Once again, we have necessity to thank for spurring invention. After years of frustration dealing with open-ended questions, researcher Jim Falk vowed to figure out a way to simplify the task of coding and working with written responses. The result is something he calls the Text Analysis Program, or TAP.

TAP uses proprietary artificial intelligence algorithms to group written responses into text clusters for easier analysis. In effect, the program does the coding and sorting of written responses for you.

As Falk, executive director of Com stat Research Corp., Cortlandt Manor, N.Y., states it in the program' s literature: "TAP 'understands' the meaning of words and groups written statements contextually. It labels each text cluster with the most representative statement and it provides summary statistics to show how many individual statements formed the cluster along with an index of internal consistency."

A typist enters the answers into the computer, inserting delimiters to separate the responses to each question and adding demographic information to each response so users can analyze how groups of respondents answered the various questions.

"If we have a four item questionnaire, for example," Falk says, "you would have the respondent's statements for each item and following that would be the demographic codes associated with that individual. And when I go to extract the data for analysis, if you give it specific demographic criteria, it will only extract that data which pertains to that demographic combination."

The program works best with short answers to specific questions. If the person rambles, Falk instructs the typist to shorten the answers.

TAP is currently available as a service from Comstat but Falk is preparing to offer it under license as a pro¬gram for the UNIX operating system. Plans for a DOS version are also in the works once Falk is able to get around the multitasking limitations of DOS.

Put to the test

The program was put to the test last year as part of a research project conducted by the human resources department of Philip Morris Companies. The research focused on determining how workers in the various Philip Morris operating companies view what's referred to as the "broken social contract" between employer and employee.

Steve Temlock is an organizational psychologist (his firm Organization Consultants Inc., is based in Westport, Conn.) who worked on the Philip Morris project. He explains the idea of the social contract: "Twenty five years ago, employees were expected to be loyal and hard-working and in return they got job security. Well, today that contract is broken in many companies because management can't promise job security. But management still needs loyal and dedicated workers and so the question is, how do you renegotiate the social contract in an honest manner with your employees?"

Big companies can have difficulty in answering this question, Temlock says, because management often assumes that all employees want the same thing. "[A big company] tends to make the assumptions that its executives make relative to the social contract. So if you have 50 year-old executives, their mental frame of reference in terms of how employees should be treated is very much driven by the kind of social contract they're accustomed to. Our research has shown that younger people are less concerned about the availability of lifetime job security than people who were around when that was still a possibility."

Anonymity guaranteed

A sample of 1,100 employees of the many companies under the Philip Morris umbrella - including Miller Brewing, Kraft, General Foods, and Philip Morris Tobacco - was recruited worldwide. All respondents were guaranteed anonymity. After volunteering to participate, they answered a brief survey that included both quantitative items and open-ended questions about their perceptions as Philip Morris employees. They then participated in a focus group to discuss issues related to the social contract.

Randy Kauto, vice president, employee relations, Philip Morris Companies, Inc., says the company didn't want to do a typical attitude survey. "We had a lot of anecdotal beliefs about how employees were feeling but no data. We wanted to do something quickly and at low cost using our own human resources people. We developed a brief worksheet that had both open-ended questions and more quantitative questions where the respondent would score us on 16 different subjects on a 1-6 scale."

"The employees cut across all kinds of jobs, age groups, and demographics," Temlock says. "We used TAP to see what the major themes were that they were thinking about. We also used it to study those ideas by demographic differences. It was a very useful program because it took qualitative data and it made it quantitative but with an enormous amount of flexibility. All of us have been in the situation where we've asked coders to look through qualitative data and code it. Once they've done it you don' t want to let people look through the data a second or third time. But this program lets you start from scratch each time. I think it's a remarkably powerful tool that will enable the richness of qualitative data to be used in quantitative ways."

The whole project, undertaken last spring, took two months from start to finish. This was the first time that Philip Morris had used focus groups for research with employees, Temlock says. "As an employee, if your company gives you a questionnaire and you usually just sit in a room and take it. You don't end up discussing it until after the results are in and sometimes not even then. But Philip Morris was willing to let people really have a conversation with each other on the subject of the social contract or what it called the Philip Morris partnership."

The employees were asked, If there was one element of the work partnership that you could change, what would that be? The predominant response across all operating companies related to employee-manager communications. "We weren't expecting that," Kauto says. "We expected to hear, 'Make our jobs more secure.' "

Philip Morris learned that employees tend to view their direct manager as "The Company" and that, in the absence of lifetime job security - which they realize is no longer a reality - they want to be able to communicate with their direct supervisors.

Blew 'em away

The real value of the TAP program came out when human resources personnel from the various Philip Morris operating companies got together to look at the research results, Kauto says. "When we delivered the reports in front of representatives of all the companies that had been working on this project, Jim was there with his laptop and he was able to do special iterations on the data and print them out on the spot, which blew everybody away.

"When we've done full scale attitude surveys in this company it's usually taken several months from beginning to end. And to then create another variation on a report adds more time. So we were all very impressed with the speed and capability of the program. For example, during the meeting one person wanted to know what women with more than 10 years service to the company said in response to a certain question. In 10 minutes she had the report in her hand. Essentially, everyone left the meeting armed with the information that they felt was most useful for them.

"Everybody has identified the broken employee contract as an issue but no one has figured out how to rewrite it, so we were without any kind of assistance in terms of other companies' experience. But our research gave us the kind of information we needed to take action."