car (version 3.0-10)

# Tapply: Apply a Function to a Variable Within Factor Levels

## Description

Applies a function, typically to compute a single statistic, like a mean, median, or standard deviation, within levels of a factor or within combinations of levels of two or more factors to produce a table of statistics. This function provides a formula interface to the standard R `tapply` function.

## Usage

`Tapply(formula, fun, data, na.action = na.pass, ..., targs = list())`

## Arguments

formula

a two-sided formula of the form `variable ~ factor.1 + factor.2 + ... + factor.n` or, equivalently, `variable ~ factor.1*factor.2* ... *factor.n`. The variables on the right-hand side of the formula are normally factors or are otherwise coerced to factors.

fun

a function, like `mean`, `median`, or `sd`, that takes a vector first argument and typically returns a single number as its value.

data

an optional data frame within which to find the variable and factor(s).

na.action

a function to handle missing values, as in statistical modeling functions like `lm`; the default is `na.pass`.

arguments to pass to the function given in the `fun` argument, such as (commonly) `na.rm=TRUE`.

targs

a list of optional arguments to pass to `tapply`.

## Value

The object returned by `tapply`, typically simply printed.

## Details

The function given by `fun` is applied to the values of the left-hand-side variable in `formula` within (combination of) levels of the factor(s) given in the right-hand side of `formula`, producing a table of statistics.

## References

Fox, J. and Weisberg, S. (2019) An R Companion to Applied Regression, Third Edition. Sage.

## See Also

`tapply`.

## Examples

```# NOT RUN {
Tapply(conformity ~ partner.status + fcategory, mean, data=Moore)
Tapply(conformity ~ partner.status + fcategory, mean, data=Moore,
trim=0.2)

Moore[1, 2] <- NA
Tapply(conformity ~ partner.status + fcategory, mean, data=Moore)
Tapply(conformity ~ partner.status + fcategory, mean, data=Moore,
na.rm=TRUE)
Tapply(conformity ~ partner.status + fcategory, mean, data=Moore,
na.action=na.omit)  # equivalent
remove("Moore")
# }
```