The Problem of Bias

Bias distorts truth. It is wrapped up with our humanity, a bug in the system that we don’t often recognize. Bias can be as straightforward as unconscious preference and it can be simple or sinister. But, however they manifest, the biases we harbor interfere with our ability to understand the real environment around us. Facts just aren’t facts when they’re skewed by bias.

When it comes to market research, bias is the most challenging obstacle. Market researchers strive for total objectivity, but reaching that pinnacle is not an easy task. That doesn’t mean that you shouldn’t conduct research. There are steps you can take to minimize the bias that is present in research by going directly to the sources of insight: market players, decision makers, and customers. 

What’s so Bad About Bias?

What makes bias so bad? We all have opinions, which are a form of inherent bias. However, bias can be relatively manageable when it’s known and understood, but it becomes a major impediment to objectivity when it is not recognized. Awareness is everything. If you are unconscious of your bias, it will creep into your research study. On the other hand, if you understand your subjectivity, you can overcome it, if you take the right steps.

Forms of Bias

Bias in market research can come in various forms but there are four that stand out as primary roadblocks:

The researcher can manage all these biases except proxies. With proxies, you entirely turn your market research over to a third party and step back,  trusting that they will construct the studies to support your unique needs. This can be a disconnect. Without direct access to the analysts and consultants who are doing research on your behalf, it is impossible to uncover and validate/invalidate their assumptions.

As a result, researchers should go directly to the ‘source’. They need to speak directly to customers practitioners and users to obtain data that can enable smart decision making. The ‘source’ has its own biases, but the researcher can ask precise questions to analyze assumptions and/or demographics derived from bias and thus validate/invalidate the source as appropriate data.

You Could Be Your Own Bias

In the same vein, researchers should consider their own biases when developing a research vehicle. Does the respondent know you? If yes, the respondent might answer differently in tone, manner, or even context than if they had talked to a stranger. Therefore, it is usually better to survey or interview ‘blinded’ respondents to ensure that filters are kept to a minimum.

Bias Inherent in the System

Surveys and interview guides can have their own biases, particularly in the method and flow of questions. Researchers need to be careful in crafting these tools, as they can influence the respondent in ways that may distort the ‘truth.’

For example, a question like “How bad are market conditions?” suggests that marketing conditions are poor, forcing the respondent to answer from that perspective, resulting in an ‘echo chamber’ that needlessly reinforces bias. Researchers should instead ask, “What are current market conditions?” which allows the respondent to answer from their point of view.

Lies, Damned Lies, and Statistics

Lastly, bias can come during the “analysis phase,” when the researcher pulls out the most significant information, synthesizes them into ‘meaningful’ insights, and presents these insights to an audience. Communication of insights can lead to misinterpretation. The best way to neutralize these biases is to articulate underlying assumptions along the way and to ensure that the audience understands how these assumptions inform the insights.

The key to good research is to go to the ‘source’, avoiding uncontrollable bias while ensuring unfiltered input from those who have the most at stake – the users.