OLS Regression - Bootstrap
Overview
One of the typical assumptions used when performing inference on an OLS regression is that the error term
is
normally distributeed.
Once this assumption has been made, it is possible to derive analytic expressions for
the distributions of the coefficients in the regression.
There are often good reasons to assume that the error term is normal. However, if the error term is not normal, the
confidence intervals created using the standard inference procedures can be misleading.
Resampling
An alternative to making the normality assumptions, one can use
resampling
to assist in statistical inference.
