Are one-sided confidence intervals taught in intro courses?

[cwolock]

Unlike t.test from stats package, our ttest does a two-sided confidence interval regardless of the null hypothesis. I assume this is intentional, since it involves some extra steps in the code. My recollection is that intro courses (at least BIOST 511) teach only two-sided intervals. I am fine keeping this the way it is - just wanted to check since it means ttest doesn't correspond to t.test in this case. Thoughts?

[adw96]

Hi Ken & Jim! (@kenrice and @jphughes9) Do you need this functionality or would you like it? I don't discuss one-sided intervals in 515/518, though I think I assume they see them in 514/517.

[adw96]

@cwolock -- thanks for starting this conversation! I got responses from Ken and Jim by email.

Ken: "But one-sided tests are the lego blocks from which we can build basically anything! (https://doi.org/10.1111/rssa.12496) ;) Seriously, I don’t mind. I mention one-sided tests but only very briefly, and as non-examinable material. Either way is fine."

Jim: "I’m more of a tinker-toy guy myself …. I didn’t use one-sided confidence intervals in 511"

So sounds good -- let's keep them out 👍 Thanks!

[cwolock]

Thanks for sending along! I just realized there's a typo in my original post - should say "regardless of the alternative hypothesis." That is to say, if you run a one-sided test, you still get a two-sided interval. I gather everyone understood what I meant, but just wanted to clarify. I will leave it as-is.

[jphughes9]

Charlie Indeed, that is how I interpreted the question. Jim -------------- Jim Hughes Professor of Biostatistics University of Washington Email: ***@***.******@***.***> Mailstop 351617 Phone: 206-616-2721 Seattle, WA 98195 Fax: 206-616-2724 Cell: 206-498-2456 From: Charles Wolock ***@***.***> Sent: Tuesday, August 3, 2021 4:11 PM To: statdivlab/rigr ***@***.***> Cc: Jim Hughes ***@***.***>; Mention ***@***.***> Subject: Re: [statdivlab/rigr] Are one-sided confidence intervals taught in intro courses? (#47) Thanks for sending along! I just realized there's a typo in my original post - should say "regardless of the alternative hypothesis." That is to say, if you run a one-sided test, you still get a two-sided interval. I gather everyone understood what I meant, but just wanted to clarify. I will leave it as-is. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub<#47 (reply in thread)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AFPGUCS5FTSBKAGBBP265ADT3BZPBANCNFSM5BPW6IFA>. Triage notifications on the go with GitHub Mobile for iOS<https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android<https://play.google.com/store/apps/details?id=com.github.android&utm_campaign=notification-email>.