SEM error - convergence not achieved r(430)

[Wendy Jane]

Hi I am getting an error when trying to run a confirmatory factor analysis.
The error is "
convergence not achieved
convergence not achieved
an error occurred when bootstrap executed sem
r(430);
"

The model ive built in the SEM builder is this:

The only options ive changed when running the model is to have 50 bootstrap replications and to report standardised coefficients, such that the syntax generated is: sem (L1 -> pa1, ) (L1 -> ee3, ) (L1 -> ee2, ) (L1 -> ee1, ) (L2 -> ee5, ) (L2 -> pa2, ) (L2 -> ee4, ) (L2 -> dep1, ) (L3 -> pa3, ) (L3 -> dep3, ) (L3 -> pa4, ) (L3 -> ee6, ) (L4 -> pa7, ) (L4 -> ee9, ) (L5 -> ee7, ) (L5 -> dep4, ) (L5 -> ee8, ) (L5 -> pa5, ) (L5 -> pa6, ), covstruct(_lexogenous, diagonal) vce(bootstrap, seed(1) dots(1)) standardized latent(L1 L2 L3 L4 L5 ) nocapslatent

Can anyone tell me where I've gone wrong - have I included too many latent variables or something?

I've included an example dataset below from https://youtu.be/bYrO3tSwGik (not my dataset)

