endless series of shortfalls, including water, gasoline, grapes, and, most recently, electricity.

But last December, Ron Komers was facing an entirely different kind of shortage. Komers, who oversees the hiring of technology workers for Riverside County in Southern California, needed to fill a growing number of IT job openings at the municipality. In the past the county’s full-time IT recruiter, Melanie Hanisco, had generally been able to fill tech vacancies as they arose. But a high turnover rate — coupled with a mushrooming need for computer specialists — suddenly left Komers with 60 vacancies scattered across several departments.

At first Komers went through the standard channels, posting the listings in newspapers and on Internet job sites. Although he received a number of responses, he didn’t fill many of the positions. After several weeks of frustration and few hires, Komers decided to take an entirely different tack. In January the county HR staff began administering a preemployment IT test to current workers — most of whom had little or no tech experience. The online exam, from ePredix, measures cognitive ability, vocational and mathematical skills, and verbal reasoning. Komers’s reasoning was that the test might uncover a hidden well of IT talent right there among the clerks and typists and office administrators.

Drive Business Strategy and Growth

Drive Business Strategy and Growth

Learn how NetSuite Financial Management allows you to quickly and easily model what-if scenarios and generate reports.

He was spot on. “We had social workers and stock clerks who scored well on the [aptitude] tests,” reports Komers. “That proved to us that there’s a lot of untapped talent out there.” Guided by the test scores, Komers has selected some current, nontech staffers who are now being trained to fill 20 of those 60 vacancies.

While administering preemployment examinations to current employees is fairly unusual, applicant testing is not; companies have long been using various psychometrics to rate job candidates. In fact, in a survey published in Human Resource Executive, a trade magazine, 69 percent of the respondents said they used some form of preemployment testing to screen potential hires.

Mostly, however, the tests have been given to candidates applying for jobs that involve customer contact (service representative) or excessive stress (air traffic controller). Since IT positions don’t usually fall into either category, and since technical work usually requires considerable expertise, employers have tended to look more at résumés than Rorschachs. “You’ve got managers who think, ‘This person has three years’ experience with Unix and this person has five, so I’ll hire the person with five years’ experience,’ ” notes Joy Hazucha, senior vice president with Personnel Decisions International, a human resources and consulting firm. “The problem is, they don’t consider that these people also have to work with a team, and in some cases, as a project manager or a software developer.”

With the salaries of IT workers skyrocketing, however, some employers are now attempting to get inside the heads of job applicants. Sprint, the Kansas City—based communications company, administers a battery of psychological tests to prospective tech employees. According to Bill Donkersgoed, manager of selection systems at Sprint’s national staffing and technology group, the exams include cognitive and motivational testing combined with structured interviews. By conducting these assessments, Donkersgoed says, the company’s managers get a more complete picture of a candidate.

As the IT function grows in importance, the wide-angle lens becomes even more crucial. “The most common mistake employers make is to hire someone just on technical skills,” notes Hazucha. “These days, IT workers need more skills than just the technical ones.”

The Testers and the Testees

In theory, psychological testing enables companies to match the right applicant to the right job. It can also help employers steer clear of slackers and malcontents — or worse. “It’s the personality that does the job,” insists Jack Cammarata, president and CEO of Handwriting Analysis Inc. “It’s not experience or education. Those are funneled through the brain.”

But actually putting a number on personality traits — concrete concepts like conscientiousness and integrity — remains at best an inexact science. Critics claim that some applicants simply lie when they are confronted with thorny questions about their ethics and work habits. Moreover, psychological profiling, which is designed to gauge reasoning abilities and emotional stability, can be pricey. While some industrial-organizational (IO) psychologists charge as little as $50 per candidate for administering a standardized test, others charge upward of $5,000 for a more in-depth analysis.

Still, some experts say that’s chump change compared with the price of hiring the wrong tech worker for a position. “The cost of replacing one person is one-and-a-half to two times that of a person’s base pay,” notes Carl Van Horn, director of the John J. Heldrich Center for Workforce Development at Rutgers University. “When you look at it that way, spending $150 on a test may be a worthwhile thing.”

In addition, IO psychologists claim that an experienced test administrator can generally smoke out liars and fakes. Says Leaetta Hough, an IO psychologist at testing specialist Dunnette Group: “If you look for a spike or an overall high score, you’re usually going to find them out.” The best way to avoid the problem altogether, adds Hough, may be for the examiner to warn candidates ahead of time not to distort their answers.

Test-taking itself doesn’t take long, usually a few hours. Devising the exam…well, that’s an entirely different proposition. Some consultants take as long as a month to prepare the questions. Test administrators often start the process by interviewing managers in the department with the opening. The thinking, of course, is that someone who oversees a job has a pretty good idea of what the job entails and the attributes needed to excel in that position.

After factoring in those discussions, IO psychologists put together a working version of the test. Before giving it to applicants, though, examiners often administer the test to the best employees working in that same department. This step is crucial, say experts; the results serve as a kind of template when assessing applicants’ test scores and their likely job performance. Notes Frank Merritt, president of ComputerPsychologist.com. “Basically, you want to look into your organization and identify the 12 to 15 people you’d like to clone.”

