MAJOR -Cooks distance rslts different RR vs OR - bug or math

General help and assistance with jamovi. Bug reports can be made at our issues page: https://github.com/jamovi/jamovi/issues . (If you're unsure feel free to discuss it here)
Post Reply
apavlo
Posts: 9
Joined: Thu Aug 18, 2022 2:00 pm

MAJOR -Cooks distance rslts different RR vs OR - bug or math

Post by apavlo »

Hello so I get a significant outlier detected with cooks distance when I do OR but not when I do RR. Is this a bug or is there a specific mathematical reason why that is? What should I report?

Regards,

Achilleas
Last edited by apavlo on Tue Aug 23, 2022 5:30 pm, edited 1 time in total.
User avatar
Ravi
Posts: 194
Joined: Sat Jan 28, 2017 11:18 am

Re: MAJOR -Cooks distance rslts different RR or OR - bug or

Post by Ravi »

Hi Achilleas, I'm not sure what you mean with "OR" and "RR" but if you fit different types of models the cook's distance is gonna be different because it is dependent on the model.

Cheers,
Ravi
apavlo
Posts: 9
Joined: Thu Aug 18, 2022 2:00 pm

Re: MAJOR -Cooks distance rslts different RR or OR - bug or

Post by apavlo »

Hello Ravi. First, thank you for responding!

What I meant was odds ratio (OR) vs relative risk (RR). Not a math expert by any stretch but it seems to me that in essence they are measuring the same thing so the Cooks distance analysis should produce similar results. And if not then I would like to know whether it is mathematically preferential to run cook's distance with OR or RR values.

Regards,
User avatar
Ravi
Posts: 194
Joined: Sat Jan 28, 2017 11:18 am

Re: MAJOR -Cooks distance rslts different RR or OR - bug or

Post by Ravi »

So what specific analysis are you talking about here?
apavlo
Posts: 9
Joined: Thu Aug 18, 2022 2:00 pm

Re: MAJOR -Cooks distance rslts different RR vs OR - bug or

Post by apavlo »

Risk Ratio/Relative Risk (RR) vs Odds Ratio (OR) meta-analysis with random effects (if I recall, it might have been fixed-effects)... If you recall, in MAJOR you get a written output of the results. When I run odds ratio it says that Cooks distance detected 1 outlier study (e.g., Ravi, 2022) when I run OR but when i do RR Cooks distance doesn't detect this study.
User avatar
Ravi
Posts: 194
Joined: Sat Jan 28, 2017 11:18 am

Re: MAJOR -Cooks distance rslts different RR vs OR - bug or

Post by Ravi »

Ah yeah, I didn't realize you were talking about the MAJOR module. It's probably best to contact the developer of the MAJOR module, Kyle Hamilton (https://kylehamilton.com/), because I don't know much about his analysis.
Post Reply