Writing a survey can seem like a walk in the park. After all, how hard could it be to write a decent question and set of responses to that question?
Honestly, it’s harder than it might seem. Here are some tips for how to write a great question.
What exactly are you trying to ask?
First things first: what is the primary purpose for your survey? This is where it all starts. If you can focus this, half your work is done. I’ve sat in client meetings where a myriad of purposes were presented, and we had to work at determining what was the most important thing they wanted to learn from the questionnaire. It’s entirely natural to see writing a questionnaire as the opportunity to have an in-depth discussion with your respondent, so to speak. However, the longer the questionnaire, the less likely your respondent is to make it all the way through.
Once you have your primary purpose identified, it’s time to brainstorm the questions. This might seem contradictory, but let the sky be the limit here. List everything you can possibly think to ask.
Whittle it down
Now that you have your brainstorm list of questions, it’s time to go back to that primary purpose and eliminate anything that doesn’t meet the purpose. Next, did you ask the same thing multiple times? There are times when this is useful, but a basic survey should eliminate redundancy as much as possible. You want to keep this questionnaire as straight-forward as possible.
Some phraseology tips
Time to work on the phrasing of your questions. This is where writing a good question can be tough. Here’s some stuff you want to avoid:
- Bias. This is one of the most difficult aspects of writing a good question for a survey. Introducing bias is easily done without even realizing it. Here’s an example of a question that introduces bias: “I enjoyed my experience.” The answer options would likely be an agreement scale. By already using a positive sounding statement, you’ve introduced bias; the respondent is more likely to answer positively.
- Scales that don’t match. Avoid using answer options that don’t match the question being asked. For example, don’t use the “likely” scale when the question has to do with whether the respondent agrees with something or not.
- Double-header questions. There’s an easy way to determine if you’re asking two questions in one — check if you’ve used the word “and” or “or.” For example: “What did you think of our communications and customer service?” Another example: “Was the support you received helpful and on-time?” In both examples, you’ve asked two items in the same question. Split them out into separate questions, or just choose one.
- Using too many words. If your question needs to be written as a full essay before you can show the answers, you might want to re-think your question. Please note this also includes writing out a scenario to set the stage for the respondent. Try to keep everything as succinct as possible while still accomplishing your objective.
- Using terminology, your audience doesn’t understand. In any organization, it’s easy to end up using terms that you understand, but that your audience doesn’t understand. If your questionnaire is being designed for an internal audience, that’s fine, but avoid it otherwise.
- TLAs. Not spelling out acronyms can leave your audience in the dark. Spell out any acronyms used, even if it seems like it’s obvious what the acronym stands for: many acronyms have multiple meanings, and you want to be clear which one you mean in your survey. (By the way, TLA = three letter acronym.)