95% ci

[jnine]

Hi SDN'ers

I'm modifying my thesis for publication and wondering if I should/need to report the 95% confidence interval instead of means/SD/and hypothesis tests? I have a relatively simple dataset calculating mean test score as a function of age and gender.

I went through a style guide (not the new/official one) and there's no mention of the CI. My perception is that the current style is to report 95% CI's, but I dont think that it's any easier to understand my data when they're listed.

I thought there may have been a previous thread on this but couldnt find it.

Thanks and any help is much appreciated.

cheers,
j9

---

Members don't see this ad.

 

[Therapist4Chnge]

This came up the other day (weeks/months?) ago. I like having more data available, but I think it may be personal preference.

 

[JockNerd]

Report 95% CIs for things like line graphs to show moderation, or other places where they're typical (logistic regressions, indirect effects in SEM/path analysis). If you're just reporting descriptive statistics, then M and SD and associated tests are typical. It sounds like you're doing a t-test, in which case I would not report CIs.

 

[robinsena]

The current style is to report 95% CI but to my knowledge, that's for tests that are more complicated that a t-test (e.g. regression analyses). Actually, that's more than a style now and nearing a requirement for publication in some journals. But, in your case, you may not need to since you are not doing analyses of that type.

I would definitely check a peer reviewed journal and look for a paper that may have used the same tests you did, to see what stats they chose to report.

Hi SDN'ers

I'm modifying my thesis for publication and wondering if I should/need to report the 95% confidence interval instead of means/SD/and hypothesis tests? I have a relatively simple dataset calculating mean test score as a function of age and gender.

I went through a style guide (not the new/official one) and there's no mention of the CI. My perception is that the current style is to report 95% CI's, but I dont think that it's any easier to understand my data when they're listed.

I thought there may have been a previous thread on this but couldnt find it.

Thanks and any help is much appreciated.

cheers,
j9

 

[jnine]

thank you!