Questionnaires/Surveys

Where to Start

What information do you need?

Which group will you use to gather the information? The group should be designed to be representative of the population being surveyed.

How will you apply the survey? Will the questionnaire be self-administered or interviewer-administered?

How will the questionnaire be delivered?

Design the survey.

Do a pre-test.

Collect data.

Analyze data.

Know Your Audience

Match the survey to the audience

Shorter surveys are better than longer ones

Keep questions clear and concise

Avoid use of technical words

Questions

Make the survey easy to answer. Questions that are closed-end, with answer blanks to be checked or circled facilitate responses and are easier to tabulate.

Closed-end question: "How many hours did you watch television yesterday?"

Closed-end response (circle one): none 1 2 3 4 5 6 or more

Open–ended question: "What benefits or advantages do you see in surveys or opinion polls?"

Open-ended questions take a lot of time to get entered into the computer. Generally, try to phrase an open-ended question as several smaller close-ended questions. For example, instead of asking:

"What is your opinion of the President?"

This question could be rephrased into: "The President is a doing a good job running the country."

Strongly Agree Somewhat Agree Somewhat Disagree Strongly Disagree

There are times when an open-ended question is unavoidable. Try to hold those until the end of the survey, and minimize the total number of them.

Provide for all likely answers. Although some questions can be answered by a yes or no, many are somewhere in between. Provide for categories as "I don’t know" or "no opinion". For some questions, allowing for shading of opinions, such as "strongly agree, agree, no opinion or strongly disagree" permits more descriptive responses.

Ask only for information that the respondent is likely to remember.

Keep the questionnaire as short as possible by asking only for needed information. Remember, the respondent is doing you a favor, and the longer the questionnaire, the less likely the respondent will be to make thoughtful replies

Avoid questions that touch on personal bias or those that appear to be prying. Such areas include age, income, education, morals and personal habits. While some types of personal questions may be necessary, limit questions to answering within ranges ( e.g., between the ages of 20-30).

Arrange questions in a logical order. Hopping back and forth between questions can confuse respondents.

Avoid leading/biased questions that by their wording suggest an answer. For example, "Don’t you believe that business majors earn the largest salaries?" leads the respondents to answer yes.

Avoid double-barreled questions. " Do you believe that air bags are unsafe and expensive?" The respondent may answer no when he thinks that air bags are safe and expensive. Solve the problem by breaking this into two questions.

Avoid the halo effect. Watch for questions that link a position with particular person or group. The respondent ’s attitude toward the person or group may influence their attitude toward the question. "Do you support Governor Bush’s plan to reorganize the Florida Education Department?" A better question would be, "Do you support the reorganization of the Florida Education Department?"

Provide specialized instructions where appropriate. If respondents should select only one answer, rank items, or skip to a certain question based on the current question, special directions must be provided to direct respondents correctly.

Provide for the respondent’s additional comments, if any, that allow for comments in the respondent’s own words. In addition, open-ended questions may encourage the respondent to speak more freely.

Pre- test

Try the questions out on many different people as the questions are evolving to identify weaknesses and potential difficulties. It is better to be aware of problems after the first five surveys are completed than when you have already collected fifty surveys. Surveys may have problems with unclear wording, confusing design, questions that are difficult to understand and even typos that could be corrected by a pilot test.

Steps:

Find five respondents. Be sure that they will fit the profile of actual respondents, which you will be questioning later.

Have the respondents complete the survey without any special instructions.

Speak with each respondent to gauge his or her response to the survey. Was it clear and easy to understand? Were they comfortable answering the questions?

Record the answers

Analyze the data collected and decide if this is the kind of information you wished to receive. Are there responses, which should have been included but were not? If you find that too many questions were answered with N/A or ‘don’t know’, then you may have to reevaluate the survey

Summarize data and discuss with group or team

Edit survey as needed.

Review the survey as a whole:

Does it convey to the respondent what the survey is about?

Does it indicate why the questions are being asked?

Does it include interesting and easily answerable questions to gain attention?

Does the conclusion express thanks for the respondent’s time and efforts?