Quick answer: R Commander’s Graphs menu covers the plots that matter: Histogram for distribution shape, Boxplot for group comparison and outlier spotting, Bar Graph for categories, Quantile-comparison for normality checks, and Scatterplot for two-variable relationships — every one auto-generating the R code in the script window. Graphs are how you see the descriptives; this guide matches each plot to its job. Menu overview first: what R Commander is.
The right plot for each question
| Question | Graphs menu choice | What to look for |
|---|---|---|
| What is the shape of this variable? | Histogram | Bell vs skew vs bimodal; bins that hide detail |
| Do groups differ in spread and center? | Boxplot (with groups) | Median lines, IQR boxes, points beyond whiskers |
| Are counts balanced across categories? | Bar Graph | Order effects, dominance of one level |
| Is this variable normal enough for t-tests? | Quantile-comparison plot | Points hugging the diagonal line |
| Do these two numbers relate? | Scatterplot | Direction, tightness, curvature, outliers |
Two options pay for themselves immediately. In the Boxplot dialog, tick Identify outliers with mouse — click suspicious points and R Commander labels them with row numbers, turning outlier hunting from guesswork into lookup. In the Histogram dialog, set Number of bins manually (try 10–20): the auto default can smooth a real skew into an innocent bell. Both dialogs write code you can reuse — Hist(), Boxplot() — which is the quiet way R Commander teaches base R graphics, and the summaries behind these plots are in our descriptives guide.
The pre-test plot ritual
Before any inferential test, three plots in ninety seconds: histogram (shape), quantile-comparison (normality), boxplot-by-group (equal spread). A clean trio licenses the t-test or ANOVA; a skewed, wild-spread trio routes you to the non-parametric shelf instead. Why shape matters so much — and what distributions model in the first place — is in the distributions guide, and the tests these plots gate-keep are mapped in the statistical tests catalog. jamovi users get the same checks as tick-boxes, shown in the jamovi first-analysis guide.
For one-to-one coaching that turns plots into defensible analysis decisions, Ampersand Academy teaches R and statistics one-to-one.
Frequently asked questions
How do I make a histogram in R Commander?
Graphs menu, Histogram, choose the variable and OK. Set the number of bins manually for honest shape detection, and the R code appears in the script window for reuse.
How do I compare groups with a boxplot in R Commander?
Graphs, Boxplot, pick the numeric variable and move the grouping factor into the Groups box. Tick identify outliers with mouse to label extreme points with their row numbers.
Which plot checks normality before a t-test?
The quantile-comparison plot: sample quantiles against a theoretical normal line. Points hugging the diagonal support normality; systematic curvature or heavy tails argue against it.
Why does my histogram shape change with bin count?
Binning is a resolution choice – too few bins hide real skew, too many exaggerate noise. Compare a few settings between 10 and 20 bins before concluding anything about shape.
Can I save or export the plots R Commander makes?
Yes – right-click the R graphics device to save as PNG or PDF, or use the device menus. The generated code can also be rerun with file-targeted graphics parameters for reproducible output.