Telling a complete story with your data
I’m a structural and analytical thinker so often rely on data for decisions. It wasn’t until I worked with AI models for private health fund data that my eyes were opened to the importance of the human element. Data is important but without human-context, it can lack meaning or lead to incorrect assumptions.
Suppose you’re analysing data of immunisation rates in geographical locations across a State. You have access to demographic data including ethnicity, education, religion and income. The data suggests a correlation between ethnicity and immunisation rates. You may then assume that either; (a) a particular ethnicity is not open to immunisations; or (b) they are not aware of the benefits of immunisations. You might then start a campaign to educate the community about the benefits of immunisation. The campaign only makes a slight improvement. Instead, consider introducing the human element. Someone spends time in the community speaking with individuals and families regarding issues/concerns with immunisations. Throughout this process you hear about the challenges of leaving work during business hours for doctor appointments; some practices are closed weekends; and other medical centres charge a premium for weekend appointments.
The data assumes the correlation of ethnicity and immunisations is the cause of lower immunisation rates. When in fact, the cause is a lack of medical centres providing economical services outside of business hours. Correlation is not causation. As Gillian Tett (Anthropologist) states “Big Data can explain what is happening. It cannot usually explain why.” (Tett, 2021).
Understanding why requires deeper research. Data generally will not explain why your top performer outperforms the rest of your team, however taking an ethnographic approach may provide insight. Ethnography is the study of people through observation. An ethnographic approach involves spending time with your top performer and understanding how he/she approaches his/her day. For example, you might notice email notifications and phones are turned off for 2 hours each morning for periods of productivity; or you might find he/she groups similar tasks together; you may even find he/she takes regular breaks allowing time for more intense focus periods… This causation is rarely found in data.
What does this mean?
Data is incredibly important but not the entire story. Combining data analysis with ethnography provides a more complete story.
Consider the following to assist in building a more complete story:
Know your audience and what’s important to them;
Identify correlations in your data (but remember, correlation is not causation);
Make assumptions about causation;
Investigate your assumptions with an ethnographic approach. Remember, “the least questioned assumptions are often the most questionable” (Paul Broca);
Identify potential actions, not just insight.
An approach with this dual lens of analyst and ethnographer helps you tell a more complete story and uncover potential opportunities.
Notes / References:
Note, example provided is fictional.
Tett, G, 2021, Anthro Vision – How Anthropology can explain business and life, Penguin Random House, United Kingdom.