If the way a question is worded leads someone to the idea that a particular response is the “correct” reply or the one that you expect them to give, then it is biased. Similarly, if a question is confusing, then it can cause people to answer in a way that they would not normally respond. This is a major threat to the reliability of your data and will often change the way that participants respond. You can read through the following examples that showcase language that might indicate bias in your questions and how to reword them.
Ethics Check
It’s important to handle sensitive topics carefully in your survey. If your research tackles sensitive topics such as alcohol consumption, law breaking, or death, you need to make sure that you are mindful and considerate of your participants when writing your questions (Choi & Pak, 2005).
For example, if you want to conduct research on mental health during the past winter break, instead of asking the question directly: “Have you been having mental health issues during the past winter break?” You can start by first inquiring about the participant’s health in general, which puts the participant in the proper mindset to answer questions of this variety. First ask “How would you rate your general health: great, good, fair or poor?” Then you can focus on questions more directly related to mental health — “Over the past winter break, how often would you say the following problems have troubled you: nervousness, anxiety or the feeling of being on the edge?” (Household Pulse Survey, 2020).
You can use the tips on this page and the previous ones to make sure that your questions are the best that they can be. Go back and review these tips, or move onto the next section to learn about how to choose which questions you will use in your survey.
