t-critical value in confidence interval

[elashoffb]

For the alternative hypothesis ">", the t-critical value in the formula for the confidence interval is computed using the area= alpha/2, even though it is a one-sided test. Is there a reason you did that?

[AntoineSoetewey]

Dear Barbara,

Thanks for your question.

In this app, only bilateral (two-sided) confidence intervals are computed. For hypothesis tests, bilateral and unilateral tests are computed.

So if you want to compute a unilateral confidence interval, you should indeed use alpha instead of alpha/2.

The reason for that is that the code of the app is already quite complex, so I decided to not do unilateral (one-sided) confidence intervals (and also because they are less common). But if you find the time to improve the app, feel free to submit a pull request.

I have added a note in the app to make this clearer, see: 36da729.

Regards,
Antoine

[elashoffb]

Dear Antoine,
Thank you for that explanation! I figured you must have had a reason.
Best,
Barbara