* Example generated by -dataex-. For more info, type help dataex
clear
input byte(ee1 ee2 ee3 pa1 dep1 ee4 pa2 ee5 pa3 dep2 dep3 pa4 ee6 ee7 dep4 ee8 pa5 pa6)
4 4 5 4 4 2 7 2 7 2 2 6 3 4 1 1 7 6
2 2 1 7 2 1 7 2 7 1 1 4 2 2 1 1 7 4
6 6 7 7 3 5 6 6 5 4 5 4 4 4 2 5 5 6
7 7 7 7 1 1 7 7 7 1 1 1 1 4 1 1 7 5
6 6 6 6 4 2 6 2 7 6 6 7 4 7 5 2 7 7
2 2 2 7 6 2 7 1 7 1 1 7 3 2 1 2 7 6
6 6 6 7 2 7 7 7 4 1 1 1 6 6 1 7 6 2
4 6 3 7 2 3 7 3 7 3 3 6 3 3 3 3 7 6
6 5 5 7 4 3 7 4 6 4 4 6 3 6 4 4 6 6
4 6 2 4 2 2 6 1 7 2 2 6 3 5 1 2 6 6
6 6 3 6 1 3 6 6 6 4 1 6 6 4 1 2 6 4
2 2 2 7 1 1 7 1 7 1 1 7 1 1 1 1 7 7
4 6 2 7 2 1 7 1 6 2 2 5 3 7 2 1 7 5
6 6 4 6 6 2 7 4 7 2 1 6 4 3 4 2 6 6
6 6 3 4 2 3 5 4 7 3 7 6 4 5 1 4 5 7
2 2 1 7 1 1 7 2 7 2 1 7 2 2 1 1 7 6
7 7 6 7 1 4 6 5 6 1 1 6 5 3 1 2 7 6
6 6 3 4 6 6 6 6 6 6 6 4 5 4 1 6 6 7
4 6 4 7 1 2 7 4 7 2 1 6 4 6 1 1 7 6
4 4 2 7 1 2 7 2 7 1 1 7 1 1 1 1 7 7
7 7 6 7 3 4 6 7 7 2 1 6 7 7 1 6 7 6
2 2 2 7 1 1 7 1 7 1 1 7 1 1 1 1 7 7
7 7 4 7 1 1 7 6 7 6 3 4 6 7 1 1 7 6
2 3 2 7 1 2 7 2 7 1 1 6 2 3 1 2 7 7
6 7 6 4 1 1 7 6 7 1 1 6 6 5 1 1 7 7
2 6 2 3 1 1 3 6 7 1 1 7 5 7 5 1 7 6
4 5 4 4 1 2 6 4 6 4 3 6 4 2 1 4 7 4
5 6 2 7 1 2 7 2 7 2 3 6 3 4 1 2 7 7
1 2 3 7 1 1 7 1 6 1 1 7 3 2 1 1 7 6
6 7 7 7 5 1 7 4 7 4 7 6 7 7 4 4 7 6
2 2 2 7 1 1 7 1 7 2 1 7 1 1 1 1 7 7
2 3 5 5 1 1 5 5 2 4 2 6 6 2 3 1 7 6
6 6 2 7 1 2 7 2 7 2 2 6 2 2 1 1 7 7
4 4 3 6 1 1 7 2 5 1 1 5 1 1 1 1 7 7
4 3 2 7 1 2 7 1 7 1 1 7 1 1 1 1 7 7
5 5 2 7 1 1 7 2 7 1 1 7 2 7 1 1 7 7
6 6 5 4 6 6 6 6 6 5 1 6 7 4 3 4 6 3
3 3 3 6 2 3 6 3 6 3 3 4 2 4 2 3 6 4
3 7 2 7 1 4 6 2 7 1 1 7 3 3 1 3 7 5
4 5 3 6 2 5 6 4 5 4 3 5 5 3 1 4 3 2
6 5 4 6 2 4 5 5 6 2 2 6 4 5 1 4 7 5
3 7 2 7 1 4 6 2 7 1 1 7 3 3 1 3 7 5
4 5 3 6 2 5 6 4 5 4 3 5 5 3 1 4 3 2
6 5 4 6 2 4 5 5 6 2 2 6 4 5 1 4 7 5
6 6 6 6 6 3 6 4 6 6 6 5 5 3 3 1 6 5
7 7 4 7 4 3 5 3 7 5 1 6 1 5 1 3 6 6
7 7 5 5 5 6 6 7 4 4 4 4 4 4 3 5 5 6
4 6 6 6 4 2 6 4 4 1 2 5 3 4 1 2 6 6
4 4 5 7 2 5 6 3 6 4 5 6 5 3 2 3 6 5
7 7 6 6 6 6 6 7 6 5 5 4 7 6 6 6 6 6
6 7 6 7 1 4 6 6 7 1 1 3 4 4 1 6 7 6
2 6 5 7 5 5 6 4 6 1 1 6 3 2 1 4 6 4
4 5 3 7 3 4 6 5 4 5 3 4 4 5 4 4 7 5
2 4 2 6 3 2 6 2 4 2 1 4 2 1 2 2 7 5
3 3 1 7 1 1 7 1 7 1 1 7 1 2 1 1 7 7
3 4 2 7 1 2 7 2 7 3 2 6 2 2 1 2 6 4
5 6 5 6 5 4 6 3 5 1 3 6 6 5 3 3 6 5
2 5 2 7 1 3 7 2 6 2 1 6 5 5 1 3 6 7
3 6 6 7 1 1 7 2 7 1 1 6 3 6 1 4 7 7
4 4 2 6 2 2 6 2 6 2 2 6 3 2 1 2 7 6
2 3 2 6 2 2 6 2 6 1 2 7 3 2 3 2 7 6
5 3 3 7 1 1 7 2 7 1 2 7 2 3 1 1 7 7
2 5 2 5 1 2 7 2 7 1 1 4 2 2 1 2 7 7
5 5 3 6 3 3 6 2 6 1 1 6 3 5 2 1 6 5
7 6 7 7 1 6 7 6 7 2 2 7 3 7 7 3 6 7
4 4 4 5 3 3 6 4 5 4 5 5 4 4 2 2 6 5
2 3 1 5 1 1 6 1 4 1 2 6 2 2 1 2 6 4
3 5 6 7 3 3 7 1 7 1 1 3 3 7 1 1 7 7
7 6 6 7 2 2 7 7 5 1 1 6 7 6 3 2 6 5
5 4 3 7 6 1 6 2 7 1 3 7 2 6 7 1 7 7
