Hi everyone,
As robustness check, I want to further evaluate the Null results of my probit models. Therefore, i want to proceed as follows:
for each model that produced a null result I want to simulate 10 000 draws from a multivariate normal distribution using the estimated coefficients and covariance matrices. Further, I want to calculate a baseline probability by setting the values of the control variables to their means or modes and setting the key variable of interest variable at the mode, too. This should result in 10 000 predicted probabilities for each model, and I calculate the mean probability as the baseline.
Then I want to change the value of each variable of interest to its maximum and calculated the mean change in the probability from the baseline, with a 90% confidence interval.
What I got so far:
xtprobit paramilitary VDem $controls lag_paramilitary, re vce(robust) nolog nocons
matrix beta=e(b)
matrix var=e(V)
drawnorm paramilitary VDem $controls lag_paramilitary, n(10000) means(beta) cov(var)
*calculate a baseline probability by setting the values of the controls and IVs to their means/modes
I do not know how to use the values to calculate the 10000 probabilities...
Does anyone has a solution?
As robustness check, I want to further evaluate the Null results of my probit models. Therefore, i want to proceed as follows:
for each model that produced a null result I want to simulate 10 000 draws from a multivariate normal distribution using the estimated coefficients and covariance matrices. Further, I want to calculate a baseline probability by setting the values of the control variables to their means or modes and setting the key variable of interest variable at the mode, too. This should result in 10 000 predicted probabilities for each model, and I calculate the mean probability as the baseline.
Then I want to change the value of each variable of interest to its maximum and calculated the mean change in the probability from the baseline, with a 90% confidence interval.
What I got so far:
xtprobit paramilitary VDem $controls lag_paramilitary, re vce(robust) nolog nocons
matrix beta=e(b)
matrix var=e(V)
drawnorm paramilitary VDem $controls lag_paramilitary, n(10000) means(beta) cov(var)
*calculate a baseline probability by setting the values of the controls and IVs to their means/modes
I do not know how to use the values to calculate the 10000 probabilities...
Does anyone has a solution?
