SEM with mediator

Natti1201 · Post by **Natti1201** » Thu Jul 25, 2024 9:35 am

Hi together,

i want to do a SEM with mediator and get the following error messages. Can somebody help me out?

Kind regards
Natalie

: Bildschirmfoto 2024-07-25 um 11.31.49.png (14.9 KiB) Viewed 5006 times

mcfanda@gmail.com · Post by **mcfanda@gmail.com** » Thu Jul 25, 2024 12:10 pm

It is very likely that your model cannot be estimated. It could be misspecified, missing values, sample size, or many other reasons. If you share a omv file we can check the reason.

Natti1201 · Post by **Natti1201** » Thu Jul 25, 2024 12:49 pm

The file is attached.

Thank you for your help!

FYI: i have no missing values and I think that the sample size is also not the problem.

The model should be the followed:

mediator = Glaubwuerdigkeit
Independent Variable: Provokation (GruppeA_B)
Dependent Variable: 3: EinstellungWerbung, EinstellungMarke, Kaufabsicht

mcfanda@gmail.com · Post by **mcfanda@gmail.com** » Thu Jul 25, 2024 11:31 pm

Remove all the constraints in "custom model settings". Set the structural model in "Endogenous models". Deselect "Latent vars" in "Variances and Covariances" (you define their path, so no covariance is needed).
here is an example that works (but check if the model is the one you need)

Rohdaten_jamovi_questionfile_mc.omv: example; (615.88 KiB) Downloaded 272 times

Natti1201 · Post by **Natti1201** » Fri Jul 26, 2024 8:08 am

Thank you very much! I think the pain point was that I tried to specify the model under custom model settings! Can you explain to me when I use the part "custom model setting" and when I use the part "endogenous model" to specify the model?

Do you have experience in reporting a SEM in a thesis? Which values are necessary to report?

Kind regards

jamovi