|
|
Bootstrapping and extremely large SE |
|
Message/Author |
|
|
I am running a path model with 6 obs vars. One of them is binary (1/0) so declared as categorical.The model is just identified (RMSEA=0/CFA=TFI=1). I am mainly interested in indirect effects. 1) estimator=WLSMV & parameterization=theta: I get huge SE for some paths, like larger than 1000 (for other paths, less than 1). However, CI results seem to provide decent results. I wonder if I can simply report CI results, not SE results in this case. Or I cannot trust any numbers from this analysis? 2) estimator=WLSMV & parameterization=delta: I get normal SE for all paths. The CI results also seem to agree with the CI results in 1). Can I use a delta method when one of the variables (dependent var) is binary and testing indirect effects? Please advise! |
|
|
We need to see your full output to say - send to Support along with your license number. |
|
Back to top |
|
|