COEFPLOT, MLABEL(): Line break appears on PNG but not PDF

[Masaru Nagashima]

Hello all,
I am using COEFPLOT and trying to add a line break to the marker label. Specifically, at each marker, I am adding a text like `coef (se)' with a line break between coef and (se). It works great on the graph window and also in the exported PNG, but somehow the line break disappears in the exported PDF. The following code replicates the issue (at least in my environment). Any idea how to make the line break work in the PDF just like it does in PNG?

. sysuse auto
(1978 Automobile Data)

. reg price

      Source |       SS           df       MS      Number of obs   =        74
-------------+----------------------------------   F(0, 73)        =      0.00
       Model |           0         0           .   Prob > F        =         .
    Residual |   635065396        73  8699525.97   R-squared       =    0.0000
-------------+----------------------------------   Adj R-squared   =    0.0000
       Total |   635065396        73  8699525.97   Root MSE        =    2949.5

------------------------------------------------------------------------------
       price |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
       _cons |   6165.257   342.8719    17.98   0.000     5481.914      6848.6
------------------------------------------------------------------------------

. coefplot, vert mlabel(string(@b,"%9.2f") + char(13) + char(10) + "(" + string(@se,"%9.2f") + ")")

. gr export "reg_price.png", replace
(file reg_price.png written in PNG format)

. gr export "reg_price.pdf", replace
(file reg_price.pdf written in PDF format)

reg_price.pdf

[Masaru Nagashima]

Any thoughts, anyone...? BTW sorry for the pdf link that doesn't appear to work. I hope everyone can replicate the issue.

[Andrew Musau]

I think you will want to convert from ps to PDF (or image to PDF). For a converter, see https://web.mit.edu/ghostscript/www/Ps2pdf.htm. There does not appear to be much evidence that anyone knew how to include a line break in a marker label prior to your post. See, e.g., https://www.statalist.org/forums/for...ariable-labels.

[Andrew Musau]

As you want your standard errors directly underneath your coefficients, a coefplot (SSC) workaround is possible. Here is an example:

sysuse auto, clear
reg price weight mpg turn, nocons
est sto D
coefplot (D, mlabel(string(@b,"%9.2f")) mlabc(blue) mlabpos(3)) ///
(D, mlabel("(" +string(@se,"%9.2f") + ")") mlabc(blue) mlabpos(5) ///
mlabgap(0.05in) offset(0)), vert nokey ciopts(lc(blue)) mc(blue) scheme(s1color)

This should export fine in PDF format.

[Masaru Nagashima]

Dear Andrew Musau,
Thank you for your posts. The second one, the COEFPLOT workaround, appears pretty clever. Basically you are plotting markers and their labels twice, first the coefficient values and second parentheses and the standard errors, with a slight adjustment in their positions and colours, right? I have slightly modified your code to my match taste, but this is awesome. Thank you so much!

. sysuse auto
(1978 Automobile Data)

. reg price

      Source |       SS           df       MS      Number of obs   =        74
-------------+----------------------------------   F(0, 73)        =      0.00
       Model |           0         0           .   Prob > F        =         .
    Residual |   635065396        73  8699525.97   R-squared       =    0.0000
-------------+----------------------------------   Adj R-squared   =    0.0000
       Total |   635065396        73  8699525.97   Root MSE        =    2949.5

------------------------------------------------------------------------------
       price |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
       _cons |   6165.257   342.8719    17.98   0.000     5481.914      6848.6
------------------------------------------------------------------------------

. coefplot (, mlabel(string(@b,"%9.2f")) mlabpos(3) mc(navy)) (, m(none) mlabel("(" + string(@se,"%9.2f") + ")") mlabc(navy) mlabpos(5) offset(0) noci), vert nokey

. gr export reg_price_update.png, replace
(file reg_price_update.png written in PNG format)

. gr export reg_price_update.pdf, replace
(file reg_price_update.pdf written in PDF format)

.

reg_price_update.pdf

[Andrew Musau]

Basically you are plotting markers and their labels twice, first the coefficient values and second parentheses and the standard errors, with a slight adjustment in their positions and colours, right?

Yes. Adding the option -noci- to the second plot as you do is a better approach as you only need to match the color of the labels.