jamovi

Hi Mauricio,

I had decided against log-linear model because I wanted to compare between two groups where one of the variables was held constant. For example regarding the data I attached, compare accident in the four different regions across the different timepoints. So I can make the conclusion that hypothetically "From 2003-2012 to 2016, the accident rates in the South rose significantly more than the accident rates in the North." However, I don't believe my current method allows such a statement.

I would definitely appreciate your opinion on what would be a more suitable method of analyses, as I am a novice when it comes to this!

Thanks again

Hi @nas2782,

my suggestion is always to look at your 3 categorical variables with log-linear regression and link you to my answer in this forum, for a request similar to yours.
viewtopic.php?f=5&t=2018&hilit=Log+linear#p7369

Cherrs,
Maurizio

Hi Maurizio,

Is there a way to perform these analyses in jamovi without having the data already formatted with counts? Perhaps using Rj to create a count table and then performing the subsequent poisson?

Thanks!

Is there a way to perform these analyses in jamovi without having the data already formatted with counts?

Hi @nas2782,

of course, counts is optional and you use it only if you have a summary data and do not have the variables with the collected data.
Put your three variables in the Factors field (leave the Counts field blank) and concentrate on building the models of interest, etc.

Note:
If you use the cloud version and create many models, you may notice some small slowdowns because the calculations are done directly by Jonathon with its new 12th generation abacus.

Hi Mauricio,

Thanks again, that was extremely helpful. One thing I'm having trouble with is comparing two groups where one variable is held the same. For example in the sample dataset, can I compare (assault-assault) x (2016-1997) x (midwest-east) using log-linear in jamovi? I tried doing this using a dataset where I made the counts and used GAMLj (attached), but post-hoc test results would not appear.
Thanks again for your help.

Hey @nas2782,

continuing on this old topic we risk being OT.
If you want we can continue from the forum, but as a private message (click on my Balloon).

Cheers,
Maurizio