Inconsistent b0 in linear regression between Jamovi and R

Discuss statistics related things

by jonathon » Fri Dec 13, 2019 11:18 am

the 1.1.9 is out!

jonathon
User avatar
jonathon
 
Posts: 1689
Joined: Fri Jan 27, 2017 10:04 am

by uab » Fri Dec 20, 2019 10:01 am

Great! Thanks so much for this.
uab
 
Posts: 3
Joined: Tue Nov 19, 2019 7:22 pm

by bgodde » Thu Mar 04, 2021 10:19 am

I just want to add two thoughts to the discussion:

1. Probably it is particularly good for teaching - an important learning outcome - that different procedures and algorithms come to different results and that it is not about true or false but about different approaches that are all justified.

2. You will find different results (particularly with regards to is an effect significant or not and how strong) if the effect itself is not very robust. Same can happen with doing analyses with or without covariates. Probably it would be a good idea to report in a paper both results (with / without covariates or those obtained with both software packages). At least one needs to be completely transparent here and should define (and justify) which package one will use for analysis before doing it.

The more important question is, how can we be sure a software package does the calculations correctly as it is supposed to do it?

Ben
bgodde
 
Posts: 2
Joined: Thu Mar 04, 2021 10:02 am

Previous

Return to Statistics