Eligibility

Eligibility refers to whether or not a sampled unit is eligible to have data gathered from it—that is, is the unit part of the survey's target population or is it not? For example, the target population for a survey might be all adults who are 18–34 years of age. As such, if a household is sampled and screened via random-digit dialing and no one living there fits the age criteria, then the household is ineligible. If there is at least one person ages 18–34 years, then the household is eligible. Ultimately, eligibility versus ineligibility is central to the issue of how well a sampling frame and a sample drawn from that frame “cover” the target population and whether or not coverage error results. Eligibility also is linked to survey costs, since samples drawn from frames that contain a large portion of ineligible units are much more costly to process. As straightforward as it may appear to determine eligibility for a survey, it often is not at all easy to do and many mistakes (errors) may occur in the process. Mistakes in determining eligibility may lead to coverage bias.

For example, most surveys have geopolitical boundaries for their samples, as they are not national in scope. In each of these surveys, the target population typically is limited to those residents living within the geopolitical boundaries (e.g. a particular county). If the mode of sampling and data collection is the telephone, as it often is, then some form of geographic screening must be instituted for interviewers to determine the eligibility of the household or person being contacted. In the case of boundaries that are commonly known and well understood by the public (e.g. one's county of residence), eligibility is readily determined without many errors, as long as the respondent does not know what answer will make her or him eligible or ineligible for the survey. (If the respondent knows this in advance of answering the screening questions, some respondents will self-select themselves in or out of the interview erroneously.) On the other hand, if the geographic boundaries that define eligible residency are not well know, (e.g. a school district or a police district), then screening a sample for eligibility via the telephone can be fraught with error. Some people will mistakenly say, “Yes,” they live within the boundaries when asked the screening question(s) when in fact they should have said, “No” (errors of commission), and others will say, “No,” when they should have said, “Yes” (errors of omission). This will occur even if the screening sequence carefully defines the eligibility boundaries, because many people are “geographically challenged” and will not understand the boundaries they are being asked about.

There are many other criteria than geography that are used to define eligibility in various surveys. For example, a major health survey conducted annually since 1994 for the U.S. government interviews only parents of children ages 19–35 months to learn about the immunization history of the children. (The exact eligibility definition for a household in this survey essentially changes every day of interviewing, since children age each day.) Researchers therefore must pay very careful attention to how eligibility is denned and how it is explained to respondents, whether that be in an interviewer-administered survey (e.g. in person or telephone) or via a self-administered survey (e.g. via mail or Internet).

...