Survey
by Janet Waters
Planning the Survey:
Choose a topic for your survey. Before you begin drafting your questions, read about the topic (e.g. the chapter in your text). Check your text web site for some suggestions for topics & questions. The first thing you must do is to determine the research question - what is it you wish to know? Keep your topic in mind & make sure all your questions are relevant to it.
Guidelines for Drafting the Survey: Some guidelines for drafting the questions on your survey:
Start with a brief introduction: Briefly introduce your topic & who you are, but avoid overly directing your subjects' answers. You may then ask demographic questions (gender, age, etc.). Either at the beginning or end (or both), make sure you include a thank you to the participant for their time!
Your questions should be clear, understandable, & inoffensive: Make sure your questions are clear, that they ask what you mean them to ask, & are unlikely to be misunderstood. Also, ensure that they are unbiased, inoffensive in their language, & easily understandable by your participants (not too technical, difficult, or unfamiliar).
Don't be too obvious: Carefully word your questions to be as innocuous and unthreatening as possible, & try not to "lead" your subjects to the answers you are looking for. This is particularly important when your topic is a sensitive one, which subjects may find embarrassing, or one which they are likely to be dishonest about.
Check the order of your questions: In general, it's a good idea to ask the most important & interesting questions early in the survey (before the subject gets too bored or impatient with the questionnaire). However, you may need to gradually lead up to your most important question, so do use your judgement.
Keep the survey short: Maximum 2 pages, in most cases.
Choose the best response alternatives: There are numerous response alternatives you can use; use the one that seems most appropriate for your questions. For example, to the question: "You and your parents argue ....." Response sets could be closed, which is easier to score. For example you could give a) a list of possible answers to the question (e.g. several times daily, daily, several times a week, weekly, etc.). Or b) frequency rates: (e.g. all the time, a lot, often, frequently, sometimes, seldom, etc.). Or c) frequency scores on an interval scale: (draw a scale of 1 to 5 (or 7), 1 being "rarely", and 7 being "most of the time"). Or questions could be open ended (leave an empty line for the subject to fill in themselves). Each method of responding has its advantages & limitations.
Construction of the survey: Make sure your survey is typed, has no embarrassing spelling or grammatical errors, & is laid out in a way that is clear & easy to read.
Pretest your survey: After you've constructed a draft of the questionnaire, pretest the survey with a willing "confederate", & ask for his/her feedback about their understanding of the questions, response alternatives, etc. Ask them to "think aloud" while answering the questions, & ask them how they interpreted each question & the list of possible answers.
Conducting the Survey:
Sample size: Generally survey a minimum of 30 subjects. You will need more if you want to do a comparison of genders, ages, etc. (e.g. 20-30 males & 20-30 females).
Sample source: It's crucial to be as unbiased as possible in your sample source. However, in some cases (e.g. when you need permission from a parent or guardian to do any research study with minors) an unbiased random sample will likely be impossible to find. In that case, do note in your research report that your sample was biased, & your findings may not reflect the opinions of the average member of that population.
During your survey:
Try not to subtly bias the respondent's answers by inadvertently showing approval or disapproval of certain answers, or by hinting in any way at the answer you are expecting. If you survey your subjects individually or in several small groups, make sure you introduce & explain the survey the same way each time you give it out.
Analyzing the Survey:
Once you have your completed surveys, you will have to make some sense of your data, & present your results in the Results section of your report. Depending on your response alternatives, you could total the answers & present the totals for each answer as a frequency or a percentage in a table form, or code & compile the open-ended answers. If you have interval scales (a scale from 1 to 7, for example), you can compute means & standard deviations (easy using Excel), & present these in a numerical form & in a graph or table. You may even be able to do correlations between pairs of numerical variables (with Excel).
Presenting & Discussing your Survey:
The standard APA style lab report can be used to present your survey. In the Introduction, briefly review past research & theory in your topic question (e.g. briefly summarize current research on stress). Use APA referencing style to cite your sources. In the Method section, under Participants, present a general description of the participants (Mean age, gender, sample source, etc). In the Materials section, describe your survey. Include a copy of the actual survey in an Appendix of the report, & refer the reader to the Appendix in the Materials section. In the Procedure section, note your general research strategy was a survey, & describe your methods of data collection (e.g. in person vs telephone or on-line survey, etc.).
In the Results section of the report, present your Results as noted above. Use a table if possible, & add a verbal description of your results. Generally, avoid interpretation of your findings case in the Results section. Simply present the findings here; discussion of them occurs in the Discussion section.
Finally, in the Discussion section, you should integrate your findings to theory & past research findings in your topic. Relate your survey findings to what you have learned from the text or other sources about that topic. For example, you could compare your participants' reports about the causes & effects of stress to what research in the text found. If your findings contradict current research or theory, or if you observed something unusual or unexpected, you could suggest reasons why your observations may differ from the expected findings.
@Janet Waters