Coefplot for coefficient of various estimates

[Tariq Abdullah]

The following code is supposed to create a graph ( like the attached one here ) but unfortunately, I'm getting an error saying - invalid 'size'
r(198);

any help or guidance from anyone

*******COEFPLT Dr. DUBE 
global SDUMMY     i.statefip i.year
eststo clear
reghdfe ln_weekearn  ///
cdme ///
hks sks /// 
aj lks  /// 
rising  $SDUMMY [aw=earnwt] , absorb(i.statefip) vce(cluster statefip) 
eststo e1 

coefplot (e1 , mcolor(red%50) msymbol(O) msize(large) ciopts(recast(rcap)) color(red%50)) keep(cdme)) ///
(e1 , mcolor(green%50) msymbol(S) msize(large) keep(hks sks) ciopts(recast(rcap)) color(green%50)))  ///
(e1 , mcolor(purple%50) msymbol(T) msize(large) keep(aj)  ciopts(recast(rcap)) color(purple%50))) ///
(auto_dube , mcolor(blue%50) msymbol(D) msize(large) keep(rising)  ciopts(recast(rcap)) color(blue%50))) , ///
scheme(plotplain) xtitle("COEFOLOT"), size(large) /// 
xline(0, lcolor(gs7) lpat(solid) legend(off) grid(none)) ///
yline(3 10 18, lcolor(gs7%40)) lapt(dot)) ///
groups(cdme ="{bf:Local}") ///
hks sks ="{bf:Spectrum}") ///
aj lks  = "{bf:Industry}") /// 
rising = "{bf:demand}") 
gr export "coefplot.pdf", replace

 * Example generated by -dataex-. For more info, type help dataex
clear
input float(ln_weekearn cdme hks sks aj lks rising) double earnwt byte statefip int year
       . -.1 0 1 1 0 0          0 1 1996
       . -.2 0 1 1 0 0          0 1 1996
       . -.3 0 1 1 0 0          0 1 1996
6.520998 -.3 0 1 1 0 0 11868.2414 8 2002
       . -.1 0 0 1 0 0          0 4 1998
       . -.2 0 0 1 0 0          0 4 1998
       . -.3 0 0 1 0 0          0 4 1998
 7.23574 -.3 0 0 1 0 0 10622.8448 8 2002
       . -.1 0 0 0 0 0          0 6 1999
       . -.2 0 0 0 0 0          0 6 1999
       . -.3 0 0 0 0 0          0 6 1999
       . -.3 0 0 0 0 0  9430.3123 6 2004
       . -.1 0 0 0 0 0          0 1 1996
       . -.2 0 0 0 0 0          0 1 1996
       . -.3 0 0 0 0 0          0 1 1996
       . -.3 0 0 0 0 0 10478.1475 6 2004
       . -.2 0 0 1 0 0          0 6 1997
       . -.3 0 0 1 0 0          0 6 1997
       . -.3 0 0 1 0 0          0 6 1997
5.990339 -.2 0 1 1 0 0 13736.5457 4 2008
       . -.2 0 0 1 0 0          0 1 1996
       . -.3 0 0 1 0 0          0 1 1996
       . -.3 0 0 1 0 0          0 1 1996
 6.19718 -.2 0 0 1 0 0  9774.6278 4 2008
       . -.2 0 0 0 0 0          0 8 1999
       . -.3 0 0 0 0 0          0 8 1999
       . -.3 0 0 0 0 0          0 8 1999
       . -.2 0 0 0 0 0  9183.4515 2 2010
       . -.2 0 0 1 0 0          0 1 1996
       . -.3 0 0 1 0 0          0 1 1996
       . -.3 0 0 1 0 0          0 1 1996
 6.19718 -.2 0 0 1 0 0 11644.6988 2 2010
       . -.3 0 0 0 0 0          0 1 1996
       . -.3 0 0 0 0 0          0 1 1996
6.846935 -.2 0 0 0 0 0 11644.6988 4 2012
       . -.2 0 1 1 0 0          0 1 1996
       . -.3 0 1 1 0 0          0 1 1996
       . -.3 0 1 1 0 0          0 1 1996
6.873564 -.2 0 1 1 0 0 11005.0987 4 2012
       . -.2 0 0 0 0 0          0 1 1996
       . -.3 0 0 0 0 0          0 1 1996
       . -.3 0 0 0 0 0          0 1 1996
       . -.2 0 0 0 0 0   10399.04 1 1996
       . -.3 0 0 0 0 0          0 1 1996
       . -.3 0 0 0 0 0          0 1 1996
       . -.2 0 0 0 0 0          0 1 1996
       . -.2 0 0 0 0 0 11189.5283 1 1996
       . -.3 0 0 1 1 0          0 8 1996
       . -.3 0 0 0 0 0          0 8 1996
       . -.2 0 0 0 0 0          0 8 1996
       . -.2 0 1 1 0 0  14571.094 1 1996
       . -.3 0 0 1 1 0          0 1 1996
       . -.3 0 0 1 1 0          0 1 1996
       . -.2 0 0 1 1 0          0 1 1996
5.943821 -.2 0 0 1 1 0 14568.3216 6 1998
       . -.3 0 0 1 0 0          0 1 1996
       . -.3 0 0 1 0 0          0 1 1996
       . -.2 0 0 1 0 0          0 1 1996
5.999409 -.2 0 0 1 0 0 10463.1051 8 1999

