0
$\begingroup$

Suppose we repeat an experiment identically and independently 100 times. Each time we construct a 99% confidence interval for $\mu$ via the t-distribution. Let X = the number of times the confidence interval fails to contain the true value of $\mu$. The distribution of X is:

Normal with $\mu=0$ and $\sigma^2 = 1$, is this correct? Since in Normal, the mean should be zero and the variance is always one.

Since this is a t-distribution, it should be Normal. I thought another answer might suit it being Normal with $\mu = 99$ and $\sigma^2 = 0.99$

Could someone explain which one could be correct, if correct at all?

$\endgroup$
  • 3
    $\begingroup$ Think about this, what values can $X$ take on? Can $X$ be negative? Can $X$ be a non-integer? $\endgroup$ – Cam.Davidson.Pilon Apr 19 '14 at 1:15
  • 1
    $\begingroup$ @Cam.Davidson.Pilon So re-thinking about this, X can take on only integers that indicate where the CI succeeded or failed. It cannot be negative and must be an discreet variable. Since it's either success or failure, then this is a Binomial with $n=100$ and $p=0.01$, Does this make sense? $\endgroup$ – GivenPie Apr 19 '14 at 1:22
  • 2
    $\begingroup$ So perfect you should write it as an answer yourself! $\endgroup$ – Cam.Davidson.Pilon Apr 19 '14 at 1:27
1
$\begingroup$

So re-thinking about this, X can take on only integers that indicate where the CI succeeded or failed. It cannot be negative and must be an discreet variable because we essential are counting the number of successes which must be a whole integer. Since we are gauging whether success or failure, then this is a Binomial with n=100 and p=0.01. p = 0.01 since we that's the chance to fail in the 99% confidence interval.

$\endgroup$
  • $\begingroup$ You're right, but it for completeness' sake it would help to state explicitly where you used the assumptions of "identically and independently." Can you construct counterexamples to your conclusion when either of those assumptions does not hold? $\endgroup$ – whuber Aug 20 '18 at 11:36

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.