Fit indices is different between lavaan and AMOS

[Asghar Minaei]

I am doing a measurement invariance analysis of a questionnaire between two groups of girls and boys. I performed configural invariance analysis with two software lavaan and AMOS. The estimated values ​​of the parameters in both software are exactly the same, but the values ​​of the fit indices such as GFI, AGFI, and especially RMSEA are different. Is there a reason and justification for  difference in the values ​​of the fit indices between these two software, or these indices should be exactly the same?

My R Codes-

mod.inva <- "
    behavior =~ adapt1 + adapt2 + adapt3 + adapt4 + adapt5
    customer =~ delight1 + delight2 + delight3
    word =~ WOM1 + WOM2 + WOM3
 
"
fit.config <- cfa(mod.inva, data =CustomerDelight,  group = "experience")

I attached some parts of softwares outputs.

Thanks in advance for your help.

Asghar.

[Christian Arnold]

Although only weakly related to your question, it should be noted that you cannot produce valid statements about configural invariance by comparing fit indices: Jorgensen, T. D., Kite, B. A., Chen, P.-Y., & Short, S. D. (2018). Permutation randomization methods for testing measurement equivalence and detecting differential item functioning in multiple-group confirmatory factor analysis. Psychological Methods, 23(4), 708--728

A practical application of the permutation test can be found for example here: 

https://link.springer.com/article/10.1007/s12208-021-00284-5

---

Von: lav...@googlegroups.com <lav...@googlegroups.com> im Auftrag von Asghar Minaei <asghar...@gmail.com>
Gesendet: Montag, 26. September 2022, 08:26
An: lavaan <lav...@googlegroups.com>
Betreff: Fit indices is different between lavaan and AMOS

--
You received this message because you are subscribed to the Google Groups "lavaan" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lavaan+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/lavaan/1f178810-a280-4c50-828c-3208fa1a5a8dn%40googlegroups.com.

[Mauricio Garnier-Villarreal]

Asghar

I also notice that the AIC and BIC are different, it sounds like all the LL based indices have differences. First make sure you have the exact same model, same df etc. And if it is the same model, my guess would be that AMOS uses a sliglthy different calculation for chi-square, like dividing be N instead N-i, for example. Look at their manual to see what formulas they use

[Jeremy Miles]

Are the chi-square and df the same?

Lavaan doesn't provide GFI and AGFI (AFAIK).

--

[Asghar Minaei]

Thank you Mauricio. The df in both software is exactly the same. Even the values of the model parameters are the same.

Asghar

You received this message because you are subscribed to a topic in the Google Groups "lavaan" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/lavaan/xZi6bnBuHPo/unsubscribe.
To unsubscribe from this group and all its topics, send an email to lavaan+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/lavaan/338491c4-b78f-481f-9a0f-1622e09f1946n%40googlegroups.com.

[Asghar Minaei]

Thank you Jeremy Miles. Yes, the chi-square and df are exactly the same in both softwares. It seems that lavaan produces 39 fit indices including GFI and AGFI, which can be extracted with the following command.

fitMeasures(fit)

Asghar.

You received this message because you are subscribed to a topic in the Google Groups "lavaan" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/lavaan/xZi6bnBuHPo/unsubscribe.
To unsubscribe from this group and all its topics, send an email to lavaan+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/lavaan/CAMtGSxm4kfjG_ZdzMCCBN0WXAwOnxGNw_VhiGDJVPx-2HZpTLA%40mail.gmail.com.