$W_1, \ldots, W_5$ be iid $Normal(3.1, 5.2^2)$. What is the probability that $\left(\frac{\bar W - 3.1}{\hat \sigma / \sqrt{5}} \geq 2.1\right)$?

I am not sure how to start. Is this equivalent to the probability that $Z \leq 2.1$?


You're close, but it's actually $1-P[t \le 2.1]$.

The distribution of $\bar{W} = \frac{1}{5}\sum_i^5 {W_i}$ is

$$ \bar{W} \sim \mathscr{N} \left( 3.1, \frac{5.2^2}{5} \right) . $$

Normally, $ (\bar{W} - 3.1)/\left( \frac{5.2^2}{\sqrt{5}}\right) $ would be the standardization of $\bar{W}$ into a standard normal $Z \sim \mathscr{N}(0,1)$. Because of the use of $\hat{\sigma}$ in the question, this is presumably coming from a sample of $5$ observations of $W$, and we are using a $t$-distribution with $4$ degrees of freedom, not a standard normal.

So, we can rewrite the probability we want in terms of $t$:

$$ P \left[ \frac{\bar{W} - 3.1}{\frac{\sigma}{\sqrt{5}}} \ge 2.1 \right] $$ $$ = 1 - P \left[ \frac{\bar{W} - 3.1}{\frac{\sigma}{\sqrt{5}}} \le 2.1 \right] $$ $$ = 1 - P \left[ t \le 2.1 \right] . $$ $$ \approx 0.052 .$$

  • $\begingroup$ Can you explain how you got 0.052 from the second to last step. (1-P[t <= 2.1]) $\endgroup$ – Jeremy Feb 25 at 2:43
  • $\begingroup$ You can't calculate it by hand, you have to use a numeric solver (available in any statistical software or Excel) or a t-table. Here it is using WolframAlpha. $\endgroup$ – Amaan M Feb 25 at 18:59

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.