Blog

Your Cultural Biases Can Impact Research Design

September 22, 2020 Author: Mario X. Carrasco

It’s common for many of us to feel as if we have no biases. To make that assumption, however, would be woefully incorrect and naive. 

The hard truth is that we are human, and our cultural biases make their way into every aspect of our lives, including our work. 

If you’re doing market research, that’s a bitter pill to swallow. We like to think of ourselves as some of the least biased people in business. We are more aware of consumer behaviors and the psychological implications of marketing messages. Nonetheless, unconscious bias often creeps into our research design. 

It’s not all doom and gloom, though. Numerous studies pinpoint the most pervasive cultural biases present when conducting market research. Being aware will help you filter them out and deliver a more unbiased research report.

Here are the top biases to be on the lookout for:

Questionnaire design bias. Various cultural biases find their way into questionnaires, ranging from the effect of acculturation to meanings of words in assessments. 

To illustrate, let’s look at one of the most infamous examples of bias found in an SAT analogy question. Below, we see how the meaning of words can differ by culture:

RUNNER: MARATHON
(A) envoy: embassy
(B) martyr: massacre
(C) oarsman: regatta
(D) horse: stable

The correct answer is C. But as the 1994 book “The Bell Curve"  suggests, the answer requires knowledge of words and concepts that are more likely to be present in the lives of affluent white students than in the lives of less affluent minority students.  

You can easily see how this type of word choice could make its way into your questionnaires. When it does, it runs the risk of alienating a portion of your sample base because they don’t understand the terminology. 

So, be mindful of the language you use. Let team members from diverse backgrounds review questionnaires, especially those including analogies and metaphors. Doing so may limit the amount of culturally specific references you may unconsciously make.

Sampling bias. The fast-growing multicultural population in the U.S. has made it increasingly difficult for those without multicultural expertise to create nationally representative sample plans. 

The U.S. Hispanic population represents over 20 countries of origin. The Asian American population has over five languages with 1 million speakers or more. These two groups create complexity in sample plans that are often overlooked, which can lead to insights that are skewed and unactionable.

Even worse, companies making decisions on these insights can cause irreparable damage to their brands among key consumers. 

Reporting bias. The well-known and overused statement “But that’s the way we’ve always done it” applies here. This is particularly apt when it comes to market research reporting. Many custom research companies rely on reporting templates to enable them to grow in an industry where it’s difficult to scale. However, with that scalability comes shortcuts. And when it comes to reporting on multicultural data, shortcuts are dangerous. 

The three biases mentioned are by no means comprehensive. Reducing bias requires constant questioning of the decisions you are making in your research design. In addition to checking your assumptions, surrounding yourself with a diverse market research team will help further eliminate the biases that make it into your research and ultimately deliver more representative and actionable insights.

This blog post was originally published on MediaPost.