Introductory Stata 19: Postestimation Statistics Part 1 (predict, kdensity, rvfplot, lvr2plot)
In this video, we will examine the postestimation commands after OLS regressions. After we regress a model, we can use a lot of postestimation commands, including marginal effects of a covariate, linear tests of parameter estimates, tests for heteroskedasticity, diagnostic plots, and many others.
*******************************
19. Postestimation Statistics
*******************************
capture log close
log using postestimation.log, text replace
set showbaselevels on
webuse nlsw88.dta, clear
describe
summarize
regress wage grade age i.married i.union i.race
*normality of error term
predict red1, residuals
predict xb1, xb
summarize wage xb1 red1 if e(sample)
graph twoway (kdensity wage) (kdensity xb1)
kdensity red1, normal
*residuals vs fitted values
graph twoway (scatter red1 xb1, yline(0))
rvfplot, yline(0)
*influential outliers
lvr2plot
*multicollinearity
graph matrix wage grade age, ms(point)
log close
#predict #kdensity #rvfplot