Quick Tips to Improve Your Results
• • • • • •
There are many ways to improve your data quality by rewording your questions to remove the opportunity for bias, test your respondents’ awareness, and keep them from being overwhelmed by choices. Here are our top tips for making sure you get the best results possible:
Never ask yes/no questions.
Asking a yes/no question essentially telegraphs to respondents what you want to hear. By selecting yes, they know that they’ll get more questions from you, which incentivizes them to misrepresent the truth. Instead, ask a question with a variety of answers that requires users to consider all options.
Don’t ask users to “select all that apply.”
Often, the cognitive burden of considering every single answer option can lead respondents to pick incorrect or unlikely answer options. Try asking them to pick their top three instead.
Keep answer options limited.
It can be tempting to include a comprehensive list of all brands or products in a category, but that amount of information can get overwhelming very quickly. Instead, try providing a maximum of 10 answers.
Reduce lookback time frames.
Respondents may have a difficult time remembering what they did a year ago or longer. Try limiting the time frame of a question to three or six months. For example, “Which of the following brands of toilet paper have you purchased in the last three months?”
Include “none of the above” and “other” options.
It’s possible that out of a list of all options, that some respondents may not find something that applies to them. Be sure to include None of the Above and/or Other options. Otherwise, you may be forcing a consumer to lie, thus skewing your results.
Randomize answer options.
People are often predisposed to selecting the first answer option in a list just because it happens to be the one on top. When setting up a multiple-choice question in Suzy, toggling on answer randomization means that each respondent will get options in a different order, which reduces this bias in your results.
Include “ghost” or “red herring” options.
Asking a test question with an obviously fake answer option is a simple way to screen out respondents who don’t or can’t answer your question faithfully.
Be as detailed in your question as possible.
It may be tempting to ask an open-ended question like “Which of the following brands have you seen lately?”, but the wording is so ambiguous that respondents may interpret your meaning differently. Instead, try asking something like, “Which of the following brands have you seen advertisements for in the last three months?” so there’s no ambiguity around what you’re asking.