Quick question, in the textbook "Introductory Econometrics", the variance of a Regression Coefficient is given as:

$var(\hat\beta_j) = \frac{\sigma^2}{SST_j(1-R_j^2)}$

where, $SST_j$= $\sum_{i=1}^n (x_{ij}-\bar x)^2$

$\sigma^2$ is the variance of errors of y against its predicted values

$R_j^2$ is the strength of fit when regressing the jth predictor against the other predictors.

So am I right in assuming that the Standard Error of the Regression Coefficient is the square root of this $var(\hat\beta)$ value?

  • 7
    $\begingroup$ You're correct. "Standard error" is used in two senses, one to refer to $\sqrt{var(\hat{\beta}_j)}$ and another to refer to the estimate that results from replacing $\sigma^2$ with an estimate in $\sqrt{var(\hat{\beta}_j)}$ $\endgroup$
    – user179309
    Jan 8 '18 at 4:34

Yes that is correct. You may be confused because, while we give separate terminology to standard deviation and standard error, the square is called a "variance" in both cases.

  • 2
    $\begingroup$ It might be worth adding that (of course) the standard error is a standard deviation, the standard deviation of the sampling distribution of the statistic. Which may be another reason why this is so confusing. $\endgroup$
    – mdewey
    Jan 8 '18 at 13:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.