In OLS, given the regression equation y = B0 + B1X, why do I often read that B1 represents the average effect of a predictor? I don't get that.
For example,
data <- data.frame(sex=c("male","female","male","female","male","female","male","female"), DV=c(22,32,34,16,66,34,77,23))
The average effect should be mean(data$DV) = 38
However,
m <- lm(DV~sex, data=data)
m$coefficients
(Intercept) sex
26.25 23.50
B1 is 23.50, which is not the average of the gender predictor...
For example,
data <- data.frame(sex=c("male","female","male","female","male","female","male","female"), DV=c(22,32,34,16,66,34,77,23))
The average effect should be mean(data$DV) = 38
However,
m <- lm(DV~sex, data=data)
m$coefficients
(Intercept) sex
26.25 23.50
B1 is 23.50, which is not the average of the gender predictor...