Mall Intercept Survey

The mall intercept survey is used most frequently for market research. The name is derived from its traditional deployment at shopping malls, where there are many readily available persons to sample and from whom to gather data. Mall intercept studies rarely use probability sampling methods, and therefore the results of such a survey cannot be used to determine scientifically the attitudes and opinions of the target population. To address this limitation, researchers often attempt to conduct the survey in a number of locations and at varying times to try to ensure that the variability within the population of interest is represented.

Generally, an intercept study entails selecting respondents by stopping them (i.e. intercepting them) in a public place (e.g. a shopping mall). Once a potential respondent is stopped by an interviewer, he or she can be screened for the criteria that determine the eligibility for the particular study. In a mall intercept study, qualified respondents are often taken to an interview facility within the mall to complete the main body of the questionnaire. Mall intercept studies are particularly useful in cases where respondents need to view or handle materials, and the method often can provide an easy and cost effective way to locate “live” respondents. Additionally, mall intercept studies can use longer and more complex questionnaires than could reasonably be implemented using a telephone or mail survey.

The key disadvantage of a mall intercept study is that it generally relies on nonprobability sampling for respondent selection, and therefore the results cannot be used to make statistical determinations about any identifiable target population. Because respondents must be present at the mall (or public place) and generally are approached in a nonrandom fashion (i.e. one that is at the convenience of the interviewer), the researcher cannot definitively determine how well the sampled respondents represent any population, not even the population of those at the mall on the day(s) of data collection. Research on mall intercept surveying has shown them to underrepresent lower-income and older persons. This does not mean that the results cannot be used to say anything about the population of mall attendees, only that the results cannot be used with confidence to make any statistically reliable estimates about them or any other population. For example, no statistical calculation of sampling error is justified when using a mall intercept sample, because there is no known nonzero probability of selection that can be assigned to the members of any target population.

A number of procedures can be put into place when using a mall intercept study that may help to increase [Page 449]the representativeness and reliability of the survey. First, it should be determined whether a representative proportion of the population of interest would actually be found at the mall or other selected intercept locations. For this reason, some intercept studies are not conducted at malls but are instead conducted outside of grocery stores, movie theaters, and/or other places where the target population is most likely to be found. Second, the survey researcher should carefully consider the geography and number of intercept locations at which data will be gathered. For example, to generate credible findings for a study of the adult population in the United States, it would not be wise to select only one mall in California and one in New York. Once the broad geographic areas are selected, the researcher ideally would randomly select the malls to be used for the study against bias being introduced in selecting only malls or locations of a particular quality or character. Third, interviewing should be conducted over a variety of days (weekdays and weekends) and should also be conducted at a variety of times of day and evening to ensure greater diversity in respondents. Fourth, a systematic sample of respondents at the mall(s) or other location(s) should be selected rather than allowing interviewers to approach respondents at the interviewers' own convenience. Finally, the researcher can consider the value of conducting a small-scale probability sample (using telephone or mail) to determine how the demographics of the mall intercept respondents and those selected at random differ, if at all, and whether or not these differences have any implications for the survey's findings.

...