|
|
 |
| Nonnormal DV in analysis type TWOLEVE... |
 
|
| Message/Author |
|
|
|
|
Hi Dr. Muthen,
I am trying to run a 2-1-1 mediation MSEM model where the analysis is TWOLEVEL RANDOM. However I am running into the following issue.
I have a continuous DV that is highly non-zero inflated (positive integer) with a positively skewed distribution. The DV is on a 5 point likert scale from 1= very slightly or not at all to 5= extremely. A majority of participants endorsed "very slightly or not at all". Given that my DV is continuous, I tried to use DISTRIBUTION = SKEWT" but it doesn't seem to work, maybe because I'm not using a mixture model. Defining the DV as log10 also does not address the problem, I think because the data is so inflated. How else might I handle my DV variable in mplus?
Thanks! |
|
|
|
|
You can handle this in several ways:
- treat the DV as categorical
- treat the DV as censored
- model the DV using two-part modeling (see UG). |
|
|
|
|
Hi Dr. Muthen,
Thanks for your suggestion. I'd like to try to model the DV using a two-part model. I wondered if you could please take a look at my syntax and let me know whether I specified everything correctly.
DATA TWOPART:
NAMES= y;
BINARY= bin1;
CONTINUOUS = cont1;
CUTPOINT=1;
TRANSFORM=NONE;
VARIABLE:
NAMES= id x m y;
USEVARIABLES ARE
id x m bin1 cont1;
MISSING=.;
BETWEEN IS x;
CLUSTER IS id;
ANALYSIS: TYPE IS TWOLEVEL RANDOM;
MODEL:
%WITHIN%
m bin1 cont1;
bin1 ON m;
cont1 ON m;
%BETWEEN%
x m bin1 cont1;
m on x(a);
bin1 on m(b1);
cont1 on m(b2);
bin1 on x;
cont1 on x;
MODEL CONSTRAINT:
NEW(indb1);
NEW(indb2);
indb1=a*b1;
Indb2=a*b2;
OUTPUT: TECH1 TECH8 CINTERVAL; |
|
|
|
|
| Please send your output to Support along with your license number. |
|
|
| Back to top |
|
|
