center: Center Predictor Variables.

Description

Perform column centering and allow for centering by groups. By centering, I simply mean to subtract a constant from every value of a variable $(x - \bar{x})$, which produces an equivalent as scale(x, scale = FALSE) in the base package.

Usage

center(x, by)

Arguments

A variable name whose values is to be centered.

A factor variable.

Value

Scaled values

Details

Essentially, what it does is redefine the 0 point for that predictor to be whatever value you subtracted. It shifts the scale over, but retains the units. In a linear model, the effect is that the slope between that predictor and the response variable doesn't change at all. But the interpretation of the intercept does. When 0 is out of the range of data, the intercept value is meaningless. But when x is centered, the mean of the response when all predictors = 0, becomes actually 0; thus the intercept becomes the mean of Y at the value you centered on.

Examples

Run this code

require(ggplot2)
#have random normal distributions with means of 2 and 5, respectively.
n <- 20
df = data.frame(id=1:n, x=rnorm(n, mean=2, sd=.5), y=rnorm(n, mean=5, sd=2),
age=rnorm(n, mean = 40, sd = 30), female=sample(c(TRUE, FALSE), n, rep = TRUE) )
mod = lm(y ~ x, data=df)
summary(mod)
ggplot(df, aes(x=x, y=y)) + geom_point()
ggplot(df, aes(x=x, y=y)) + geom_point() +
geom_abline(intercept=mod$coefficients[1], slope=mod$coefficients[2]) +
stat_smooth()

# centering
mod = lm(center(y) ~ center(x), data=df)
summary(mod)
 # center and z-score: (x - xbar)/sd(x)
 # mod = lm(rescale(center(y)) ~ rescale(center(x)), data=df)
 # as.beta(mod) # after fitting

 with(df, center(y, as.factor(female))) # center by group