The presence of the interviewer undoubtedly has some effect on the accuracy of the data obtained; the effect may be both good and bad. It is more tense when interviewers are present in person than when they are on the telephone; but when an interviewer is present, it is hard for respondents to report incorrectly if the environment does not support their statements (e.g. do they live in a house or an apartment). Interviewers can probe and explain when they think it is useful and they can do this better in person than on the telephone.
On the other hand, much effort is often put to limiting the interviewer in adapting to individual interviews. The interaction between interviewer and respondent’s replies, and the personal interests and attitudes of interviewers will cause them to interpret responses differently.
Personal interviews at home can have a positive effect on the accuracy of information obtained by asking respondents to check the specific brands they may have on hand rather than relying on memory. Mail permits more flexibility on the part of the respondent in this regard, but it does not provide the same degree of motivation. Telephone interviews presumably have the same advantage as home interviews in this regard, but most researchers do not want to send respondents away from the telephone because some won’t come back.
A final consideration on the quality of information obtained is that of cheating on the part of interviewers. Personal interviewers, whether working door-to-door or in shopping centers, work alone and under little supervision in most cases. Some are tempted to fill in questionnaires in full or in part without completing the interviews. Procedure for check backs with respondents and for testing patterns of response enable most well planned studies to minimize cheating, but telephone interviewing done at centralized locations and under direct supervision provides the best control.
Control of Sample: The largest difference among the three methods of communication is in the amount of control they permit over the sample from which information is collected. A method is weak if it permits the interviewer to determine the individuals from whom data are obtained. There are two elements involved; the original designation of individuals from whom it is proposed to obtain information and the final determination of the individuals from whom information is actually obtained. The first element will be considered here while the second element will be considered in the following section on response rates.
Theoretically it is easy to designate a precise probability sample of consumers for a study using personal interviews at home since all members of the population can be reached with the method. In practice, it is usually necessary to have a list of all members of the universe, if a representative sample is to be obtained. For most large universes such as all consumers in the United States or even all households in St. Louis such lists are not available. It is possible to substitute lists of areas (e.g. city blocks) and so to obtain adequate samples; but to do so for a national project is difficult, time consuming and expensive. Few marketing research studies now use personal interviews with such elaborate area samples. Where lists of the universe to be studied are available e.g. the change account customers of a department store, strict sample control is feasible.
When personal interviews are made at shopping center locations, no good control of the sample is possible. Usually respondents are selected at different locations within the shopping center, at different hours of the day and no different days of the week. Nevertheless, samples selected at shopping malls tempt them to be composed of respondents who are much younger than the average and are more likely to have children younger than 18, have higher than average incomes, and spend more time shopping.