lm_test {tidyrstats} | R Documentation |
Linear Model Testing for Grouped, Nested, or Ungrouped Data
Description
Applies a linear model to a data frame and returns tidy model summaries. Supports ungrouped, grouped (dplyr::group_by()), and nested (tidyr::nest_by()) input data.
Usage
lm_test(input_data, formula)
Arguments
input_data |
A data frame or tibble. Can be ungrouped, grouped, or nested. |
formula |
A model formula, either quoted or unquoted (e.g., y ~ x * z , or "y ~ x * z"). |
Details
Designed to allow seamless 'in-line' chaining to fit linear models to columns of a tibble. Compatible with ungrouped, grouped or nested input. Compatible with native and magrittr pipe. Uses broom::tidy() to extract model summaries.
Value
A tibble with tidy model output sorted by p value, including:
- term
Model term (e.g., intercept, predictors, interactions)
- estimate
Estimated coefficient / beta
- std.error
Standard error of the estimate
- statistic
t-statistic
- p.value
p-value for the hypothesis test
If the input is grouped or nested, group identifiers are retained in the output. In the nested case, nested terms are relocated to the left-most column of the tibble.
Examples
library(ggplot2)
library(dplyr)
# Ungrouped
mpg |> lm_test( cty ~ hwy * cyl)
# Grouped
mpg |> group_by(class) |> lm_test(cty ~ hwy * cyl)
# Nested
mpg |> nest_by(class) |> lm_test(cty ~ hwy * cyl)