I just found that there is no "missing" option under "analysis: type=" (p519, Mplus user's guide). For the version 4.1 I am working on, there is such an option. I want to know the reason. I also would like to know which are the appropriate estimators for a SEM analyses with coninuous obersved variables when there are missing data. Also, when the observed variables include categorical variables, which are the appropriate estimators? How do I apply FIML to the missing data using version 6?
Starting with Version 5, it is not necessary to specify TYPE=MISSING. It is the default. You do have to specify missing value flags using the MISSING option of the VARIABLE command.
Sue Lee posted on Thursday, February 03, 2011 - 2:03 pm
I'm using Mplus version 5.21. I converted all missing data-points to period (.) and added the following command under the VARIABLE and ANALYSIS:
VARIABLE: MISSING ARE ALL .;
ANALYSIS: TYPE = MIXTURE MISSING;
And I keep getting the following warning after the input ends.
*** WARNING in ANALYSIS command Starting with Version 5, TYPE=MISSING is the default for all analyses. To obtain listwise deletion, use LISTWISE=ON in the DATA command. 1 WARNING(S) FOUND IN THE INPUT INSTRUCTIONS =============== I do not want listwise deletion (i.e., excluding an entire record from analysis if any single value is missing).
Then, do I have to rerun the data with a different command under the ANALYSIS or can I just use the current analysis result? I don't understand what the warning exactly tells me to do.
You are getting TYPE=MISSING; It is the default. You are not getting listwise deletion. The warning is not telling you to do anything. It is informing you of a default change in the event you want to get listwise deletion.