In order to obtain accurate and factual data, a survey instrument must be well constructed and a survey planned carefully and conducted under strict conditions. Of importance in planning a survey is a determination of previous and on-going research through an extensive literature review. In designing the survey, the researcher should aim at precision, logic, and efficiency in translating research objectives into specific questions, and at obtaining a large percentage of reliable returns. The subject of the research should justify the effort and the information should be otherwise inaccessible. The questionnaire should be brief, contain clear instructions, be organized in a neat and logical order that facilitates progress from question to question, and should be worded understandably and objectively, but in such a way as to be challenging and interesting. Various questionnaire styles--"closed,""open,""agree/disagree"--have particular advantages and disadvantages that should be considered when selecting one. In distributing the questionnaire, it is essential to identify the exact nature of the study population and the representativeness of the sample chosen. Questionnaires should be mailed at the same time, and first returns should be expected within a week. Follow-up procedures should begin at the end of the third week. Determination of the appropriateness of computer use in processing data should be included in original planning. (AYC)