Internet Data Collection

Active Data Solicitation

A researcher using active Internet data collection methods actively targets specific individuals or identifiable, restricted populations and directly asks for their opinion, viewpoint, or feedback. Examples of the type of population that might be targeted by active data solicitation include everyone who attended an [Page 555]infection control training session, all the pediatric physicians on a listserv, and all new enrollees in a WIC program (the federal nutritional supplement program for women, infants, and children). Active data solicitation includes techniques such as Web and e-mail surveys where the researcher sends an e-mail with an embedded electronic survey or sends a link to a Web site containing the survey. In some instances, the link for the Web site could be mailed or presented (e.g., at a conference) for individuals to access directly. Regardless of how potential responders get to the survey, active data solicitation requires the researcher to get respondents to use their own time to provide information to the researcher. In addition, all respondents must have sufficient access to a computer and the Internet to be able to respond. A current uncertainty that often arises with this method lies in understanding the response rate, that is, the relationship between the number of potential respondents and the number of those who actually participated. The researcher knows how many people responded to the survey, which yields the numerator, but what is the denominator? To put it another way, how can you determine the number of individuals who are part of the population you targeted but who did not respond to your survey? This is less of a problem when a survey is e-mailed to specified recipients, although there may still be issues such as those raised by inactive e-mail addresses. However, when a survey is posted on a listserv, calculating the response rate presents a major problem. In many cases, the researcher does not know how many people actively read and participate in a particular listserv, so it is unclear what number should be used for the denominator in calculating the response rate. In addition, standards for acceptable response rates for Internet surveys still need to be established.

Passive Data Solicitation

Passive electronic data solicitation, in contrast, does not target specific respondents. For example, a researcher could post a link to his or her survey on a popular news Web site so that anyone accessing the Web site could complete the survey. In this example, the researcher does not specifically solicit information from a known person or restricted population; in fact, the respondents are self-selected rather than targeted by the researcher, although there is a general concept of targeting in the choice of the Web site where the link is posted. For instance, some Web sites are more likely to be accessed by persons with a liberal or progressive political point of view, others by persons with a conservative political point of view. As with active data solicitation, respondents are still required to spend their own time and effort to complete the survey.

...