How IBM changed the research world

By Heidi Dietzsch

We continue our overview of the basic methods of data collection and how they are evolving. Last month we discussed paper-and-pencil (PAPI) interviewing.

After decades of struggling with PAPI – time-consuming, error-prone and costly data collection method – the advent of the computer was a real game changer for market research.

When IBM launched the personal computer in 1981, the potential of quantitative research increased exponentially. Computer-assisted telephone interviewing (CATI) and computer-assisted personal interviewing (CAPI) became the primary quantitative data collection methods. CAPI replaced the clipboard and paper questionnaire in face-to-face fieldwork, these days often done with tablet computers, while CATI improved and simplified traditional telephone interviews.

CAPI and CATI are similar, but with CAPI the interview takes place in person instead of over the telephone. With CATI, trained interviewers in a specialised call centre conduct the interviews. All interviews are recorded.

One of the advantages of CAPI and CATI is that they use sophisticated software that allows automatic questionnaire routing where irrelevant questions or sections can be skipped. This simplifies the interviewer’s work – they no longer have to struggle through complex routing in paper questionnaires. Rather, they can focus on guiding respondents through the questions.

The software can also integrate a variety of formats – multiple choice, rating scale and verbatim questions. Rating scale questions usually use the Likert scale, named after the American psychologist, Rensis Likert, who invented it in 1932. Likert scales are designed to measure attitudes and opinions. Respondents are provided with a statement and indicate to what extent they agree or disagree with it. Usually five or seven ordered response levels are used. For example: “Please agree or disagree with the following – My private banker always ensure that all my needs are met.” 1) Strongly agree 2) Agree 3) Neutral 4) Disagree 5) Strongly disagree.

One area requires extra vigilance: to set up questionnaires on CAPI and CATI systems, without any errors, is time consuming. Questionnaire routing can be complex and computer programmers need to ensure that this works flawlessly. Before fieldwork can start, researchers have to do thorough questionnaire testing. One small error in a questionnaire can lead to massive problems with the data.

Data collection using CAPI and CATI is more efficient and accurate because the interviewer enters answers directly into the computer rather than on a paper questionnaire. This eliminates the separate data capturing step where mistakes can creep in. Furthermore, with these two methods interviewer errors resulting from incorrectly following questionnaire routing and skip patterns are eradicated.

Despite their similarities, CAPI poses some restrictions that CATI does not, and vice versa. For instance, CAPI interviews sometimes have to take place in some strange places – crowded malls, hospitals, remote villages or busy inner-city street corners. This means interviewers can face challenging and even dangerous situations. Of course, failing battery power and intermittent internet coverage create their own problems.

While CATI interviewers are permanently stationed in call centres and thus do not struggle with the above, they also experience challenges. For instance, when there are many verbatim questions, interviewers need excellent typing and spelling skills. Poor data entry may lead to time-consuming data cleaning. Also, interviewers need to capture answers within the context it was intended, otherwise it could lead to compromised data integrity.

CAPI and CATI definitely made the lives of market researchers easier compared to older methods. Although these two methods have some disadvantages, they are widely used and most market research companies have invested heavily in them. At Intellidex we use CAPI and CATI interchangeably, together with digital collection methods.