Dummy variables for ordinal logistic regression on Jamovi

jordanwater · Post by **jordanwater** » Mon Apr 22, 2024 9:21 am

Hey guys!

So i'm in a state of confusion as to whether Jamovi automatically takes into account of how most logistic regressions need dummy variables for each value/level/category of their variable. I've seen that a lot of people do that for their logistic regressions on similar studies of mine but where they use a different programme. While i'm watching tutorials on YouTube I do not see anyone making dummyvariables for their variables, and was therefore wondering if Jamovi has this already programmed into the software where it takes account of all that already?

For context, my independent categories are nominal and therefore don't have an average. (Also might as well pop another small question in here to verify, ordinal logistic regressions only require the dependent variable to be ordinal, right?

Hope my short question made sense... and thank you for taking your time to help me out.

Post by **jonathon** » Mon Apr 22, 2024 10:14 pm

hi,

yes, when you specify a variable as a factor, jamovi takes care of the dumming coding for you automagically. if you want to do it manually, you can dummy code the variable yourself, and then specify the dummy coded variables as 'covariate's.

ordinal logistic regressions only require the dependent variable to be ordinal, right?

yup.

jonathon

njain@iiphg.org · Post by **njain@iiphg.org** » Tue Apr 23, 2024 5:43 am

I am writing to discuss a challenge we encountered when we ran the meta-analysis in the JAMOVI cloud.

We wanted to run a proportions meta-analysis to calculate the pooled prevalence. However, after several attempts, we were not able to perform the test.

A message “the condition has length >1” is displayed.

Please let us know if you have any thoughts or recommendations on how we can proceed.

Post by **jonathon** » Tue Apr 23, 2024 5:45 am

hi njain,

can i get you to start a new thread?

kind regards

jonathon

jamovi