exporting pvalues (computed using ritest) to final regression table using esttab

[Sanj Ari]

Hi,

I am new to Stata. I need to export the pvalues that I have computed using ritest (randomization inference), which I have saved in a matrix, to the final regression table alongside the std errors and coefficients. The following is the code I have for now.

...

eststo: quietly areg GAD7_line_average_el CBT_el IHT_el GAD7_line_average, absorb(factcode) robust 

sum GAD7_line_average_el if PC==1
estadd scalar Control_Mean = r(mean)

* Randomization Inference GAD-7 *

ritest CBT _b[CBT], reps(100) seed(125) strata(factcode): quietly areg GAD7_line_average_el CBT_el IHT_el GAD7_line_average, absorb(factcode) robust

matrix pvalues_CBT_GAD7 = r(p)
mat colnames pvalues_CBT_GAD7 = "Prevalence of Anxiety (GAD-7>=8)"
mat rownames pvalues_CBT_GAD7 = "Skills Training"

ritest IHT _b[IHT], reps(100) seed(125) strata(factcode): quietly areg GAD7_line_average_el CBT_el IHT_el GAD7_line_average, absorb(factcode) robust

matrix pvalues_IHT_GAD7 = r(p)
mat colnames pvalues_IHT_GAD7 = "Prevalence of Anxiety (GAD-7>=8)"
mat rownames pvalues_IHT_GAD7 = "Health Education" 

** Regressing PHQ9 **

eststo: quietly areg PHQ9_line_average_el CBT_el IHT_el PHQ9_line_average, absorb(factcode) robust 
sum PHQ9_line_average_el if PC==1
estadd scalar Control_Mean = r(mean)

* Randomization Inference PHQ-9 *

ritest CBT _b[CBT], reps(100) seed(125) strata(factcode): quietly areg PHQ9_line_average_el CBT_el IHT_el PHQ9_line_average, absorb(factcode) robust 

matrix pvalues_CBT_PHQ9 = r(p)
mat colnames pvalues_CBT_PHQ9 = "Prevalence of Depression (PHQ-9>=10)"
mat rownames pvalues_CBT_PHQ9 = "Skills Training"

ritest IHT _b[IHT], reps(100) seed(125) strata(factcode): quietly areg PHQ9_line_average_el CBT_el IHT_el PHQ9_line_average, absorb(factcode) robust 

matrix pvalues_IHT_PHQ9 = r(p)
mat colnames pvalues_IHT_PHQ9 = "Prevalence of Depression (PHQ-9>=10)"
mat rownames pvalues_IHT_PHQ9 = "Health Education" 

matrix pvalues = pvalues_CBT_GAD7,pvalues_CBT_PHQ9\pvalues_IHT_GAD7,pvalues_IHT_PHQ9

estadd matrix pvalues = pvalues

** Tabulating the output **

esttab using table8_RI.csv, cells(b(fmt(2) star) se(fmt(2) par) pvalues(fmt(2) par([ ]))) label title(Table 8: Impact on Workers'Psychological Well-Being) nonumbers mtitles ("Prevalence of Anxiety (GAD-7>=8)" "Prevalence of Depression (PHQ-9>=10)") star(* 0.1 ** 0.05 *** 0.01) coeflabels(CBT_el "Skills Training" IHT_el "Health Education" Control_Mean "Control Mean") keep(CBT_el* IHT_el*) stats(Control_Mean N ar2, fmt(2 0 2) labels("Control Mean" "N" "Adj. R-sq")) 

eststo clear

Ideally, the output should be in the following format (the values in square brackets is what I need)

[Nils Enevoldsen]

Does this example help?

sysuse auto, clear

eststo reg1 : reg price trunk

ritest trunk _b[trunk], reps(100) seed(12345) : `e(cmdline)'
matrix pboot = r(p)
mat colnames pboot = trunk
estadd matrix pboot = pboot : reg1

esttab reg1, cells(b se(par) pboot(par([ ]))) label nocons

[Sanj Ari]

@Nils Enevoldsen thank you! How would I replicate the above that you had mentioned when running for multiple regressions as in my case? Since I would have mutiple matrices to save the data each time I run a ritest. I am sorry I am a little confused about this.

