If you had the choice, would you rather be in a job interview answering questions of your prospective employer face-to-face, or all alone, inputting your answers onto a computer keyboard?


If you're like most people and consider job interviews as pleasurable as a visit to the dentist, then the latter would be your preference.

For market researchers, computer interviewing is becoming more and more their preference as well for assessing things like product demand. Personnel officers prefer it to screen job applicants.

Why its popularity? Faster gathering and analysis of data, say its users. Furthermore, some seem to think respondents answer more truthfully to a computer than face-to-face to another human being.

"The technique is state-of-the-art," says Ted Evans, manager of planning and analysis at Chevron Corp.'s Ortho Consumer Products unit in San Francisco.

"It seems to make the respondent more comfortable and provides more time and reflection in answering the questions. Compared to personal interviews or mail surveys, we really believe the results are more representative of people's attitudes and opinions."

Chevron used the tool, provided by a software supplier, about two years ago on its field marketing people. Chevron asked the 140 employees nationwide to assess the company's marketing strategy as well as questions about their job.

The computer interviewing technique was opted over the paper and pencil method to prevent the employees' supervisors from "hovering around and looking over their shoulders." Says Evans, "That can be very intimidating."

Reducing intimidation

Because many sales people work out of their homes and don't have PCs, respondents were asked to go off the work-site to complete the approximately 30 minute interview. If they didn't have access to a personal computer, they were instructed to go to a computer store and, if necessary, pay for the use of the machine. Needless to say, only two were required to hand over the $5 or $10 "rental" fee.

Disk and letter

The floppy disk each participant received stored both open-ended and scaler-type questions. In addition, each respondent received a letter which guaranteed their anonymity and provided the basic instructions for completing the interview. The interview was to be completed and the floppy disk returned in two weeks.

Machine-readable

"One of the big advantages of the data being on a floppy disk is that it's in machine-readable form," says Evans. "This speeds up the data analysis tremendously. With hand-written surveys, the data would have to be key-punched and cross-tabulated. Not only is the paper and pencil method time-consuming but it leaves room for transcription errors."

Company adjustments

The results from the research showed that Chevron needed to improve its point of sale materials, adjust its advertising and make some minor shifting of workloads. Getting this kind of employee participation is high on the company's priority list, says Evans.

"We like active employee input at all levels of the organization on how we might improve as a whole. We can't assume the study was fully responsible, but we believe there's been an improvement in employee performance and productivity. That's the name of the game."

Evans warns those who seek employee input to remember a very important premise.

"You must genuinely want this kind of information and be willing to use it to improve your company. You must also be willing to accept the fact that there are some good ideas out there, ideas besides your own."

Product potential

For yet another purpose, Smith Kline & French Laboratories in Philadelphia, manufacturer of pharmaceutical and hospital products, uses computers to evaluate the potential for certain products within a particular market. Depending on what is being tested, their respondents are either doctors or hospital pharmacists.

The technique they use on the computer is called "conjoint analysis." It involves asking a series of questions, ones which are based on the answers to previous ones, to find out what features the respondents are willing to give up to get an improvement on another product feature.

One study the company recently conducted involved packaging options for a prescription drug and price-sensitivity. The objective was to find out from hospital pharmacists what would happen to the sales volume of the product if its price was raised or lowered.

Another study evaluated the market potential for a new product. The doctors were asked a series of questions about the trade-offs they would make in choosing to prescribe a new drug. For example, a doctor might be asked if he would be willing to accept a product with slightly more side effects if that product was also slightly more effective.

The whole process of evaluating the potential for a new product usually takes about three months, says Don Marshall, manager, quantitative analysis for Smith Kline.

It begins with a product management request to investigate a new competitive product on the market. The product analyst then helps develop a hypothetical drug profile. Next, a questionnaire is created and then fielded at a medical convention and other locations where the respondents are chosen. After analyzing the results, a simulation model can be devised which will indicate what the market share of a new product with that profile should be.

The value of computer interviewing is that it can easily handle more drug attributes than the paper and pencil method.

"The paper and pencil survey is also very costly and time-consuming because all the answers need to be coded and key-punched and then a simulation model developed," explains Marshall. "With a computer, all the data is right on the disk. We don't have to wait for coding and key-punching."

Marshall gives other pluses for using the machine interviewer.

"Doctors like to play with a computer. They seem to think harder and pay more attention to it. We seem to be generating better data too."

Technique and equipment

Respondents can answer computerized questions in two ways. The first is to type responses onto a keyboard to questions displayed on a video screen. Another is to respond to a mechanical voice at the end of a telephone line by pressing numbers on a telephone. In most cases, you'll need a personnel computer and a color video?display terminal for attractiveness.

Some limitations

Not all interviewing problems are solved by using a computer. Unlike humans, computers are unable to explain misunderstood questions, recognize unclear answers or prod respondents to elaborate. And unless the respondents are good typists, the computer can't usually give lengthy responses.

Additionally, there is a problem with the technique when the respondents are asked to be interviewed at a central location such as at a convention or shopping mall. The problem is who is being interviewed, a restriction which could drastically limit the value of the results.

Some are solving this dilemma by telephoning people at home using automatic-dialing equipment and asking questions with a mechanical voice. Since these randomly dialed numbers are considered by some as an invasion of privacy, many states have forbid this method and restraints have been suggested in many others.

Computer proponents

Proponents of the method, however, point out its non-human qualities are really a plus. Computers don't press but wait patiently while a respondent takes time to think about a question before answering. It's inability to show facial expressions or emotion make the machine unbiased.

Beyond this, computers put questions to each respondent all in the same way yet at the same time show flexibility. A respondent's answers can bring up new questions based on earlier responses.

The potential for computer interviewing to "replace" human interviewers is a possibility, but probably not for sometime. For now job interviewees may still have to speak face-to-face with the company boss.