5 Don'ts For Writing Survey Questions

By: Amanda Barna


Surveys are used for many reasons - to measure employee or customer satisfaction, to gauge public awareness of an agency or service, or evaluate the effectiveness of a program, training or class. Administering a survey can be a tremendous benefit to any organization. Knowing what you do well and how/where you can improve can make a measurable difference.

One of the dangers when putting together a survey is not asking the right questions the right way which can result in data that cannot be used, or, even worse, can lead you in the wrong direction.

Here are five tips of what NOT to do when writing survey questions:

  1. Do NOT use slang, abbreviations or jargon. People may interpret these differently, changing the meaning of the question based on that interpretation. As a general rule, don't use slang, or jargon and identify even commonly used abbreviations. This is becoming more of an issue with the increased use of text messaging and tweeting (survey questions don't need to be asked in 140 character or less). "Have you bought a CD in the past year?" If this question is asked of a young adult it may mean something different than if you ask this question to a banker.
  2. Do NOT write leading questions. These questions can lead people to one answer over another simply because of how it is worded. "You have heard of the Center for Marketing and Opinion Research, haven't you?" People who haven't heard of the agency are more likely to say that they have because the question is worded in a way that makes those who have not heard of the agency feel uninformed, out-of-touch, or even stupid.
  3. Do NOT ask double-barreled questions. In other words, questions that include more than one issue or question. "How would you rate the economy and job opportunities in Ohio?" Not everyone will relate these two issues together so it is best to be asked in two separate questions.
  4. Do NOT ask questions that sound TOO smart. "At what age was your child ambulatory?" This may be appropriate for medical professionals, but not for the general public. Most survey questions for the general public should be written at the 6th grade level. Decide who you intend to administer the survey to and write questions to that audience's level of understanding.
  5. Do NOT use overlapping or unbalanced response choices. Asking a question that does not allow for all possible responses can be confusing and frustrating. Response choices should be:
    • Mutually exclusive- response categories should not overlap. "How many days in the last week did you exercise? None, 1-2, 2-3, 4 or more?" There are two possible response choices for those who have exercised twice.
    • Comprehensive- there should be a choice for all possible responses. "How many days in the last week did you exercise? 1-2, 3-4, 5 or more?" What about those who have not exercised in the past week, how would they answer?
    • Balanced- there should be as many positive options as negative. "How satisfied are you with your job? Very satisfied, satisfied, or somewhat satisfied? What if they are not satisfied?" This question assumes that everyone is satisfied on some level, which usually is not the case.

If you are unsure if you are asking the right questions the right way or if you are using the survey data to make key decisions in terms of programming or expenditures, you should have your survey looked at by a professional. Many research firms will consult with organizations on survey design for a small fee. This small investment can end up saving you time and money in the long run. Remember, a survey is only as good as the questions it contains.