## Request: Pairwise contrast in estimated marginal means

Everything related to the development of modules in jamovi
ricclo
Posts: 5
Joined: Thu Mar 24, 2022 2:49 pm

### Request: Pairwise contrast in estimated marginal means

Thank you for a great software!

I would like to request that pairwise contrasts and their inference is incorporated into the modules. In the current form it is not possible to compare groups.

All the best
jonathon
Posts: 2716
Joined: Fri Jan 27, 2017 10:04 am

### Re: Request: Pairwise contrast in estimated marginal means

do you mean post-hoc tests?

jonathon
ricclo
Posts: 5
Joined: Thu Mar 24, 2022 2:49 pm

### Re: Request: Pairwise contrast in estimated marginal means

Yes, please. I am trying to replace the need for SPSS with Jamovi and one of the functions that I tend to miss is the posthoc pairwise comparison in the estimated marginal means (i.e., corresponding to the emmeans in R).

It would be useful with a transformation ability to the estimated marginal means contrast in generalized linear models (i.e., for odds-ratios to risk ratios in logistic regression, similar to contrast(regrid(emmeans())) ).

All the best,
Rico
jonathon
Posts: 2716
Joined: Fri Jan 27, 2017 10:04 am

### Re: Request: Pairwise contrast in estimated marginal means

hi,

i think you'll find jamovi already provides post-hoc tests and estimated marginal means for linear models.

jonathon
ricclo
Posts: 5
Joined: Thu Mar 24, 2022 2:49 pm

### Re: Request: Pairwise contrast in estimated marginal means

Hello,

Here comes a late addition to this request. Is it possible to incorporate posthoc marginal means to the binomial logistic regression models?

Currently, one can get the estimated marginal risk for a variable (e.g the probability of outcome=1 for man and woman separately), but as far as I can see, point estimate and 95%CI is not provided for their relationship, i.e the marginal risk ratio/marginal risk difference of the two.

In R: marginalmeans::avg_comparison can be used to compute such output.

Thank you for a great program.
best
mcfanda@gmail.com
Posts: 504
Joined: Thu Mar 23, 2017 9:24 pm

### Re: Request: Pairwise contrast in estimated marginal means

Module GAMLj3 does them
ricclo
Posts: 5
Joined: Thu Mar 24, 2022 2:49 pm

Great Thank you