1 Your first table: the basics of tbl_summary()
1.1 Looking at the data
1.2 Table 1 in one line
With just this,
- continuous variables (
age,marker, etc.) → median (Q1, Q3) - categorical variables (
stage,grade, etc.) → n (%) - 0/1 variables (
response,death) → the number and proportion of 1s on one row - missing values → an Unknown row
are detected and summarized automatically. The footnote below the table states the summary statistics used.
1.3 Choosing variables: include
In include, besides listing column names directly, you can use selection helpers such as starts_with("t") and where(is.numeric).
1.4 Changing labels: label
In gtsummary, per-variable settings are written as variable ~ value (a formula). Several can be specified together with list(). You will see this notation again and again.
1.5 Changing statistics: statistic
To show continuous variables as mean (SD):
all_continuous(): all continuous variablesall_categorical(): all categorical variables (including 0/1 variables)- Write statistic names inside
{ }; any other characters are displayed as is
| Available statistics (continuous) | Meaning |
|---|---|
{mean}, {sd} |
Mean, standard deviation |
{median}, {p25}, {p75} |
Median, 25th and 75th percentiles |
{min}, {max} |
Minimum, maximum |
{N_nonmiss}, {N_miss} |
Number non-missing, number missing |
| Available statistics (categorical) | Meaning |
|---|---|
{n}, {N}, {p} |
Count, denominator, percentage (%) |
You can also specify individual variable names, as in statistic = list(age ~ "{mean} ({sd})", marker ~ "{median} ({p25}, {p75})"). This lets you, for example, use mean (SD) for roughly normal variables and median (IQR) for skewed ones.
1.6 Showing on multiple rows: continuous2
To display several statistics for one variable on separate rows, set its type to continuous2.
1.7 Number of decimal places: digits
1.8 Specifying variable types: type
gtsummary detects types automatically from the kinds of values, but when it doesn’t do what you want, fix it with type.
| Type | Display |
|---|---|
"continuous" |
Summary statistics on one row |
"continuous2" |
Summary statistics on several rows |
"categorical" |
n (%) for each level |
"dichotomous" |
Only one level (by default 1, TRUE, or “Yes”) on one row |
1.9 Displaying missing values: missing
With missing = "no", the missing row disappears, but the denominator of the percentages is the number excluding missing values. Papers are often required to report the number missing, so if you hide it, supplement it in a footnote or similar.
1.10 Direction of percentages and sorting
tbl_summary()alone gives you the groundwork for Table 1include(variable selection),label(labels),statistic(statistics),digits(decimal places),type(type),missing(missing values)- Settings are written by combining
variable ~ valuewithall_continuous()/all_categorical()