Anonymity

Anonymity is denned somewhat differently in survey research than in its more general use. According to the American Heritage Dictionary, anonymity is the quality or state of being unknown or unacknowledged. However, in survey research, the concept is more complex and open to interpretation by the various organizations that conduct surveys.

In the form closest to the standard definition, anonymity refers to data collected from respondents who are completely unknown to anyone associated with the survey. That is, only the respondent knows that he or she participated in the survey, and the survey researcher can not identify the participants. More often, anonymity refers to data collected in surveys in which the respondents are de-identified and all possible identifying characteristics are separated from the publicly available data. Many survey research organizations provide data and data summaries to individuals outside their organizations. These data are considered anonymous if those outside the survey organization cannot identify the survey participants.

However, for many surveys denned as anonymous, the survey organization could, if needed, identify the respondents. For example, in a survey that uses pure random-digit dial procedures, limited information about the respondent is available to the survey organization. Through the use of various databases, the organization could possibly determine the household associated with the telephone number. Survey organizations would rarely do that.

Survey researchers have developed a number of procedures for designing anonymous surveys. For example, many surveys conducted in classrooms or other gathered events use unnumbered questionnaires and do not contain questions that could identify respondents. For some classroom surveys, identifying information is collected on a sheet separate from the questionnaire.

A procedure sometimes used in postal surveys is to include a return postcard along with return envelope. The unnumbered questionnaire is returned in the envelope, and the postcard is sent separately to let the researchers know that the questionnaire has been returned.

Survey researchers have developed many techniques for conducting completely anonymous surveys. For example, Internet surveys offer multiple methods for anonymous participation. Some surveys may not require authentication to access the survey. Invitations are sent to potential participants but with no control over who participates nor how often. A more sophisticated recruitment method is to completely separate the database used for authentication from the database that contains the survey responses. Another method is for one organization to send the recruiting requests and a second to collect the data.

A similar method can be used for telephone surveys. The telephone numbers can be stored in a database that has no direct link to the survey responses. This method can be used with random-digit dial telephone number samples to further separate the identifying information from the survey responses.

However, the procedures for ensuring anonymity can conflict with other important survey quality control procedures. For example, sending unnumbered paper questionnaires with postcards in postal surveys allows respondents to return the questionnaires but not the postcard. As a result, follow-up requests cannot be limited to nonrespondents only. Respondents who did not return the postcards may believe their [Page 28]first questionnaire did not reach the survey organization and respond a second time.

...