Troubleshooting Error Message - MxTryHard()

Posted on
No user picture. Hal Joined: 09/10/2021
##Hi,

##I'm running the univariate ADE model using Hermine's script.
##When it comes to testing the assumptions of equal means and variances, I keep getting the following error message:

```
Running oneEMVOc with 6 parameters
Error: The job for model 'oneEMVOc' exited abnormally with the error message: fit is not finite (The continuous part of the model implied covariance (loc2) is not positive definite in data 'MZ.data' row 526. Detail:
covariance = matrix(c( # 2x2
0.8, 2.37180447958956
, 2.37180447958956, 0.8), byrow=TRUE, nrow=2, ncol=2)
)
In addition: Warning message:
In model 'oneEMVOc' Optimizer returned a non-zero status code 10. Starting values are not feasible. Consider mxTryHard()
```

##I've used mxTryHard and it still will not run

```
Error in h(simpleError(msg, call)) :
error in evaluating the argument 'x' in selecting a method for function 'print': non-numeric argument to mathematical function
```

##I also tried changing the start values manually, and I get the same error message listed above.

Replied on Thu, 07/07/2022 - 11:47
Picture of user. AdminRobK Joined: 01/24/2014


covariance = matrix(c( # 2x2
0.8, 2.37180447958956
, 2.37180447958956, 0.8), byrow=TRUE, nrow=2, ncol=2)
)

You will definitely need to set the start values such that the off-diagonal elements are no greater than the main diagonal elements.