Dear fellow researchers
I would like to run an a-posteriori Monte Carlo study to determine the necessary sample size for my SEM in Mplus with categorical/ordinal data but I cannot wrap my head around the systax I need.
I have 13 latent factors with 5 indicators (on average) and a total sample size of n = 450.
Any help is greatly appreciated.
Please note that I cannot use R.
Many thanks
Marcel
I haven't done this myself (with ordinal data). I would assume that you do it like a Monte Carlo study for continuous data, but with the requisite parameters for ordinal data.
Chapter 12 of the Mplus manual in combination with Examples 5.2 and 5.10 might be helpful, as well as this paper:
Muthén, L. K., & Muthén, B. O. (2002). How to use a Monte Carlo study to decide on sample size and determine power. Structural Equation Modeling, 9(4), 599-620.
While, of course, I don't know the purpose of your study, I want to mention that post-hoc power analyses are usually of limited value.
Dear Heiko
many thanks for your swift reply and the recommended literature. In the meantime, I have managed to tackle the problem with the chapters you mentioned. Mostly, anyway. At the time when I conducted my main study, I did not even know what Monte Carlo studies were, so I thought I could demonstrate that I am aware of them and know how to run one, whatever the result might be.
Greetings from Göttingen
Marcel
http://www.statmodel.com/discussion/messages/23/57.html?1548621130
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size