4 5 2 4 2 1 6 3 6 3 3 7 5 3 1 3 7 7
6 6 7 7 7 6 6 4 7 7 6 6 4 5 7 7 7 6
7 7 5 7 2 4 6 2 7 1 3 6 2 7 1 4 7 6
4 7 3 7 5 2 6 2 7 1 3 7 1 4 2 2 7 6
6 6 2 7 2 1 7 1 7 1 1 7 4 5 2 1 7 7
2 2 1 5 2 2 4 2 4 2 5 7 3 5 1 2 2 2
3 6 2 7 1 2 7 4 7 2 1 6 3 4 1 2 7 6
2 2 1 7 1 1 7 1 7 1 1 7 1 2 1 1 7 7
7 6 4 7 4 7 7 4 7 2 2 5 6 4 6 5 7 6
2 2 1 7 1 1 6 1 7 1 1 7 1 2 1 2 7 6
5 6 6 5 4 4 4 5 7 2 2 5 5 5 3 3 7 5
4 5 1 7 3 4 7 2 6 1 1 6 4 4 3 2 7 7
4 4 4 6 2 3 7 2 6 4 3 6 4 4 2 2 7 6
7 7 7 7 1 7 7 7 7 2 7 7 7 7 4 1 7 4
5 5 2 7 1 2 7 2 7 1 1 7 3 4 1 1 6 6
6 6 4 6 3 3 6 5 5 4 3 4 4 7 3 2 6 6
2 2 3 7 1 1 7 1 7 1 2 6 2 2 1 1 7 7
2 5 4 6 1 3 6 2 7 1 1 6 3 4 1 4 6 6
2 4 2 6 1 2 7 1 7 1 1 6 2 3 1 1 7 7
4 6 3 7 1 2 6 2 6 1 1 6 3 4 1 1 7 6
4 4 3 7 4 4 6 3 6 4 1 6 3 4 2 2 7 6
6 4 4 7 1 4 7 3 7 1 2 6 6 6 1 4 7 7
3 5 1 2 1 1 6 2 6 1 1 6 1 3 1 1 6 6
2 6 2 7 2 2 7 3 7 1 1 7 3 5 1 2 7 6
6 6 5 5 4 4 6 4 4 3 3 2 4 2 3 3 5 4
2 3 2 5 1 5 6 2 5 2 2 5 2 4 3 2 6 5
6 6 5 6 5 4 6 5 7 1 1 5 4 6 2 3 5 5
4 4 5 6 2 3 6 2 6 3 1 7 6 4 1 2 6 7
5 5 3 7 3 3 7 4 4 4 4 6 6 6 4 2 4 4
2 4 3 6 1 2 6 2 5 1 1 5 2 5 1 2 5 4
end

Thank you very much.

[Weiwen Ng]

One issue is: is the model identified to begin with? L4 only has two indicators, and I think normally the minimum is 3 indicators to identify a latent variable. I suspect this is the cause.

If the overall model is identified, then I would assume that one of the bootstrap replications is failing for some reason. I am not sure how to advise on that.

[Wendy Jane]

Thank you very much!
I don't really understand the rules of whether a model is identified or not but that helped thank you

[Weiwen Ng]

Thank you very much!
I don't really understand the rules of whether a model is identified or not but that helped thank you

you aren’t alone, identification is hard to understand. If you can do matrix algebra, you could be able to check identification that way, but I forget the math and I don’t really know matrices very well. I typically rely on these more hard and fast rules by David Kenny. You want to look under the section on measurement models (as opposed to structural models, whereas you specified a purely measurement model above). Part 2a says what I said, that you normally need three indicators for a latent construct.

Part 2b says that you can have only two indicators if you are willing to constrain their loadings to be equal. That seems a bit of a strong assumption to make, but if SEM researchers have some generally agreed burden of proof that you can use to justify that assumption, you can specify it with a constraint. if you go this route, make sure you know what you’re doing. If you have no idea what this paragraph means, definitely don’t do it.