|
|
|
Steps to Carry Out a Survey
Objectives of the study. The first step in any survey is specifying what needs to be learned. What is the survey supposed to accomplish? In concert with the client, firmly establish exactly what information is needed from the survey, why it is needed, and what the results will be used for. The first draft of the questionnaire is based on this information. Sample design. The population to be studied must first be identified. A good sample frame -- the population available for sampling -- must be devised. Then sampling decisions are made, and the sample is drawn. Whenever the results are to be generalized beyond the sample, probability (random) samples are required. Questionnaire construction. A questionnaire is prepared to carefully translate specific issues into questions understandable to the respondents. Questionnaires are designed to maximize validity and reliability (e.g., via multiple measures for key variables; clear, unambiguous wording for every question). Use mostly closed-ended questions because they have proven to be more efficient and more reliable than open-ended questions, but include one or two open-ended questions in the survey to allow respondents to fully express their own opinions on central issues. In writing each question, follow well established rules: ensure the question is relevant to the objectives of the project; make sure respondents have the knowledge, opinions, or experience needed to answer; check whether there is a standard question used by other surveyors; use appropriate vocabulary; use correct grammar; ensure the question has a single, clear focus; try to limit the question length to 20 words or less; avoid biased words or phrases; include positive and negative alternatives in the question; ensure the response format fits the question; ensure answer categories are exhaustive and mutually exclusive; include clear directions for answering each question; anticipate the needs of the analysis; and ask what we are taking for granted in wording the question. The draft questionnaire is pre-tested. Gather data. The usual methods of collecting data are phone surveys and mail surveys. The choice between methods depends on such factors as the time frame, available budget, information available on the sample, literacy level of the sample, and so on. Edit and enter data. Once the data have been collected they must be checked thoroughly for errors, coded for computer entry, and computerized. Analysis. Analysis begins at a descriptive level and moves to a statistical analysis of inter-relationships among key variables.
|
|
André Paradis Tel: (905) 335-0307 Fax: (905) 335-0080 andre@measureresults.com Webmaster |