As data analysts, we pride ourselves on being unbiased or objective, and when we are designing research, we diligently seek to reduce sources of bias. Constructing research this way is considered a best practice of our trade. But is that the only bias that exists? Are we also considering the unavoidable social and cultural forces that subconsciously shape human behavior and perceptions?
These other sources of bias are called systemic or systematic bias, and they reflect the inner workings or relationships within the cultural systems in which we live, work and socialize. They can occur naturally and unintentionally, such as with family rituals or customs, or they can be deliberate attempts to influence an outcome, such as starting food kitchens in food deserts. These sources of bias are simply like the air we breathe; pervasive and so fundamental to our identity that we don't even realize their existence. And if we ask a person in this system a question about their behavior or motivation, their answer would not reflect these forces because they would not be conscious of their effect.
These underlying forces are why we analysts must think about bias in two ways: that which we can avoid, and that which we can leverage. Avoiding bias is a technical design element of any study design with plenty of codified knowledge, so there is no value I can add in that regard. But leveraging unavoidable bias is an opportunity to enhance our understanding, and presents itself in two ways:
- The first opportunity is to consider an analysis of ourselves as important as the analysis of a dataset. Understanding what we've come to believe is "normal" helps in revealing our biases and assumptions when reaching conclusions.
- The second is incorporating an understanding of the cultural systems that influence participants in the study, which provides essential context for any data collected and analyzed.
Combining this understanding of personal bias with an assessment of the impact of cultural systems offers the potential for more powerful and meaningful insights from data. Insights and implications become more than reported numbers absent of any context; they begin to incorporate the fundamental human and social truths that explain why people do what they do. Then, analysis becomes significantly more useful and not just an illusion of precision based on decimal points in a spreadsheet.
Next: Embracing the Bias Part 2: Analyst, Know Thyself