Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Supported by

Unable to specify model predictors in hierarchical regression JASP 0.95

I saw a thread in the discussion forum from 2016 where someone said they were unable to specify predictors for each model in a hierarchical regression, and I am having the same problem now almost 10 years later. I was wondering if I am doing something wrong or if this has just not been updated in later versions of JASP?

Problem: I want to perform a hierarchical regression with 5 predictors, predictors A and B in model 1, and predictors C, D, and E in model 2. All predictors automatically show up in model 1, with none in model 0. I can move predictors from model 1 to model 0; however, once predictors appear in model 0, they cannot be removed, or moved to model 1 again. If I add a model 2, I cannot move any predictors from either model 0 or 1 to model 2.

Comments

  • From my point of view there doesn't seem to be any bug (just tested it). You write: "however, once predictors appear in model 0, they cannot be removed, or moved to model 1 again". Predictors in earlier models automatically and necessarily show up in later models (otherwise models wouldn't be nested). This is normal in hierarchical regression and is not specific to JASP.

    "If I add a model 2, I cannot move any predictors from either model 0 or 1 to model 2." That's because predictors in models 0 and 1 are automatically included in model 2 and later models.

    For example, I've included FACTWGT_b_W142 in model 0, and it gets automatically added to models 1 and 2 and cannot be removed:


  • My issue is that I can only move predictors from model 1 to model 0. I want to leave model 0 as the intercept with no predictors. If I add a model 2, I cannot move predictors from model 1 to model 2. I can only move predictors from model 1 to model 0. Once I have entered predictors into model 0, I cannot remove them.

    I want 0 predictors in model 0, 2 predictors in model 1, and all predictors in model 2. There seems to be no way to do this in JASP.

  • You have to remove predictors from model 1 first, and then add them manually to model 2:

    https://i.imgur.com/vCArdDk.gif

  • Also, there was a bug regarding this functionality in 0.95, and this was solved in the hotfix. So please try 0.95.1, if you haven't already!

    EJ

  • That solved it EJ, thank you!!

Sign In or Register to comment.