Coefeasy
is a simple R package under development for making regression
coefficients more accessible. With this tool, you can report linear and
logistic regression coefficients accurately and instantly. The output of
the functions here is only readings in natural language and helps to
standardize the phrasing of widely used models. The package offers
results not only in English, but in multiple languages, and can
automatically control for heteroscedasticity, and offers the capability to
choose the type of standard error for the correction and group errors by
cluster—all in a single function.
Coefeasy
is a straightforward package designed primarily for, but not
limited to, people learning or teaching introduction to statistics or
econometrics. Historically, although an interpretation of coefficients
should follow a consistent design, the underlying knowledge to read the
results of a regression can be quite technical and unambiguous in its
presentation. Acknowledging this challenge, coefeasy
offers a
systematic solution, primarily serving linear and logistic models. As
evidence-based decision making continues to gain importance, coefeasy
aims to help close that knowledge gap, providing clear and accessible
statistical learning, and reinforcing the importance of automation and
standardization.
Coeffeasy can be installed using the development version of the package from github with:
install.packages("remotes")
remotes::install_github("mateoservent/coefeasy")
In this example, we use coefeasy_lm()
to interpret coefficients from a
linear regression on the mtcars dataset, analyzing how car weight
affects miles per gallon when controlling for horsepower and quarter
mile time.
library(coefeasy)
# Fit the linear regression model
model <- lm(mpg ~ wt + hp + qsec, data = mtcars)
# Use Coeffeasy's function to interpret the coefficients
# Using the default settings, which automatically deduce variable names and use an alpha of 0.05
coefeasy_lm(model)
# Specifying the names of the response and predictor variables for a clearer interpretation, adjusting the significance level, and selecting a longer, more descriptive result
coefeasy_lm(model, y = "Miles per gallon", x = "Car weight (1000 lbs)", alpha = 0.01, short = F)
# To display results in another language, such as French, use the argument 'language="fr"' (beta).
Second example, we use coefeasy_logit()
to interpret coefficients from
a logistic regression on the same dataset. We will analyze how the
weight of the car affects the likelihood of a car having an automatic
transmission (1 = Automatic; 0 = Manual).
# Fit the logistic regression model
model_logit <- glm(am ~ wt, data = mtcars, family = "binomial")
# Using the default settings, which automatically deduce variable names and use an alpha of 0.05
coefeasy_logit(model_logit)
# Specifying the names of the response and predictor variables for a clearer interpretation, and adjusting the significance level
coefeasy_logit(model_logit, y = "Automatic Transmission", x = "Car weight (1000 lbs)", alpha = 0.01)