how can I know there is a causal relationship between the variables before drawing the SEM model? i mean how can i know there is relationship between the predictors variables to link them in sem model? I don't have any latent variables in my data. just I have predictor variables and 2 dependant variable. and how can I predict the event in sem? thank you
thank you for informing me
On
It sounds like you are trying to infer a Structural Causal Model (SCM) from data. This task is typically referred to as "Causal Discovery", or "Causal Structure Learning". The python library causal-learn has some of the best-established methods for doing so, and there are many other approaches and software packages.
An important note of caution: Causal Discovery requires very strong assumptions on the data generating process, and I recommend you take great care to check your assumptions and the plausibility of results when applying causal discovery to real-world data.
Try just running a binary correlation matrix. Or, run a standard regression for each of your DVs and check the variance inflation factor to see if the IVs are multicollinear. Either technique will establish relationships among your variables.