How to use Predict() a linear model given the value of B1

Question

library(wooldridge)
library(tidyverse)
library(stargazer)
setwd("C:/Users/Charlie/Desktop/R Homework")

data(wage1)

reg_wage1 <- lm(lwage ~ female + educ + (female * educ), data = wage1)

stargazer(reg_wage1, type = "text")

female_at_zero <- data.frame(female=0)
pred <- predict(reg_wage1, female_at_zero)

stargazer(pred, type = "text")

My problem is when I try to run this code it keeps asking me to put in a value for educ, but I do not want to change education, I only want to see the model's results in stargazer if female is equal to 0.

Answer 1

The issue that you're running into is that you've defined your model reg_wage1 with educ so it won't know how to make a prediction without an educ value. lm has defined the female variable based on the values of the educ variable (not to mention the interacted variable you added to the model). If you want to see the effect on lwage with just female = 0 , you're going to have to redefine your model.

Another option is to simply call female_at_zero <- data.frame(female = 0, educ = 0) . By setting educ to 0, you're essentially looking at just the effect of female . But this solution is redundant because the value of pred will simply be the value of Constant from the output of stargazer(reg_wage1, type = "text") . This is due to the fact that a linear model ( lm ) defines its intercept ( Constant ) as the point at which all independent variables are equal to 0.