Function map • gtstats

Choose by task

Question	Start with	Use next when needed
What is in my dataset?	`describe_data()`	`assess_distribution()` and `assess_variance()` for selected continuous variables
How do I describe participants by group?	`summary_table()`	`add_p()`, `add_total()`, or `add_proportion()` only when justified
How do I report outbreak or surveillance measures?	`epi_table()`	Review its denominator and effect-measure components
How do I format final results calculated elsewhere?	`as_stats_table()`	`customise_table()` and `save_output()`
Do groups differ for one variable?	`compare_groups()`	`effect_size()` for magnitude; audit helpers for review
Are two continuous variables associated?	`correlation()`	`plot_correlation()`
What is a proportion or rate?	`proportion_stats()` or `rate_stats()`	—
What are RR, OR, and risk difference?	`crosstabs()`	—
How do I change the finished table?	`to_gt()`	`customise_table()`, `to_flextable()`, `save_output()`

The core distinction

summary_table() describes many variables for a descriptive table. compare_groups() tests one variable across one group variable. assess_distribution() describes the empirical shape of selected continuous variables, not an inferential test choice.

assess_variance() describes group SDs, variances, and spread ratios for selected continuous variables. It is diagnostic only: Welch methods do not require equal variances, and the function does not select an inferential test.

Objects remain inspectable

Every main function returns an object containing the data behind the display. The default print method is the publication-ready table. Components such as $summary, $inferential, $recommendations, and $notes are available for programming and review; they are not extra steps required for routine use.

result <- compare_groups(mtcars, variable = mpg, group = am)
result$inferential
diagnostics_stats(result)

A minimal decision tree

Participant-level data
  └─ describe_data()
      ├─ assess_distribution()  → how to present selected continuous variables
      ├─ assess_variance()      → how observed spread differs across groups
      ├─ summary_table()        → Table 1 / participant characteristics
      └─ compare_groups()       → one inferential question

Outbreak or surveillance data
  └─ epi_table()                → events, denominators, risks, or rates

Already calculated results
  └─ as_stats_table()           → preserve values; style and export only

The Start here article explains the minimal workflow; the birth-weight case study applies it end to end. Missing data and denominators explains how every analysis population and displayed percentage is formed.

For a compact, user-facing table of arguments and defaults for every exported function, see Function options.

For the complete and auditable test = "auto" decision table—including Welch versus Student/ANOVA, marked-skew rank routes, sparse categorical tables, and paired/repeated methods—see Inferential tests and assumptions.