March 30, 2004
Student's t Distribution
Student's t distribution accounts for the fact that most confidence intervals and hypothesis tests are based on estimated variances, which are an additional source of uncertainty.
We considered some entries from the table. If you observe the rules, this distribution is not of major practical importance in a lot of applied economic research. The most important point is probably that statisticians refer to "t ratios" even when they are not really using this distribution.
Using the t distribution makes confidence intervals a little wider and makes rejection regions a little further from the null hypothesis.
The Linear Regression Model
We will spend more time with the linear regression model later, but for now we discussed enough material to make it possible to try out this new idea on Stata 4.
