 |
|||
 |
|||
|
FULL ARTICLE What is the reason given most often for avoiding on-line research projects? Sampling. Clients and marketing research suppliers are convinced that on-line research is not "representative," based on the fact that the majority of consumers do not access the Internet. Also expressed is the concern that "self-selection" alone could bias the results of on-line studies. How do we address these issues? When someone says that a particular sample is not representative, ask "Of what?" In traditional research, random-digit dialing telephone samples are considered representative of the universe of telephone subscribers. Studies conducted in malls are viewed as representative of the demographics of those locations. We take an "nth" select to obtain a representative sample from a list. Panels are built to reflect the demographics of a population. In each case, researchers recognize that the samples obtained are representative, to varying degrees, of the universe from which they are drawn. Similarly, on-line research samples are representative of the universes from which they are drawn, depending on the methodology used to create them. We interview Web site visitors by intercepting them at random and inviting them to participate. When we have access to e-mail lists, whose members have agreed to receive solicitations, we use "nth" sampling procedures. We also use convenience sampling by developing panels of people interested in participating in on-line research projects. They key is to be clear about what universe the sample represents. The issue of self-selection is a factor in sampling for all forms of market research. Research participation always is voluntary, and we never know what differentiates those who agree to be interviewed from those who do not. As with many types of surveys, when we conduct on-line research we attempt to increase participation with incentives. We offer sweepstakes, prizes or cash to those who qualify and participate in Web surveys or on-line focus groups. The problems of self-selection are no greater or less on the Web than with traditional methods. Furthermore, evidence shows that on-line and traditional research methods often yield similar results, both in quantitative research and qualitative studies. When our clients have asked us to run parallel studies, the results have been virtually identical. Others in the field have reported similar findings. For example, a telecommunications client conducted parallel quantitative studies. The core marketing issue was to evaluate new strategic corporate concepts. One supplier conducted the research as a mall intercept. We conducted the study as an on-line survey, in which our Random Web sampler intercepted visitors to the client’s Web site. On all key measures, the results of both studies lead to virtually the same conclusion. However, the on-line study was significantly less expensive. In another case, an advertising agency conducted a parallel qualitative study. Agency personnel moderated the traditional groups, while we conducted focus groups on-line. The agency reported that although the on-line participants were more aware of advertising, the results from both were similar. The Web, like any central location, simply serves as a meeting place. It can be used to study people who would otherwise be too expensive to survey. The Web is not only a viable place to conduct research, it’s a highly desirable place. Just as researchers never would recommend a single methodology to all clients, neither should they ignore the potential of one method. |
|
|