JASP mediation analysis

cmk

Hello,

I'm trying to do a mediation analysis and JASP and everything seems to be going smoothly, however, when I plot the model, the error seems to be incredibly high on the predictor variable (304; see image). It's a rather modest sample size (n =236) and other errors seem reasonably low, so I'm not sure what to make of it. Any ideas would be helpful.

Ihnwhi

Hi @cmk,

Thanks for your question!

In your model, it seems like the value of 304 is a variance of the variable PCI, not the error. This is because, in SEM, the variances of the exogenous latent variables are estimated. That being said, when you take a look at the descriptive statistics of the variables in your dataset, the variance of PCI should be equal to 304. Otherwise, please let us know!

Cheers,

Ihnwhi

cmk

Hi,

Thanks for you response. It was very helpful. The descriptives produced a variance of 305.144, so only slightly more than what was displayed in the diagram

All the best

Ihnwhi

Hi @cmk,

Glad to hear that the response was helpful. Also, it is a great observation that 304 and 305.144 are not exactly the same.

Please let me explain the reason for the difference between 304 and 305.144 for you. The value of 305.144 is the variance PCI in your observed data (i.e., sample statistics). On the other hand, 304 is the estimated variance of PCI.

In SEM, parameters are estimated in a way that minimizes the difference between the observed covariance matrix and the model-implied covariance matrix (more technically speaking, this minimization leads to the maximization of the log-likelihood function). Here, the value of 304 is a result of such a minimization, hence being different from the value of 305.144 from your sample.

Cheers,

Ihnwhi