Our Proven Process: How to Avoid Bias in Qualitative Research

Apr 15, 2024

How to avoid bias in qualitative research is an important question to consider in your quest to understand consumers and the humans behind them. Especially since bias can slip in from many directions.

Here's one example:

Our Chief Insights Seeker Isabelle Landreville was in the middle of a tricky situation. She was leading a focus group testing reactions to a hearty breakfast advertisement featuring construction workers. Tensions began mounting between a blue-collar worker and a white-collar participant. The white-collar man started mouthing off about how construction workers are uneducated and disrespectful. Isabelle could see this bias taking hold in the group. She knew she had to step in - not to suppress his views but to avoid letting one person’s assumptions colour everyone else's perspectives.

Managing this kind of bias is an art moderators must master to gather objective insights.

“The key is to get insights that have power and clarity while also understanding where different people in the group come from,” said Rami El-Khatib, Resident Storyteller at Sylvestre & Co. “And we do that by fostering an environment where everyone feels heard, without allowing any single person to dominate the discussion. This involves actively managing and mitigating biases”.  

The types of moderator bias in research

Before exploring solutions, it helps to understand the types of biases:

There are two key categories – bias while moderating and during the analysis.  

Bias while moderating

Confirmation bias: Seeking or interpreting information to confirm pre-existing assumptions. A moderator may unconsciously steer conversations to prove hypotheses they already hold.

Observer bias: Influencing responses through body language and facial expressions. If a moderator smiles and nods every time someone complains about a product, it could skew perspectives. Moderators can mitigate this through the best body language, tone and facial expressions.

Availability misweighing: Assuming easily recalled singular responses represent the whole story. If one person mentions an issue, a moderator shouldn’t assume that issue is widespread without more evidence.

Social desirability bias: Participants providing answers they think sound best rather than what they actually believe. This often shows up around sensitive topics when respondents want to save face. Moderators can build rapport, ask indirect questions about others' views, and avoid implying "right" answers to address this participant bias.

Cultural bias: Misinterpreting responses due to different cultural backgrounds.

Gender bias: Letting the moderator’s gender impact participant openness, especially around sensitive subjects. Some may share more freely with a moderator of the same gender.

Leading question bias: Framing questions to nudge participants toward a desired response. For example, “Don't you agree ads with real people feel more authentic?” Moderators should carefully design the order of questions and topics to avoid skewing responses.  

Stereotyping: Unintentionally pigeonholing participants based on demographics. This colours analysis and interpretations. Moderators can recruit representative sample populations with diverse backgrounds to minimize this.

Ethical bias: Unconsciously skewing conversations to fit the moderator’s internal sense of right and wrong.

Anchoring bias: Locking into initial participant responses, limiting consideration of other viewpoints that emerge.

Reporting bias

The other type is bias during the reporting phase.

Selectively including quotes and feedback to present only a positive perspective. This can bury important critiques. Analysts should analyze all data fully, even insignificant results, to avoid this and reporting bias selectively highlighting positives.

Confirmation bias: Emphasizing findings supporting pre-existing assumptions.  

Overgeneralization bias: Making sweeping judgments based on a small sample size. For example, assuming one Midwestern mom’s opinion represents nationwide views.

Cultural bias: Failing to accurately convey cultural nuances different from the analyst’s own background.

Language bias: Misinterpreting responses when not fluent in participants' communication style.

Recency bias: Focusing on recent responses while overlooking earlier feedback.

Client expectation bias: Tailoring reporting to fulfill perceived client interests rather than purely representing findings.  

Stepping Back from Bias

The last thing a researcher wants to do is run the risk of any of the confirmation biases mentioned.

In our process on how to avoid bias in qualitative researchers, we take the right steps to avoid this minefield of issues.  So, how can moderators and analysts safeguard the truth?

The crux of answering how to avoid bias in qualitative research comes back to mindfulness, confidence, and constant calibration.

The first step is self-awareness. Isabelle believes “knowing a problem or identifying a problem is fifty percent of the solution.”

By understanding different biases, moderators can intentionally design discussions to limit influence.

“First off, get out of your head and into your body...be present in the moment and mindful,” Isabelle said. Checking biases means focusing outward, not inward.

This starts with showing up authentically. Owning confidence and directing attention outward allows moderators to pick up on subtle signals from the group. By constantly calibrating the room’s energy, moderators can detect shifts and dynamics that provide clues into perspectives.  

Remain present and avoid distraction from inner self-doubt. Mindfulness about biases paired with composure paves the way for fruitful discussions.

Of course, confidence doesn’t come naturally for everyone. Some moderators fuel assurance through preparation and process. Others channel personal superpowers like quick thinking or discipline.  

“Get to know yourself,” Isabelle urges. When moderators clearly recognize individual and collective strengths, playing to skill sets is easier, balancing out limitations.

Avoiding bias in qual is possible  

In the end, unchecked bias, unfortunately, beats objective truth. From the moment of encounter, assumptions stealthily interject, shaping interpretations. And once positions harden, prying free new perspectives grows challenging.  

“We constantly challenge our preconceived notions here at Sylvestre,” said Rami. “The diversity within our team is one of the tools that enable us to confront our biases more effectively. We learn not just to recognize the biases that influence our work, but also how to leverage our diverse viewpoints.”

And that’s a great example of how we constantly evaluate how to avoid bias in qualitative research. It's certainly an ongoing process and learning experience, but with the right mindset, awareness, and attitude, it's possible to avoid or at least minimize bias so the project can drive research results that are clear and have power.

“I don’t think we’re ever fully clear of biases,” said Rami. “But we actively try to navigate qualitative research with more informed and inclusive perspectives. That in my opinion is what makes our insights richer and ultimately more nuanced.”