Survey

Surveys are instruments that are used to collect data on a wide range of behavioral, attitudinal, and cognitive issues, including respondent perceptions, attitudes, beliefs, values, purchasing practices, and other behavioral intentions. Surveys consist of a series of questions about an issue of interest to the practitioner. Practitioners may use surveys to gauge stakeholders' opinions on an issue relevant to an organization's practices, determine the effectiveness of a campaign, measure the community's perceptions of risks associated with a new production process, or assess an organization's reputation.

Surveys typically involve securing responses from a subset of the population of interest (called a sample) to which the practitioners hope to generalize. Surveys are the most popular research methodology in public relations due to their relative ease of construction, administration, tabulation, and analysis. However, practitioners must be careful when constructing survey questions and deciding who should complete the surveys (sampling) and how (e.g., selfadministered or administered by a trained interviewer). Practitioners must also decide whether their research design is cross-sectional or longitudinal.

As in all research, reliability (providing consistent measures in similar situations) and validity (responses correspond to what they are intended to measure) are concerns. The goal is to have the questions and response options mean the same thing to all respondents. Practitioners need to be able to attribute differences in answers to differences in respondents rather than to possible extraneous factors such as educational level, technical expertise, or interviewer style, which introduce error. Threats to reliability and validity can produce inaccurate results. For this reason, careful attention should be devoted to the words used in the questions and response options and how these are phrased. Technical jargon, ill-defined terms, words with multiple meanings, and loaded (“red flag”) words can create problems related to the interpretation of and response to questions and the selection of answer options. When the possibility for misunderstanding exists, precise definitions should be provided.

Other problems occur when questions are too complex and ask for several pieces of information in a single question. This often occurs when two issues contained in a question are connected by and. For example, if respondents were asked to agree or disagree with the statement “XYZ Corporation is an asset to the community and provides needed employment for the community,” they are being asked to respond to two issues (XYZ Corporation as an asset to community and as employer). Other problems arise when questions are leading and seem to suggest the desired response. For instance, the question “Is school violence the biggest problem facing our public schools today?” seems to suggest that yes is the desired response.

Finally, the method of survey administration may affect the reliability and validity of the results. There are two methods of survey administration. First, surveys may be self-administered through the use of a questionnaire that is mailed and returned, posted and returned via the Internet, or distributed to the desired sample in a particular setting and collected (e.g., distributed and collected at a town hall meeting, an information-giving session for employees, etc.). Respondents complete these surveys themselves. Error may arise from their failure to read questions thoroughly, their skimming through answer options, and their difficulty in recording responses. Additionally, there is no guarantee that the desired respondent completed the survey. Second, surveys may be administered by trained interviewers through face-to-face or telephone interviews. This method is more expensive due to the costs of training interviewers and compensating them for their work. However, training interviewers to be consistent in their behaviors increases the uniformity and accuracy of survey completion.

...