Informant

An informant in a survey is someone asked to provide information about another person, persons, or an organization, for example, when a parent is interviewed about a young child who could not answer the survey questions. Informants (also known as proxies) tend to be used in surveys when the target respondent is unable to respond or when it is not feasible to collect responses from all members of a group under study. As the use of informants to collect quantitative data has become integral to survey research, due to the cost-effectiveness of the approach, so has the study of the effects of using informants on the data that are collected. The substitution of informants limits the types of data that can be collected with accuracy, and proper selection methods should be used to minimize the resulting response bias that has been noted to occur occasionally when data collected from informants are compared to self-reported data.

There are several types of surveys in which the use of informants is the most efficient means of collecting responses. Informants are frequently used in surveys when members of the population under study are unable to provide responses because a physical or cognitive impairment prevents them from responding. Because informants are the only way to collect information about these target populations, informants are used despite the fact that several methodological experiments have shown that using informants produces response bias. Specifically, in surveys asking about disability, informants tend to overreport more obvious types of disability (such as difficulty with activities of daily living) and to underreport less obvious types (such as mental health problems).

Another survey situation in which the informant method of data collection has been used is when it is not economically feasible to interview all individual respondents, such as all members of a household or an organization. Studies using this method generally develop selection rules that ensure that the selected informant is likely to be able to provide accurate responses about others in the household or [Page 335]organization. The selection rules used by the survey are applied (e.g. randomly selecting one of the adults in the household), and the selected informant reports on the behaviors of other members of the household or organization. Further, surveys asking questions on sensitive topics have used informants to reduce bias from self-reported responses due to social desirability. These include surveys collecting information about sensitive behaviors such as alcohol, tobacco, and drug use, which tend to be under-reported by respondents. For example, the results of some experiments suggest that the informant method yields estimates of alcohol use that are closer to actual alcohol sales figures than self-reported data.

The ability of selected informants to respond accurately to surveys depends on how observable the survey subjects are by the informant and on the informant's ability to recall events. The survey topics asked of informants must take these issues into account, as informants are best able to provide accurate information about others when the informant has a high degree of knowledge about those he or she is answering questions about. For this reason, topic areas in surveys using informants that include factual or “hard” measures are preferable to “soft” measures requiring subjective evaluation by those providing responses. For example, in a survey administered to principals of an organization, demographic questions answered by an informant on “hard” topics such as age, prior experience, or race will likely produce lower item nonresponse and be more reliable than topics less likely to be observable to informants, such as net worth or total household income.

...