[Sanj Ari]

*** (2) Run the regressions ***

** Regressing GAD7 **

eststo reg1: quietly areg GAD7_line_average_el CBT_el IHT_el GAD7_line_average, absorb(factcode) robust

sum GAD7_line_average_el if PC==1
estadd scalar Control_Mean = r(mean)

* Randomization Inference GAD-7 *

ritest CBT_el _b[CBT_el], reps(100) seed(125) strata(factcode): quietly areg GAD7_line_average_el CBT_el IHT_el GAD7_line_average, absorb(factcode) robust

matrix pvalues_CBT_GAD7 = r(p)
mat colnames pvalues_CBT_GAD7 = "pvalue Skills Training"
mat rownames pvalues_CBT_GAD7 = "Skills Training"


ritest IHT_el _b[IHT_el], reps(100) seed(125) strata(factcode): quietly areg GAD7_line_average_el CBT_el IHT_el GAD7_line_average, absorb(factcode) robust

matrix pvalues_IHT_GAD7 = r(p)
mat colnames pvalues_IHT_GAD7 = "pvalue Health Education"
mat rownames pvalues_IHT_GAD7 = "Health Education"

matrix pvalues_GAD7 = pvalues_CBT_GAD7,pvalues_IHT_GAD7
estadd matrix pvalues_GAD7 : reg1

** Regressing PHQ9 **

eststo reg2: quietly areg PHQ9_line_average_el CBT_el IHT_el PHQ9_line_average, absorb(factcode) robust
sum PHQ9_line_average_el if PC==1
estadd scalar Control_Mean = r(mean)

* Randomization Inference PHQ-9 *

ritest CBT_el _b[CBT_el], reps(100) seed(125) strata(factcode): quietly areg PHQ9_line_average_el CBT_el IHT_el PHQ9_line_average, absorb(factcode) robust

matrix pvalues_CBT_PHQ9 = r(p)
mat colnames pvalues_CBT_PHQ9 = "pvalue Skills Training"
mat rownames pvalues_CBT_PHQ9 = "Skills Training"

ritest IHT_el _b[IHT_el], reps(100) seed(125) strata(factcode): quietly areg PHQ9_line_average_el CBT_el IHT_el PHQ9_line_average, absorb(factcode) robust

matrix pvalues_IHT_PHQ9 = r(p)
mat colnames pvalues_IHT_PHQ9 = "pvalue Health Education"
mat rownames pvalues_IHT_PHQ9 = "Health Education"

matrix pvalues_PHQ9 = pvalues_CBT_PHQ9,pvalues_IHT_PHQ9
estadd matrix pvalues_PHQ9 : reg2

** Tabulating the output **

esttab reg1 reg2, cells("b(fmt(2) star)" "se(fmt(2) par)" "pvalues_GAD7(fmt(2) par([ ]))" "pvalues_PHQ9(fmt(2) par([ ]))") label title(Table 8: Impact on Workers'Psychological Well-Being) nonumbers mtitles ("Prevalence of Anxiety (GAD-7>=8)" "Prevalence of Depression (PHQ-9>=10)" "pvalues GAD7" "pvalues PHQ9") star(* 0.1 ** 0.05 *** 0.01) coeflabels(CBT_el "Skills Training" IHT_el "Health Education" Control_Mean "Control Mean") stats(Control_Mean N ar2, fmt(2 0 2) labels("Control Mean" "N" "Adj. R-sq")) drop(_cons GAD7_line_average PHQ9_line_average)

eststo clear

@Nils Enevoldsen when the code is run in the manner you had mentioned the table output is displayed as below. However, I want the output to be in the form of the table I had posted earlier in #1