Diversity, equity and inclusion are becoming an increasing focus in many facets of life. A simple example of these three key areas can be applied to a classroom and a student’s birthday party. Diversity is everyone in the class is invited to the party. Equity is everyone receives the same party favor and same size of birthday cake. Inclusion is everyone has the opportunity to play or dance.
The business focus on diversity, equity and inclusion (commonly known as DEI) impacts more than just employee satisfaction. Research done by Quantum Workplace discovered that 61 percent of employees believe these strategies are essential and affect the workplace environment. What’s more, cognitive diversity can increase workplace innovation by as much as 20 percent, which affects the bottom line.
DEI is typically applied to areas like the workplace, church, communities, schools, etc. What may surprise you is that DEI can also be applied to data analysis. Companies review data to interpret actions of groups of people and use that interpretation to develop additional strategies. Because there is a human element to how and why data is collected, there are assumptions that can quickly derail DEI efforts through unknown bias. These 3 C’s of DEI data are critical to consider and can be avoided.
The Assumption: Data does not exist in a vacuum. Data must have some context around it to create a story. The number 2 does not have any context and therefore cannot be interpreted. If that number is $2, one additional data point now exists but more information is needed – is it $2 for a cup of coffee or for an acre of land? The issue becomes what context the data is put into, and how that context can make the data seem positive, negative or neutral.
Avoiding Context Assumption: In the above example, if this was an example of a price for land, did you assumed it must be in an undesirable location? What if that price was for an acre of land in 1823? Looking for additional data points helps to fill in the data picture. Avoid this assumption by adding additional data points so context is not filled in based on individual bias.
The Assumption: Data correlation is one way in which data is applied. This becomes a working hypothesis as we develop strategies based on the data and how the data points affect one another. For example, your organization is looking to increase diversity in a particular department and the job is advertised with a recruiter and includes an educational requirement. The number of qualified people of color is less than expected and the assumed correlation is that, because of the educational requirement, the pool of diverse candidates is less.
Avoiding Correlation Assumption: When the recruiter in this example is interviewed, it is found out that the advertisement targets did not include communities of color. A correlation of education to applicants was assumed that was not accurate. Unless testing for a direct correlation, take the time to dig deeper into why the data shows something unexpected. Be suspicious of possible correlation unless it can be statistically verified.
The Assumption: The idea of credence is to have a belief that something is credible or having confidence in something as accurate. Data, if tortured long enough, can give up any story required. Data simply is and has no bias. What causes bias and shades of credibility is the viewpoints of the analyzer. Put another way, people see what they want to see in the data. Because of individual experiences, data points can be elevated or diminished based on those experiences and that bias.
Avoiding Credence Assumption: In the last example, the educational requirement was given higher credence and the community target strategy was given less credence. It is a common and subconscious practice to focus on a subset of data. Before focusing on a subset of data, consider the data as equal and look at trends to help identify areas of interest. Changes in trends reduces credence assumption and highlights important data.
With so much information coming in, leveraging the learnings of past experiences and focusing on seemingly important data points should be expected. Those experiences do not need to be labeled right or wrong, they simply help eliminate data overload.
When companies recognize the assumptions that are automatically applied and become aware of bias that have the potential to change the analysis of data, those companies are removing blinders that could have them miss opportunities or worse, lead them down a path that harms the business.
These 3 C’s are some of the first biases applied to data during the analysis phase. They are the first because they are inherent and as such, become embedded in the data application DNA. Having an awareness of how those 3 C’s also affect DEI can have a ripple effect in an organization.
Lisa Apolinski is an international speaker, digital strategist, author and founder of 3 Dog Write. She works with companies to develop and share their message using digital assets. Her latest book, Persuade With A Digital Content Story, is available on Amazon.