How to avoid bias in research

One of the most important concerns we have as market researchers is how to minimize the influence researchers have on participants. We want to uncover participants’ real thoughts, feelings and motivations, but what if the researcher’s presence, or even the mere act of asking the question, has an impact on participants’ responses?

Experienced researchers have a variety of methods we use in qualitative research to reduce that impact and obtain the most accurate input possible. For example:

• Setting the tone. The intro at the very beginning of the discussion is an important tool to minimize bias. The first thing I tell participants is “I’m an independent researcher. That means you can say good things or bad things and it won’t hurt my feelings either way. My job is to get people’s honest opinions.” (The late great Naomi Henderson advised against saying “there are no right or wrong answers,” since that brings up the idea of wrong answers. I agree!)
• The intro is also the time to set the ground rules to avoid having one participant dominate the conversation, which can also introduce bias. I say “I want to hear from everyone and keep it even, so if you notice you’re going first a few times, I’m going to ask you to hold back and let someone else go first.” If a participant later starts to dominate the conversation, they can be reminded of this ground rule.
• Neutral language. When our clients have questions, they tend to be unconsciously biased in favor of their product. For example, they may want to know how a certain feature has improved customers’ experience. Instead of asking questions like “how has this this feature improved your experience?” I prefer questions like “What impact, if any, did this feature have on your experience?”
• One exception: when we want to make sure participants aren’t saying things to please us, we might play devil’s advocate to see how strongly they feel. For example, “Is that feature really that helpful? Someone was just saying in the last group that it made no difference to them. Were they wrong?” This gets participants to provide more detail about why they feel the way they feel.
• Watching our own signals. The moderator can bias things without realizing by smiling or saying something like “great answer” to a participant. We really have to watch ourselves. I try to thank everyone for their input, say things like “that’s an interesting point” and “does anyone agree or disagree?” to make sure I appear neutral. And the first time someone disagrees with another participant, I’ll praise them and say something like “that’s what I was talking about earlier – we can feel free to disagree with each other in a friendly manner.”

While we can’t completely eliminate researcher influence, the above strategies can help a great deal. As third-party researchers, one benefit we bring to our clients is the fact that we come from the outside and are therefore more able to be neutral when conducting research. It’s important to approach research discussions with curiosity and empathy, rather than advance expectations about participants’ responses.

Let’s discuss how to elicit unbiased information from your customers and prospects. Email me at info at bureauwest.com.