Mail Questionnaire

Mailing QUESTIONNAIRES to PROBABILITY SAMPLES can be a cost-effective method to collect information when addresses are available for the POPULATION of interest (e.g., individuals or establishments). Single-mode surveys can be designed that employ mail questionnaires only, or these self-administered questionnaires may be part of a mixed-mode approach that includes other modes of administration (e.g., telephone, Web-based, or in-person INTERVIEWS). A frequently employed dual-mode design is to call sample members who have not returned mailed questionnaires to attempt interviews by telephone. Multimode designs can help reduce NONRESPONSE and often increase the representativeness of the resulting sample.

The SAMPLING FRAME for any RESEARCH DESIGN that includes a mail component can be a set of addresses, such as a list of employees of a firm or members of a professional organization. Note that the lists of addresses in telephone books cannot be relied on for sampling for population studies. The representativeness of samples drawn from telephone listings is compromised by unlisted numbers and households that either have no telephone service or rely solely on cell phones.

To identify members of rare populations for a mail SURVEY, one might use RANDOM-DIGIT DIALING (RDD) techniques. When a household is identified in which a member of the population of interest lives, the interviewer requests the address and a questionnaire can be mailed. Although the advantage of using RDD is that every household with a telephone has a known chance of selection, many people are unwilling to divulge their addresses to strangers who call on the telephone.

Respondent selection can be another challenge with mailed questionnaires (e.g., in situations when addresses, but not names, are known or when parents are required to randomly select one of their children to report on). There are techniques available for designating a respondent or subject (in the case of PROXY REPORTING), but these designs must be carefully planned, and the quality of the resultant sample depends on respondent compliance with sampling instructions.

Researchers have a toolbox of methods to use to develop survey instruments to collect data that can be used to adequately address their research questions. Instrument development often begins with a literature review to identify existing instruments that address [Page 604]the issues at hand. It continues with focus groups conducted in the population(s) of interest to identify salient issues and the terminology that people use to talk about these topics. Candidate survey items are drafted about these domains of interest, and cognitive interviews, intensive one-on-ones with individuals from the target population, are conducted. The goal of these interviews is to identify survey questions and terminology that are not consistently understood. The instrument can then be revised and retested prior to fielding the survey. Instruments also often undergo expert review prior to launching the survey. Here, authorities in the field of interest are asked to propose domains and comment on the suitability of the planned instrument. The last step before fielding a mail questionnaire is a PRETEST. Individuals with characteristics that mirror the target population can be brought together for group pretests. First, participants complete the questionnaire, and then they are debriefed about their experiences with navigating the instrument and any difficulties they may have encountered in answering the questions. Group pretests also allow an estimation of how long it will take respondents to fill out the form.

...