jamovi

Hi, I was doing assumption checks for my analysis' in my thesis.
I couldn't figure it out this, could you help me:

Is there a way to handle these assumptions in this module (or in the jamovi):
1. Homogenity of variances (Homogenity of level-1 variances ?) (Equality of level-1 variances for each level-2 unit ?) -> For this, can I use the residuals-predicted scatterplot?
1.1. Is "Homogenity of variances" assumption meaning "Homogenity of level-1 variances" or "Equality (Homogenity) of level-1 variances"?

I have used Levene's test in the 'One-Way ANOVA' test and the result is p<.05. But in this, there are less missing values that deleted because I didn't delete missing values from the dataset, meaning GAMLj3 module deleted automatically missing values that came from variables that I included in the module. So variances are not equal. Is this a problem for the analysis or Levene's test is not related to the this assumption?

With residuals-predicted scatterplot, I can say level-1 variances are independent. Can I say level-1 variances are homogen? With residuals-predicted scatterplot by cluster variable, can I say "Equality (Homogenity) of level-1 variances for each level-2 unit are met because in each cluster variances are alike."?

2. Normality of level-2 residuals/errors
There is a save option for the residuals in the Options but there is no save option for the level-2 residuals (Am I wrong?). In this situation, how can I check this assumption?

3. Independence of each random effects -> For this, can I use the random effects correlations table in the result?

Thanks.

And also, I watch this youtube videos and thisperson checking 3 graph and saying "All assumptions are okey.". Is there no consensus on these assumptions in the literature?
1. https://youtu.be/1DBujFQsJ_U?t=770
2. https://youtu.be/CiYHoxXKA5w?t=1127

The residuals-predicted scatterplot:

An important aspect of the mixed model, as opposed to the GLM, is that the assumptions changed along the model. However, the model is quite robust against vialotions. Check this out: https://besjournals.onlinelibrary.wiley ... 210X.13434.
The most important thing is that the model is well specified (no missing cluster variable, no missing random effects, etc). Distributional concerns (within reasonable bounds), are less important.

Thank you, I am gonna look the link you gave me (Probably I am gonna cite this paper). I am gonna give the graph that I shared in here, and I am gonna point out "This is not a must-have".

And also this is not relevant, but I want to say while it is coming to my mind.
I had tried JReshape module for filtering and merging student and school data sets in PISA 2022. It is working but;
1. It is not saving the missing values from the source datasets to result dataset. For example, there is a data with value 99 in the source dataset, but in the result dataset, it doesn't recognize this data as a missing value, so, in this case when I doing mixed model analysis, it is not excluding the data.
2. It is not saving the description of variables.

Because of these, I have wrote two scripts for filtering dataset and merging dataset (with Python and R) and all of them working. But, if a filter a dataset with this scripts (Python or R), HLM 8 software loading wrong and showing 1 data in the dataset. If you want to look at it I can share it here (I gived this scripts MIT License).