Hi everyone,
I am in the process of trying to run a cross-lagged panel design with a pre-test and post-test segment.
While both the pre-test and post test are in STATA, the only way to separate without creating brand new, complicated variables is to do something on the lines of the following:
sem (Disorg -> mygang mydrugs mydruguse) (CE -> mygetalong mypee mytrustneigh), group(my-repost)....
In doing so, I get something that looks like this:
Structural |
myworry <- |
agerecode |
Pre-Test | .0002951 .0049864 0.06 0.953 -.0094781 .0100684
Post-Test | .0121706 .0086342 1.41 0.159 -.0047521 .0290933
mygender |
Pre-Test | .0854836 .1267767 0.67 0.500 -.1629941 .3339613
Post-Test | -.3100499 .2404826 -1.29 0.197 -.7813872 .1612874
section |
Pre-Test | .0110405 .0269222 0.41 0.682 -.041726 .0638069
Post-Test | -.0037216 .0602451 -0.06 0.951 -.1217998 .1143566
Disorg |
Pre-Test | -.4868334 .1932851 -2.52 0.012 -.8656652 -.1080016
Post-Test | .3857709 .3016247 1.28 0.201 -.2054027 .9769445
CE |
Pre-Test | -.5970583 .1761851 -3.39 0.001 -.9423749 -.2517418
Post-Test | -.7535653 .3012077 -2.50 0.012 -1.343922 -.163209
_cons |
Pre-Test | 2.413731 .2075027 11.63 0.000 2.007034 2.820429
Post-Test | 1.466647 .3823612 3.84 0.000 .7172327 2.216061
For my estat coef command I get an output similar to this (excluding other variables for clarity):
myworry <-
Disorg |
Pre-Test | -.4868334 _b[myworry:0bn.myprepost#c.Disorg]
Post-Test | .3857709 _b[myworry:1.myprepost#c.Disorg]
Stability coefficient test:
. estat stdize: test _b[myworry:1.myprepost#c.Disorg] = _b[myworry:0bn.myprepost
> #c.Disorg]
( 1) - [myworry]0bn.myprepost#c.Disorg + [myworry]1.myprepost#c.Disorg = 0
chi2( 1) = 6.93
Prob > chi2 = 0.0085
My questions are the following:
1) Is using the group command in this formula OK for doing this type of analysis or should I be using something else (IE: creating different variables).
2) If I am doing this correctly, Would I interpret this as follows: there was a positive, significant change of worry under improvements to social disorganization b = 0.86, p<0.01?
(My reasoning for this: difference 0.38- -0.48 = 0.86; P value= 0.0085?)
3) more generally, because I haven't seen a separate thread for this anywhere --> what the heck is the difference between the sort by command and the group command?
Thank you!!
I am in the process of trying to run a cross-lagged panel design with a pre-test and post-test segment.
While both the pre-test and post test are in STATA, the only way to separate without creating brand new, complicated variables is to do something on the lines of the following:
sem (Disorg -> mygang mydrugs mydruguse) (CE -> mygetalong mypee mytrustneigh), group(my-repost)....
In doing so, I get something that looks like this:
Structural |
myworry <- |
agerecode |
Pre-Test | .0002951 .0049864 0.06 0.953 -.0094781 .0100684
Post-Test | .0121706 .0086342 1.41 0.159 -.0047521 .0290933
mygender |
Pre-Test | .0854836 .1267767 0.67 0.500 -.1629941 .3339613
Post-Test | -.3100499 .2404826 -1.29 0.197 -.7813872 .1612874
section |
Pre-Test | .0110405 .0269222 0.41 0.682 -.041726 .0638069
Post-Test | -.0037216 .0602451 -0.06 0.951 -.1217998 .1143566
Disorg |
Pre-Test | -.4868334 .1932851 -2.52 0.012 -.8656652 -.1080016
Post-Test | .3857709 .3016247 1.28 0.201 -.2054027 .9769445
CE |
Pre-Test | -.5970583 .1761851 -3.39 0.001 -.9423749 -.2517418
Post-Test | -.7535653 .3012077 -2.50 0.012 -1.343922 -.163209
_cons |
Pre-Test | 2.413731 .2075027 11.63 0.000 2.007034 2.820429
Post-Test | 1.466647 .3823612 3.84 0.000 .7172327 2.216061
For my estat coef command I get an output similar to this (excluding other variables for clarity):
myworry <-
Disorg |
Pre-Test | -.4868334 _b[myworry:0bn.myprepost#c.Disorg]
Post-Test | .3857709 _b[myworry:1.myprepost#c.Disorg]
Stability coefficient test:
. estat stdize: test _b[myworry:1.myprepost#c.Disorg] = _b[myworry:0bn.myprepost
> #c.Disorg]
( 1) - [myworry]0bn.myprepost#c.Disorg + [myworry]1.myprepost#c.Disorg = 0
chi2( 1) = 6.93
Prob > chi2 = 0.0085
My questions are the following:
1) Is using the group command in this formula OK for doing this type of analysis or should I be using something else (IE: creating different variables).
2) If I am doing this correctly, Would I interpret this as follows: there was a positive, significant change of worry under improvements to social disorganization b = 0.86, p<0.01?
(My reasoning for this: difference 0.38- -0.48 = 0.86; P value= 0.0085?)
3) more generally, because I haven't seen a separate thread for this anywhere --> what the heck is the difference between the sort by command and the group command?
Thank you!!
