Confidence interval for t-test

Discuss the jamovi platform, possible improvements, etc.
Post Reply
arcaldwell49
Posts: 9
Joined: Wed Oct 11, 2017 7:03 pm

Confidence interval for t-test

Post by arcaldwell49 »

Hey guys, quick question. I noticed the confidence interval for the t-tests is for the confidence interval around the mean difference not the effect size. I think in the current format this may confuse users (myself included in that group) considering the order is (mean difference, SEM, effect size, confidence interval). Would you all be open to moving that confidence interval to next to the mean difference. Also, I have included the code for the confidence intervals for effect sizes if you all would like to add them.

Sorry for the all the comments recently :grin:

For the independent samples

Code: Select all

            LL <- d - (stats::qt(((
              100 - (ciWidth)
            ) / 100) / 2,
            res$parameter)) * abs(d) / stats::qt(res$p.value /
                                                      2,
                                                    res$parameter)
            
            UL <- d + (stats::qt(((
              100 - (100 * ciWidth)
            ) / 100) / 2,
            res$parameter)) * abs(d) / stats::qt(res$p.value /
                                                      2,
                                                    rest$parameter)
For the paired samples

Code: Select all

            LL <- d - (stats::qt(((
              100 - (ciWidth)
            ) / 100) / 2,
            stud$parameter)) * abs(d) / stats::qt(stud$p.value /
                                                      2,
                                                    stud$parameter)
            
            UL <- d + (stats::qt(((
              100 - (100 * ciWidth)
            ) / 100) / 2,
            stud$parameter)) * abs(d) / stats::qt(stud$p.value /
                                                      2,
                                                    stud$parameter)

User avatar
jonathon
Posts: 2613
Joined: Fri Jan 27, 2017 10:04 am

Re: Confidence interval for t-test

Post by jonathon »

ah yeah, that's good idea.

i've created an issue:

https://github.com/jamovi/jamovi/issues/463

jonathon
Post Reply