4
$\begingroup$

Construct a $95\%$ confidence interval for $5\beta_4$.

If this question were about $\beta_4$ without the $5$, I would absolutely know what to do. But I have to idea how the $5$ comes into play. I can't find any examples in my lecture slides or the textbook that have an numerical coefficient before the regression coefficient.

Would I construct the $95\%$ CI for $\beta_4$ and then multiply the upper and lower bounds by $5$? Or is this some clever way of asking for the $95\%$ prediction interval when the $x$ associated with $\beta_4$ is equal to $5$?

Thanks in advance for any clarification you can give!

$\endgroup$
  • $\begingroup$ Might it mean "parameter B4 from Equation 5"? $\endgroup$ – James Phillips Oct 27 '18 at 15:56
5
$\begingroup$

Confidence interval of $\beta_4$ is constructed using the the following sampling distribution. (My formula here is generic because I don't know how many coefficients you have).

$$\hat{\beta}_4 \sim N( \beta_4, SE^2)\,,$$ where $SE$ is the standard error for $\hat{\beta}_4$. Now let's multiple by 5. $$5\hat{\beta}_4 \sim N( 5\beta_4, (5\times SE)^2)\,.$$

So the standard error gets multiplied by 5. Now you can make the 95% confidence interval using this distribution.

$$5\hat{\beta}_4 \pm t_{1-\alpha, df} (5 \times SE) = 5\left( \hat{\beta}_4 \pm t_{1-\alpha, df} SE \right)\,.$$

Thus, the upper and lower get multiplied by 5.

$\endgroup$

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.