Anonymous posted on Sunday, September 11, 2005 - 5:12 am
Dear Prof. Muthen
Mplus is a wonderful and fantastic software! Many of my colleagues have converted to Mplus-er from Liserlites, AMOSer...
as Mplus 4 is coming out soon...I hope to give some suggestions :--
1) allows type=logistic to be possible in combination with type=twolevel missing random,etc. and beyond univariate analysis.
2) Allow probit analysis with both THETA & DELTA parameterization in TYPE=MIXTURE.
3) To be able to read data files from other softwares e.g., excel, SPSS, SAS format.
4) Regarding documentation, will it be possible to add in some descriptions on the model formulations in the beginnin of each chapter before the Mplus examples are listed (or another separte guide on these model formulations). In some advanced topics (multilevel and mixture modeling), sometimes it is quite hard to relate the mplus function to the actual models are being fitted.. It is not straightfoward to relate the tech. appendix with the examples in Mplus manual...
5) more flexible parameter constraints, such as allow Mplus default parameters (say threshold) be user-defined parameter.
1. Logisitc regression is avaliable throughout Mplus. The maximum likelihood estimator with categorical outcomes estimates logistic regression. TYPE=LOGISTIC; will most likely be removed in Version 4 to avoid this confusion.
2. I will add this to the suggestion list.
3. I will add this to the suggestion list but it is doubful that we will ever do this as we have more important statistical topics to tackle and creating ASCII files is a relatively simple task.
4. The technical appendices on the website contain this type of informaton. They will be updated to be more current. We will also be writing books on these topics. Our short course handouts can also be purchased at a nominal cost and they contain formulas for each topic.
5. I am not sure that I understand this suggestion. All parameters are within control of the user as far as I know. Are you talking about expanding the capabilities of MODEL CONSTRAINT?