|
|
 |
| How to deal subgroups of stratified s... |
 
|
| Message/Author |
|
|
| Youngsoo You posted on Wednesday, September 06, 2017 - 10:02 pm
|
|
|
Dear Drs. Muthén!
I have a data set collected through stratified sampling.
The number of samples in the data set is 400 and each sample has values of 11 variables, including its strata and weight.
I divided the data set into 4 groups by a variable (a0), which is neither strata nor weight variables.
I wanted to compare the descriptive statistics and the regression estimation result of each group.
So, I got a result with the following command syntax;
TITLE:
HG_feul poverty_test1
DATA:
FILE IS hg_fp_test1.txt ;
FORMAT IS free ;
LISTWISE IS OFF;
VARIABLE:
NAMES ARE id strata pw HB RF TW W_HL WIN BOT ef a0;
USEVARIABLES ARE strata pw HB RF TW W_HL WIN BOT ef a0;
MISSING are all(-999);
STRATIFICATION IS strata;
WEIGHT IS pw;
ANALYSIS:
TYPE = COMPLEX ;
iteration = 2000;
MODEL:
ef on HB RF TW W_HL WIN BOT;
OUTPUT:
stdyx;
SAMPSTAT;
I'm wondering if this syntax is right, even though the result looks correct, I guess. |
|
|
|
|
| Your input doesn't have anything related to a0. You can create 3 dummy indicators out of a0 and add those to the regression. |
|
|
| Youngsoo You posted on Thursday, September 07, 2017 - 5:57 pm
|
|
|
Oh, I already split the data set into 4 data sets and I used the same syntax with 4 data sets respectively.
Was this way wrong?
But, thank you for your advice.
I'll try it! |
|
|
| Back to top |
|
|
