Let’s load the dataset.

/* Dataset is adopt_singleevent.dta */

use "~/Dropbox/event_history/dta/adopt_singleevent.dta"

Table 6.1

We begin by estimating the Cox and Weibull models, and storing the results to make a table.

* Cox model 

eststo clear 

eststo: stcox mooneymean, exactp nohr basehc(haz_cox)

* Weibull model

eststo: streg mooneymean, dist(weib) nohr

* Reset working directory to collect output

cd ~/Dropbox/event_history/rmd

* Generate regression table output for Cox and Weibull 

esttab using ch6_cox_weib.html, replace ///
    coeflabel(mooneymean "Pre-Roe" _cons "Constant") ///
    title(Models of Adoption of Restrictive Abortion Legislation) ///
    mtitles("Cox Model" "Weibull Model") ///
    eqlabels("", none) ///
    b(2) se(2) nostar ///
    stats(ll N, label("Log-Likelihood" "<em>N</em>") fmt(2 0))

Let’s estimate the Royston-Parmar model in comparison, and display the results in a table.

eststo clear 

eststo:stpm mooneymean, scale(h) df(3)

* Generate regression table output for Royston-Parmar model

esttab using ch6_rp.html, replace ///
    coeflabel(s0:_cons "Spline 1" s1:_cons "Spline 2" s2:_cons "Spline 3" xb:mooneymean "Pre-Roe" ///
    xb:_cons "Constant") ///
    title(Models of Adoption of Restrictive Abortion Legislation) ///
    mtitles("Royston-Parmar Model") ///
    eqlabels("", none) ///
    b(2) se(2) nostar ///
    stats(ll N, label("Log-Likelihood" "<em>N</em>") fmt(2 0))

Figure 6.1

We already computed the baseline hazard for the Cox model earlier in the regression equation, so now we compute the baseline hazards for the Royston-Parmar and Weibull models.

* Compute baseline hazard for Royston-Parmar model 

predict haz_rp, haz zero

* Compute baseline hazard for Weibull based on P.H. parameterization.

streg mooneymean, dist(weib) time

* First, note that lambda=exp(-beta'x).  We can compute lambda for each confict type.

gen lambda_base=exp(-(_b[_cons])) if e(sample)

* Second, note that h(t)=lambda*p*(lambda*t)^(p-1).  We can generate the hazard rate for the baseline.

gen haz_weib=lambda_base*e(aux_p)*(lambda_base*yrtoadp)^(e(aux_p)-1)

twoway (line haz_rp yrtoadp, sort lpattern(solid)) (line haz_weib yrtoadp, sort lpattern(solid)) ///
    (line haz_cox yrtoadp, sort connect(stairstep) lpattern(solid)), ///
    legend(off) ///
    xtitle("Years Since Roe vs. Wade") ///
    title("Hazard Functions", position(11)) ///
    scheme(s2mono) graphregion(color(white) icolor(none)) ///
    saving(hazadopt.gph, replace)

graph export hazadopt.png, replace

{