[Tariq Abdullah]

This is the picture of the original code ( which produced the neat-looking graph above ) I'm trying to replicate from Dr. Arindrajit Dube.

[Andrew Musau]

coefplot is from the Stata Journal/ SSC, as you are asked to explain (FAQ Advice #12).

coefplot (e1 , mcolor(red%50) msymbol(O) msize(large) ciopts(recast(rcap)) color(red%50)) keep(cdme)) ///
(e1 , mcolor(green%50) msymbol(S) msize(large) keep(hks sks) ciopts(recast(rcap)) color(green%50))) ///
(e1 , mcolor(purple%50) msymbol(T) msize(large) keep(aj) ciopts(recast(rcap)) color(purple%50))) ///
(auto_dube , mcolor(blue%50) msymbol(D) msize(large) keep(rising) ciopts(recast(rcap)) color(blue%50))) , ///
scheme(plotplain) xtitle("COEFOLOT"), size(large) ///
xline(0, lcolor(gs7) lpat(solid) legend(off) grid(none)) ///
yline(3 10 18, lcolor(gs7%40)) lapt(dot)) ///
groups(cdme ="{bf:Local}") ///
hks sks ="{bf:Spectrum}") ///
aj lks = "{bf:Industry}") ///
rising = "{bf:demand}")
gr export "coefplot.pdf", replace

The highlighted may be the source of your error. You already are specifying options then you add a comma and another option. The fix may be as simple as removing the comma, but the general advice for debugging graph errors is to start small and then gradually build up.

[Tariq Abdullah]

That was very helpful and insightful advice for a beginner coder like me. Actually, following your suggestion, I started from scratch by doing little by little, and eventually I came with a much better result than I got in the beginning. Yet to put the finishing touch but listening to you I'm certainly gonna reach there. Again appreciate your kind advice so so much!

[Matthew Williams]

That was very helpful and insightful advice for a beginner coder like me. Actually, following your suggestion, I started from scratch by doing little by little, and eventually I came with a much better result than I got in the beginning. Yet to put the finishing touch but listening to you I'm certainly gonna reach there. Again appreciate your kind advice so so much!

Hi Tariq,

May I know where to get the replication package of Dr. Arindrajit Dube?

Thank you.

[Tariq Abdullah]

Hello Matthew,

This is the coefplot package with the help of this trpe of graph can be produced. As Mr. Musau kindly suggested - it would be better if you can start from the scratch by taking little steps then that would be helpful to build up to the final stage. And, to. produce the exact graph if you follow the thread I attached the image of Dr. Arindrajit Dube's code too - though I haven't yet been able to replicate the exact command of Dr. dube, but I've been successful to repicate a portion of it. The coefplot was contributed to the economics community by Dr. Ben jann. Hope this helps and if you can replicate the code of Dr. dube kindly give me a knock ! would appreciate that .


http://repec.sowi.unibe.ch/varia/Ode...dense-2018.pdf

[Matthew Williams]

Hello Matthew,

This is the coefplot package with the help of this trpe of graph can be produced. As Mr. Musau kindly suggested - it would be better if you can start from the scratch by taking little steps then that would be helpful to build up to the final stage. And, to. produce the exact graph if you follow the thread I attached the image of Dr. Arindrajit Dube's code too - though I haven't yet been able to replicate the exact command of Dr. dube, but I've been successful to repicate a portion of it. The coefplot was contributed to the economics community by Dr. Ben jann. Hope this helps and if you can replicate the code of Dr. dube kindly give me a knock ! would appreciate that .


http://repec.sowi.unibe.ch/varia/Ode...dense-2018.pdf

Hi Tariq,

Thank you for your reply and helpful suggestions. I appreciate that.

However, may I know where did you get the code of Dr. Arindrajit Dube? Or from which paper of him (you took a screenshot of his code, so I guess you may know the source)?

[Tariq Abdullah]

Hi Matthew
If you don't mind may I have your email id ? I'd prefer giving you the link over other platform if you don't mind.

[Matthew Williams]

Hi Matthew
If you don't mind may I have your email id ? I'd prefer giving you the link over other platform if you don't mind.

Thank you for the kind support. I sent you a private message, so please check your inbox.