# Histogram — Situational Evidence

Use to show the distribution of one quantitative variable. Use a bar chart—not a histogram—for discrete categories.

## Required

- Use equal-width contiguous bins unless unequal bins are analytically necessary; if widths differ, encode density by area and explain it.
- Label the quantitative x-axis and the count, frequency, probability, or density y-axis accurately.
- Choose bin boundaries from the data and analytical purpose, then keep them consistent across compared panels.
- Use one contrast-safe stop from the selected colourmap for one distribution.

## Allowed choices

- Use established bin rules as a starting point, then check whether important distribution features remain visible and stable.
- Overlay a second distribution with transparency only when overlap is the question and both remain separable by another cue.
- Prefer facets or outline histograms when more than two distributions must be compared.
- Add a short labeled median, mean, or threshold line when it directly answers the question.

## Forbidden

- Do not choose bins solely to make a desired pattern appear.
- Do not leave gaps between ordinary adjacent bins; gaps imply categories.
- Do not use a smoothed density curve as if it were raw data.
- Do not compare panels with different bin boundaries or scales without an explicit reason and label.