Re: "Reporting participation in epidemiologic studies: a survey of practice".

We read with great interest the article by Morton et al. (1) that addressed the alarming trend in declining participation rates for epidemiologic studies over recent years. The authors should be commended for addressing such an important issue. In addition to the possible reasons for decline listed by the authors, we would like to comment on another potential reason that may fall under their broad category of ‘‘societal factors’’ (1, p. 202). Mail and telephone marketing firms, along with various nonprofit organizations, have developed highly deceptive ‘‘survey marketing strategies’’ for eliciting purchases or donations. They are disguised as surveys and begin as such but, unlike standard market research surveys, they lead into either a request for a donation in support of a particular group or cause or questions about whether the respondent would like to purchase an alternative product or service. It is likely that use of these strategies negatively impacts the public and is one reason people are distrustful of surveys in general. Consequently, when contacted by various universities, government agencies, and other organizations that are conducting legitimate surveys for research purposes only, potential subjects refuse to participate. As scientists who depend on public participation for many of our research endeavors, we can only continue to support the use of surveys solely for research purposes and refuse to support ‘‘surveys’’ conducted for any other purpose. That said, it would be useful for epidemiologists to empirically document methodologies that boost participation rates and to ensure that participation rates associated with various study designs are consistently reported. Morton et al. (1) did not discuss differences in participation rates across studies by data collection method (i.e., in person, telephone, or mail) or by use of procedures specifically designed to increase cooperation, such as introductory letters prior to data collection contact. Since the sharpest decline in participation is among population controls in case-control studies, epidemiologists may need to move away from the use of random population control selection. Future research should examine control selection methods that may be comparatively more viable, such as the use of family, friends, and neighbors—a method discussed by Morton et al. (1). Such ‘‘snowball sampling’’ techniques have been used effectively by anthropologists and sociologists for many years. Morton et al. (1) have identified a disturbing trend and, in so doing, have raised awareness of it. Future research needs to identify the most effective strategies to curb this trend.