Distribution & Group Comparison
The most common figure type in experimental research. Choosing the wrong display for your sample size — bar charts for n=5, violin plots for n=4 — introduces systematic misrepresentation that reviewers and editors now actively flag. Licklider selects and guards the appropriate figure type based on your declared sample size and data structure.
How to Choose
Rule of thumb: n < 10 → show every point (strip/jitter or dot plot). n = 10–30 → box plot or violin + points overlaid. n > 30 → violin plot or histogram. Bar charts (mean ± error) should be used only when the summary statistic is the primary message, with individual data overlaid for n < 20.
Figure Types
Box Plot
Median, IQR, whiskers (1.5×IQR), outlier points — efficient distribution summary for n ≥ 10.
Violin Plot
Full kernel density estimate with optional box overlay — preferred for n ≥ 20.
Strip / Jitter Plot
Every data point shown — the default for small n (< 20) in biology and medicine.
Bar Chart (Mean ± SEM)
Conditional use with mandatory individual data overlay for n < 20.
Histogram
Distribution shape for a single continuous variable — bin-width optimization required.
Dot Plot (Estimation)
Gardner-Altman estimation plot: raw data paired with mean difference and 95% CI panel.
2D Density Contour
Two continuous variables with overlapping group distributions for large n (> 100).