Clean water is necessary for biological life to survive. We need it to live and stay healthy…but what if it’s evil?
- Water can be extracted from rocket fuel.
- Water is the main ingredient in pesticides.
- Water is the #1 cause of drowning.
- 100% of people exposed to water will die.
Sure, all of the above are true, but that doesn’t mean you should avoid water — that would be impossible, anyway! These are instances of presenting data unethically.
So how can you present data ethically? First, consider your message. The facts may not be distorted, but the way the information is presented may be altered to intentionally or unintentionally exaggerate or understate the facts.
Here are a few unethical strategies to avoid when you present data in graphs and charts.
Truncated Axis
In this case, the Y-axis does not start at zero, so the data is exaggerated.
In the charts below, the differences between the interest rates are negligible — they’re 0.002% different from one another. But the chart on the left is much different than the chart on the right because the Y-axis begins at 3.140% on the left.
Takeaway: Start your Y-axis at zero like the chart on the right.
Area as Quantity
With area as quantity, the data is distorted because the area of the bars or pieces within the chart that represent the data don’t match their values.
In the chart below, the people on welfare and the people with a full-time job are not significantly different — only 6.9 million — but the chart, whose X and Y axes are not labeled, makes it seem like a much more significant difference.
Takeaway: Make the area of bars, circles, or whatever you use to represent the data proportional to the values of the data.
Correlation as Causation
Sometimes when we see a chart whose data almost matches, it can be easy to think that one thing caused the other.
In this example, the data might lead you to think that the increase in murders caused more people to buy ice cream, or that murders increased because more people bought ice cream. It’s more likely they aren’t causally related at all. One didn’t cause the other; it’s simply a coincidence that they seem related. We call this “correlation.”
In actuality, the rates of murder and ice cream purchases are dependent on the weather: the hotter it is, the more ice cream is sold and the more murders are committed. That means the two are correlated, not causal.
Takeaway: In similar data, one thing might not have caused another; they could be simply correlated (coincidental).
For more funny examples, visit Spurious Correlations.
Aspect Ratio
When it comes to pie charts, it can be difficult for the brain to interpret results. In this example below, the pieces of the pies look strikingly similar from one pie chart to the other.
Comparison data is better presented in bar charts or line graphs, where your audience can easily see the differences in data.
Takeaway: For comparison analyses, use bar charts or line graphs instead of pie charts.
Now that you know how to ethically present your data, where does IntelliBoard fit in? Our next post will show you how to use IntelliBoard data to effectively tell your story and convince your audience.
Ann McGuire is an experienced marketer with more than 20 years creating content, marketing communications programs, and strategies for tech firms. She reads, writes, and lives in New Haven, CT with her husband and two needy cats.
Resources
Explore Learning Analytics Insights
Assessment for Learning: Creating Strong Feedback Cycles
Read how you can use your LMS data to deliver strong assessments for learning without being difficult to set up.
What is Regular and Substantive Interaction (RSI)?
Discover how regular and substantive interaction can facilitate high-quality learning at your educational institution.
How Universities Can Break Down Data Silos and Generate New Insights
Break Down Data Silos and Generate New Insights