STA 113 2.0 Descriptive Statistics

Misuse of Statistics

Dr. Thiyanga S. Talagala
Department of Statistics, Faculty of Applied Sciences
University of Sri Jayewardenepura, Sri Lanka

Misuse of statistics

Misuse of statistics refers to the incorrect application or interpretation of statistical data, which can lead to misleading conclusions.

Presenting proportions without counts – Reporting percentages without the raw numbers, making it hard to assess sample size or reliability.
Simpson’s paradox – Ignoring subgroup effects that reverse the overall trend when data are aggregated.
Reversing the axis – Flipping the scale direction to create a misleading impression.

Presenting summaries without visualizations – Omitting graphs that could reveal distribution shape, outliers, or patterns.
Bar charts that do not start at zero – Exaggerating differences by truncating the axis baseline.
Ignoring skewness – Using the mean when the distribution is skewed, leading to a poor representation of central tendency.
Over-reliance on a single measure – Reporting only the mean or median without spread (range, variance, standard deviation, IQR).

Too many categories → Human eyes can’t reliably compare many wedge angles. Once you get more than 4–5 slices, it becomes guesswork.
Pie charts in general → They encode data by angle, which our brains are not great at judging, especially without a baseline. A simple bar chart usually communicates the same information much more clearly.
3D pies are worse → The perspective distortion changes wedge sizes, making some slices look bigger or smaller than they really are.