dimensionality decision with bootEGA

[mrsrome20]

I am employing bootEGA to determine dimensionality, and I've observed a discrepancy in the decision when utilizing raw data compared to the polychoric matrix. The items under consideration are ordinal, featuring 4 or 5 categories each. The raw data bootEGA unequivocally indicates the presence of 4 dimensions, whereas the polychoric matrix bootEGA suggests 3 dimensions.

This inconsistency arises exclusively when employing the "glasso" method, as both decisions align when using the "tmfg" approach. Upon scrutinizing my data, I noted a lack of multivariate normality, which leads me to speculate whether this non-normality is a key factor influencing the discrepancy. Consequently, should I prioritize the "tmfg" method for its robustness in handling non-normally distributed data?

Thank you very much for dedicating your time to assist me in this matter. Your guidance is highly appreciated.

[hfgolino]

Hi. Use the total entropy fit index to decide which dimensionality structure fits the data better: the 3 or the 4-dimension solution. Cheers, Hudson *Hudson Golino* *Associate Professor* *Department of Psychology* *E* ***@***.*** ***@***.***> *University of Virginia* 1023 Millmont St, Office 115 Charlottesville, VA 22903 Em seg., 18 de dez. de 2023 às 15:25, mrsrome20 ***@***.***> escreveu:

…

Hi, I have a problem. When I use bootEGA to decide the dimensionality, the decision changes when I use the raw data or when I use the polychoric matrix. My items are ordinal and have 4 or 5 categories. Which should I trust? The raw data decision or the polychoric matrix decision? Raw data bootEGA clearly says that there are 3 dimensions and the polychoric matrix bootEGA says that there are 4 dimensions. — Reply to this email directly, view it on GitHub <#147>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABFLKG6XVSTB7A3Z2G5OJO3YKCDDBAVCNFSM6AAAAABAZ6MG62VHI2DSMVQWIX3LMV43ASLTON2WKOZSGA2DOMRQGY3TKMI> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>