Hi All - has anyone else experienced different stat testing results with custom tables when the table is weighted within the custom tables syntax vs. when it is weighted outside using the weight by command? Which is the correct way?
Here is an example -
Example where table is weighted within Ctables syntax
CTABLES
/WEIGHT VARIABLE=hvolume_val_1_OUT_MF
/VLABELS VARIABLES=s12 dcountry DISPLAY=DEFAULT
/TABLE s12 BY dcountry [COUNT F40.0]
/CATEGORIES VARIABLES=s12 dcountry ORDER=A KEY=VALUE EMPTY=INCLUDE
/CRITERIA CILEVEL=95
/COMPARETEST TYPE=PROP ALPHA=0.05 ADJUST=BONFERRONI ORIGIN=COLUMN INCLUDEMRSETS=YES
CATEGORIES=ALLVISIBLE MERGE=NO SHOWSIG=NO.
vs. same output but different stat testing when the table is weighted externally as in the example below:
weight by hvolume_val_1_OUT_MF.
CTABLES
/VLABELS VARIABLES=s12 dcountry DISPLAY=DEFAULT
/TABLE s12 BY dcountry [COUNT F40.0]
/CATEGORIES VARIABLES=s12 dcountry ORDER=A KEY=VALUE EMPTY=INCLUDE
/CRITERIA CILEVEL=95
/COMPARETEST TYPE=PROP ALPHA=0.05 ADJUST=BONFERRONI ORIGIN=COLUMN INCLUDEMRSETS=YES
CATEGORIES=ALLVISIBLE MERGE=NO SHOWSIG=NO.
weight off.
The weighting is more conservative when it is weighted within Ctables and that makes me think that perhaps its best to use that. Can someone advise on what is the correct or more accurate method.