Multiple Imputation & mixture modelin... PreviousNext
Mplus Discussion > Missing Data Modeling >
Message/Author
 Shelley Brown posted on Friday, April 05, 2019 - 4:06 am
I created 25 multiple imputed datasets within Mplus as per 11.5 example in the MPlus user guide; it worked fine.

Next, I tried to use the newly created MI datasets (named missimplist.dat) in my primary analysis (an LPA with auxiliary covariates) BUT I received this error message:


*** ERROR
The number of observations is 0. Check your data and format statement.
Data file: /USERS/shelleybrown/Documents/2017/RESEARCH/MPLUS/LPA/DATA/missimp1.dat
*** ERROR
Invalid symbol in data file:
"*" at record #: 1, field #: 85


A quick look of the first MI dataset revealed a series of * symbols for the missing data.

How do I resolve? Do I use the missflag option? If so, how and where? Also, note that I have missing data for my primary indicators used for the LPA but missingness for the LPA is addressed via the maximum likelihood option. I am only using MI to address missingness for the auxiliary variables.

Here is my MI code:

TITLE: Created 25 multiple imputed datasets for three-Step LPA-aux
DATA: FILE = lpa_312q.dat;
VARIABLE: NAMES = IDNO ……etc;
Missing are all (-999, 999);
Usevariables are sex ACE_tot YLS_his Genrecd3 …..etc;
AUXILIARY = IDNO etc etc;
DATA IMPUTATION:
IMPUTE = ACE_tot YLS_his age_2 vio_indw;
Ndatasets = 25;
SAVE = missimp*.dat;
ANALYSIS: TYPE = BASIC;
 Bengt O. Muthen posted on Saturday, April 06, 2019 - 12:59 pm
Just put * for the Missing= option in the Variable command (see UG):

Missing = *;
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: