Today’s rabbit hole is data equity. A word that encapsulates a lot of what I feel like I’ve wanted to learn about for a long time now. This is really only the start so will essentially be a summary of a cool article I found as well as a log of further reading on the topic. The article is amazing and should be read. There’s even a video version!
Unfair Comparisons: How Visualizing Social Inequality Can Make It Worse TLDR:
- Deficit framing is a bias that people who suffer from inequalities are somehow to blame for them. Deficit thinking is harmful because it encourages things like victim blaming or believing that it’s a characteristic of the group itself that’s causing the differences. This overlaps with ideas around stereotypes.
- People don’t tend to think about variance (different from uncertainty) when looking at visualizations like bar charts and that’s not helped by the chart only showing a single number for each group.
- The design of charts matters, charts that better visualize variability tend to work much better for this so that its harder or impossible to ignore.
- Confidence intervals are specifically not included in the “visualizing variability” group since it shows the uncertainty instead of the actual variability. Confidence intervals can even help to reinforce stereotypes
- Using a single number (eg mean) to represent a group is maybe a bad idea.
Connections:
- At a different level is how things like algorithms can unintentionally reinforce things like stereotypes or systemic racism so this isn’t just happening at a data vis level (link to review, link to book)
Further Reading: