Table of direct, indirect, and total effects


asdocx offers a convenient way to export tables of total, direct, and indirect effects to MS Word, Excel, LaTeX, or HTML. asdocx compiles various statistics from the stored matrices and creates a complete table including coefficients, standard errors, z-statistics, and p-values. The command syntax is given below:

Syntax


* sem model should be run first
asdocx estat teffects, [effects(te de ie) ///
title(table_title) dec() dect() decp() ///
notes(table_notes) replace label tzok]

Options enclosed in square brackets [ ] are optional. If they are not used, a table of total effects will be created.

Options


effects() – Specifies the types of effects to report. You can specify all effects like this: effects(te de ie), or a single effect like this: effects(te), where:

te: Total effects
de: Direct effects
ie: Indirect effects

title(table_title) – Table title displayed at the top of the table.

notes(table_notes) – Table notes displayed at the bottom of the table.

dec() – Sets the number of decimal places for coefficients. The default is 3.

dect() – Sets the number of decimal places for z-statistics. The default is 2.

decp() – Sets the number of decimal places for p-values. The default is 4.

replace – Replaces any existing table. The default behavior is to append.

label – Displays variable labels instead of variable names.

tzok – Displays an equal number of decimal places, even if trailing values are zeros.

As noted above, estat teffects is valid only after a sem model has been run. If you run the command directly, you will receive the error:

last estimates not found

 


Let’s use an example of a sem model that is given in its help file.

* Load example dataset
webuse sem_sm1, clear

* Run the sem model
sem (r_occasp <- f_occasp r_intel r_ses f_ses) ///
    (f_occasp <- r_occasp f_intel f_ses r_ses), ///
     cov(e.r_occasp*e.f_occasp)

* Estimate and export the effects to Word file
* Without using the effects() option, the default of 'te' (total effects) is used
asdocx estat teffects
Table: Total effects
Variables Coef. Std. Err. z P>|z|
r_occasp 0.062 0.059 1.05 0.294
f_occasp 0.295 0.147 2.01 0.044
r_intel 0.303 0.051 5.94 0
r_ses 0.19 0.05 3.8 0
f_ses 0.153 0.051 3 0.003
f_intel 0.109 0.052 2.1 0.036
r_occasp 0.225 0.177 1.27 0.204
f_occasp 0.062 0.059 1.05 0.294
r_intel 0.064 0.049 1.31 0.19
r_ses 0.12 0.048 2.5 0.012
f_ses 0.201 0.049 4.1 0
f_intel 0.392 0.05 7.84 0
Notes:

 

Example 2: Report all effects


To report selected or all effects, use the effects() option.

* Load another dataset
use https://stats.idre.ucla.edu/stat/data/hsbdemo, clear

sem (read <- write math ses socst ) ///
    (science <- read math awards ) ///
    (write <- math read awards )

asdocx estat teffects, effects(te de ie) tzok replace
Indirect effects
Variables Coef. Std. Err. z P>|z|
read
read 0.008 0.006 1.33 0.1835
write 0.001 0.001 1.00 0.3173
math 0.020 0.017 1.18 0.2380
ses 0.006 0.007 0.86 0.3898
socst 0.002 0.002 1.00 0.3173
awards 0.480 0.377 1.27 0.2041
write
read 0.000 0.001 0.00 1.0000
write 0.008 0.006 1.33 0.1835
math 0.029 0.029 1.00 0.3173
ses 0.049 0.065 0.75 0.4533
socst 0.019 0.020 0.95 0.3421
awards 0.030 0.022 1.36 0.1738
science
read 0.003 0.002 1.50 0.1336
write 0.042 0.033 1.27 0.2041
math 0.156 0.039 4.00 0.0001
ses 0.264 0.248 1.06 0.2891
socst 0.101 0.030 3.37 0.0008
awards 0.161 0.131 1.23 0.2187
Notes:
Direct effects
Variables Coef. Std. Err. z P>|z|
read
read 0.000 0.000 . .
write 0.124 0.095 1.31 0.1902
math 0.444 0.074 6.00 0.0000
ses 0.780 0.717 1.09 0.2757
socst 0.299 0.064 4.67 0.0000
awards 0.000 0.000 . .
write
read 0.063 0.061 1.03 0.3030
write 0.000 0.000 . .
math 0.133 0.056 2.38 0.0173
ses 0.000 0.000 . .
socst 0.000 0.000 . .
awards 3.831 0.246 15.57 0.0000
science
read 0.335 0.068 4.93 0.0000
write 0.000 0.000 . .
math 0.356 0.077 4.62 0.0000
ses 0.000 0.000 . .
socst 0.000 0.000 . .
awards 0.579 0.359 1.61 0.1074
Notes:
Total effects
Variables Coef. Std. Err. z P>|z|
read
read 0.008 0.006 1.33 0.1835
write 0.125 0.096 1.30 0.1936
math 0.465 0.067 6.94 0.0000
ses 0.786 0.722 1.09 0.2757
socst 0.301 0.065 4.63 0.0000
awards 0.480 0.377 1.27 0.2041
write
read 0.063 0.061 1.03 0.3030
write 0.008 0.006 1.33 0.1835
math 0.162 0.045 3.60 0.0003
ses 0.049 0.065 0.75 0.4533
socst 0.019 0.020 0.95 0.3421
awards 3.861 0.244 15.82 0.0000
science
read 0.338 0.069 4.90 0.0000
write 0.042 0.033 1.27 0.2041
math 0.512 0.071 7.21 0.0000
ses 0.264 0.248 1.06 0.2891
socst 0.101 0.030 3.37 0.0008
awards 0.740 0.374 1.98 0.0477
Notes: