Dear All,
I would like to plot the estimated coefficients from several regressions discontinuity (RD).
By coding, for instance,
if I use the coefplot routine I automatically get the plot of e(tau_bc), which constitute the stored bias-corrected local-polynomial RD estimates and my coefficients of interest.
So far so good… My problem is that each RD is run on a different point in time (i.e, year) with respect to a specific moment, acting as a reference time-point.
If I have four RDs, stored as A, B C and E, then there must be a fifth null estimate stored as D (my reference point).
I tried to add a scalar=0 in correspondence to eststo D, but when I rely on coefplot, the D parameter gets the value of the last non-null stored estimate (i.e., the “E”). Down below a simplified version of the code line.
Am I on the correct track? Any more suitable alternatives?
Minor point: the recast(connected) option does not connect the point estimates. Any hint also on this?
Thanks for considering my request
Pasquale
I would like to plot the estimated coefficients from several regressions discontinuity (RD).
By coding, for instance,
Code:
eststo A: rdrobust y1 x1, c(0) eststo B: rdrobust y2 x2, c(0)
So far so good… My problem is that each RD is run on a different point in time (i.e, year) with respect to a specific moment, acting as a reference time-point.
If I have four RDs, stored as A, B C and E, then there must be a fifth null estimate stored as D (my reference point).
I tried to add a scalar=0 in correspondence to eststo D, but when I rely on coefplot, the D parameter gets the value of the last non-null stored estimate (i.e., the “E”). Down below a simplified version of the code line.
Code:
coefplot (A) (B) (C) (D) (E), yline(0) vertical recast(connected)
Minor point: the recast(connected) option does not connect the point estimates. Any hint also on this?
Thanks for considering my request
Pasquale