Acsys has done just that. A couple of years ago this Atlanta-based staffing company began searching for a better way to assess its own potential hires. Ultimately, according to Acsys human resources consultant Lena Markovski, the company settled on a browser-based model designed by eTest. Although the eTest assessment comes with a couple of generic profiles built-in, Acsys went one step further. “We took a sample of successful people within our company and tested them to see what common skill sets they use daily,” explains Markovski. Then Acsys compares those results with the results of job candidates.

So far, Markovski says that this compare-and-contrast method has proved to be a valid predictor of on-the-job performance. “We’ve had people who have interviewed very well, but their test results come back off the chart where they should be,” she says. “Occasionally a manager will hire that person anyway, and three to six months down the line, the employee’s gone.”

Head Casing

The money spent screening that employee is gone, too. While calculations of the return on investment for preemployment testing vary from company to company, users agree that the payback period tends to be short. At Dallas-based Texas Instruments, for instance, the company tests both prospective IT hires and current IT workers. According to Steve Lyle, the company’s director of staffing, Texas Instruments generally recoups its investment in employee testing within three to six months.

Such a quick payback is critical for IT testing. Unlike other employees, the average tech worker stays at a company for only about 18 months. The short stay puts a premium not only on personality testing but also on skills assessment. “If you know a hire is probably only on board a year and a half, you can’t afford to have a six-month ramp-up,” says Mike Russiello, CEO of Brainbench, an online certification specialist. “People are hired for what they know today.”

Over the past few years a handful of companies have launched virtual services aimed at measuring what people know today. Bookman Testing Services, for one, offers a skills assessment tool called TeckChek. The online service features more than 150 Web-enabled proficiency exams, including tests for Java, data modeling, and telecommunications support and service. Other Web-based skill analyzers include SkillDrill’s BrainBuzz and Kenexa’s Prove It.

On the whole, these Internet-based assessment tools tend to be less expensive than traditional preemployment testing methods. Depending on the vendor, the cyberexams may cost anywhere from $10 to $150 per applicant. Placement specialists say that the low costs make these virtual tests particularly useful for weeding out underqualified applicants early on in the hiring process, so managers can focus on more-suitable candidates.

Still, some HR managers have doubts about placing too much stock in preemployment tests — no matter how early in the process they’re used. Sandy Jess, director of human resources for IT staffing company Matrix Resources, is one of those doubters. “Some employers might let the test make the complete decision for them,” says Jess. “That’s a big mistake.” By her lights, testing doesn’t eliminate the need to obtain job references or to conduct face-to-face interviews. “Human interaction is still key,” she says.

Even backers of psychological tests concede that the exams make for imperfect barometers. But they also point out that, when it comes to hiring, perfection isn’t necessarily necessary. “There’s no exact right person for a job,” insists Sprint’s Donkersgoed. “There are just better and worse degrees of fit.”

Karen J. Bannan is a contributing editor at eCFO.

Reading the Fine Print

While dozens of companies are turning to the Internet to help assess job candidates, others are taking a somewhat lower-tech approach. Handwriting analysis, long standard operating procedure for companies in France, is starting to catch on in the US.

Proponents of the technique say a trained graphologist — yes, graphologist — can gauge numerous character traits, such as willingness to learn, honesty, and flexibility. What’s more, backers claim that handwriting analysis is completely unbiased. They say a tester can’t completely determine a person’s age, sex, race, or religion merely by looking at a handwriting sample. That, in turn, should make fair assessments more likely.

But before hiring a handwriting analyst, employers need to read the small print. Like any preemployment test, handwriting analysis must meet all Equal Employment Opportunity Commission guidelines. If the analysis goes outside of the guidelines — say, if the tester asks personal or intrusive questions — the test is null and void as far as the EEOC is concerned.

What’s more, handwriting analysis tends to be extremely subjective.

Carl Van Horn, director of the John J. Heldrich Center for Workforce Development at Rutgers University, says the results often vary according to who’s giving the test. “I’m not a big fan of handwriting analysis,” he says flat out. “I’m not persuaded that there’s validity [in test results] from one examiner to the next.”

Indeed, several organizations for handwriting analysis currently exist in the US, and different practitioners often employ different methods. Hence, an employer could get conflicting profiles for the same job candidate — hardly a ringing endorsement for the technique.

Handwriting analysis isn’t cheap, either — a detailed report can cost anywhere from $150 to $500. Graphologists say HR managers would be well-advised to get plenty of references before choosing a handwriting expert. Also, they say corporate users should avoid hiring any analyst who doesn’t intend to study the requirements of the vacant position. “Doing open-end analysis is like trying to match an apple with an orange,” warns Jack Cammarata, the president and chief executive officer of Handwriting Analysis Inc., a testing company. “Handwriting analysis — when done right — requires a lot of preliminary work.” —KJB

4 Powerful Communication Strategies for Your Next Board Meeting