Survey research is the most common form of public relations research.  Through the use of specifically worded questionnaires and a carefully selected sample, researchers are able to make judgments about a much larger population.  The degree to which survey results can be seen as being representative of a larger population depends on the methodology used.   Survey research is generally conducted in three ways: through face-to-face interviews, over the telephone, or through self-administered questionnaires.   What follows is a checklist for the conduct of survey research:

• Define the problem or purpose. Who are the target audiences in whom you are interested and what is it that you want to find out?

• Identify population and choose sample. To embark upon this step, one has to understand several basic definitions:


>>    Sample - the selection of a segment of a population for the purpose of making observations of and drawing conclusions about the population as a whole.

>>    Sampling Frame - the list from which the sample, or some stage of the sample, is drawn.

>>    Probability Sampling - a sample will be representative of the population from which it is selected if all members of that population have an equal chance of being selected in the sample.

>>    Non-probability Sampling - when a sample is selected without regard to whether everyone in the population has an equal chance of being selected.

     There are several sampling techniques from which you can select:
 

• Simple Random Sampling - every person within the sampling frame is assigned a number.  Through a random selection of numbers, a sample is developed.
• Systematic Sampling - involves the selection of every Kth member of a sampling frame.  For example: if you are seeking a sample of 250 names from a survey frame of 2500, then K=2500/250 or 10.  In this scenario, select every 10th name for the sample.
• Cluster Sampling -  involves breaking the population into heterogeneous clusters and then selecting the sample from individual clusters.  This is often done when segment(s) of the desired population is/are overrepresented in the sampling frame.
• Census - every member of the sampling frame is surveyed.

• Create the survey instrument.  The form a survey instrument is dictated by how it is to be administered and analyzed.  Questions can take four basic forms: Dichotomous (such as true/false, yes/no),  Rating Scale (measures the range, degree or intensity of attitudes), Open-ended (the response set is not specifically defined), and Close-ended (the response set is specifically defined). There are several concerns to be considered when wording survey questions:

• Use simple words - use language that is appropriate and understood by the audience for whom the survey is intended.
• Don't be vague - use language that has a clear or specific meaning.
• Keep it short - respondents may balk at answering long questions or surveys.
• Avoid bias in the wording or ordering of questions - the manner in which a question is worded  or its placement can influence responses to questions that follow.
• Don't ask objectionable questions -  even the most sensitive of information can be obtained if questions are worded tactfully.
• Save the toughest questions for last - asking the toughest questions first could abruptly end the process or bias subsequent responses.
• Don't assume knowledge - the respondent may need  background information.
• Pretest the questionnaire - if there are any bugs in the survey  instrument, it is best to find them before distributing surveys on a large scale.  Run a small test sample first to see if the questions are understood.
• Logistical concerns - Can the personnel requirements be met?  Can the survey be administered within the desired time frame?  Does it fit into the budget?

• Train data collectors. Without training, data collectors could inadvertently influence the outcome of the survey.  They should be trained to administer the survey the same way every time.

• Compile and analyze raw data.