Write the purpose of your survey. For example, you might be gathering public opinion about a recent political candidate's debate performance, or you might be gauging the reaction to a recent movie release.
Identify the people you plan to target with your survey questions. You could limit your survey questions to a limited demographic group, such as middle-aged women. You may also seek to speak with a larger sample that is representative of the socioeconomic, gender and racial populations of a specific area.
Indicate the kinds of questions you plan to ask in your survey. Some surveys are limited to short yes-or-no questions, while other surveys have more in-depth free-form response questions mixed in with multiple-choice questions.
List your expectations for the results. For example, you may expect that a majority of respondents will say that they enjoyed a certain movie. On the other hand, you may expect a minority of respondents watching a political debate to have a negative opinion of a candidate. Make as many viable predictions as possible, writing down the reasoning. This will allow you to analyze differences between the expected and actual results, and modify your expectations for future surveys.
Describe how you plan to use the survey results. For example, you may share the debate survey results with a candidate to modify his speaking style or choice of words. You may report the movie survey data to the film studio, which may then decide whether or not to screen a movie in certain areas.