7 Tips for Designing a Great Survey

7 Tips for Designing a Great Survey


By Sofia MacGregor | Survey Methodologist

Surveys are a great way to gather reliable and usable data. They can provide valuable insight into almost any topic. When you create a survey you should always go through a few rounds of edits and testing to ensure the survey design is appropriate. Here are a few tips to help you design your next survey: 

  1. Make sure the questions are relevant to the main topic and other questionnaire topics. Check that all question are related to the main topic and avoid all unnecessary questions. Each question should be set against the other questions to make sure it doesn’t affect the others. (Some questions might create prejudice against other questions).
  2. Avoid double-barreled questions. Avoid items asked in negative form, they can puzzle respondents. An example of a double-barreled question is: “Do you think the government should increase its spending on education and national defense in 2020?” Education and national defense spending should be separated into two different questions.  
  3. Avoid negative questions. Avoid items asked in negative form. They confuse and frustrate respondents. Here is an example of a negative question: “Do you agree or disagree with the following statement: smoking should not be allowed in closed spaces.” The “disagree” option might confuse respondents. Instead, you could phrase the statement, “smoking should be allowed in closed spaces”. In this case, both the “agree” and “disagree options have straightforward meanings.
  4. Avoid leading, loaded or biased questions. A loaded question is one that contains words that are emotionally charged. A leading question directs the respondent toward a specific answer. A biased question is one that emphasizes one of the answers over the others. Here is an example of a leading question: “Do you agree that companies should do everything they can to reduce their carbon footprint ?” A better way to ask this particular question would be, “Do you think companies should find ways to reduce their carbon footprint ?” An example of a biased question is, “Buying local produce contributes to the wellbeing of the American economy. Do you prefer buying locally grown produce over imported produce?
  5. Use simple and direct language. Long questions can easily burden respondents. Short and concise is the ideal way to go.
  6. Make sure all possible answers are included. When using closed-ended questions the answers should be exhaustive and mutually exclusive. If in doubt, allow respondents to add information, typically phrased as “Other______.” 
  7. Use simple items to measure complex concepts. Use common words. Avoid jargon and professional concepts when writing questions. You can triangulate items to introduce complex concepts and ensure reliability. This strategy should be used when exploring a complex concept. A triangulation strategy consists of presenting several questions that explore the same concept, but each item is constructed or worded differently. When using triangulation, it’s always best to include an open-ended item before a close-ended one. And finally, always test for consistency and reliability.

See how Premise can help collect targeted questions and gauge attitudes on a range of questions. Visit our Sentiment & Surveys Solution or email us at info@premise.com

About Sofia MacGregor

Sofia MacGregor is a Survey Methodologist at Premise Data. She helps to formulate surveys, design experiments and incorporate academic best practices into the work. Before joining Premise, she was in academia both as a researcher and graduate student. When she’s not working, she likes to explore trails and take pictures.