Labeling a single observation on a scatter plot.

[Karishma DSouza]

I looking at the relation between years of education and vaccine skepticism across countries. I want to label just one country, France. But am unable to. I tried labeling all countries on the graph and even that fails. The Country variable is a string variable. I'm not sure what's going on here.

Code:
reg Safety School_Years

gen country1 = "France" if Country=="France"
scatter Safety School_Years || lfit Safety School_Years, mlabel(country1) legend(off) ytitle("% not agreeing that vaccines are safe")

*Labelling all countries - also failing
scatter Safety School_Years || lfit Safety School_Years, mlabel(Country) legend(off) ytitle("% not agreeing that vaccines are safe")

[Nick Cox]

Here is some technique, in the absence of a data example of yours:

. sysuse auto, clear
(1978 Automobile Data)

. sort price

. l make in L

     +--------------+
     | make         |
     |--------------|
 74. | Cad. Seville |
     +--------------+

. scatter price weight , ms(Oh) || scatter price weight in 74, ms(O) mc(red) mla(make) mlabcolor(red) scheme(s1color) legend(off) ytitle(Price (USD)) 

.

So here's a guess at code closer to what you wish

 scatter Safety School_Years || lfit Safety School_Years || scatter Safety School_Years if country == "France", mlab(country)  legend(off) ytitle("% not agreeing that vaccines are safe")

The main problem in #1 is there are no markers to label in lfit. The option belongs on a scatter call, but your || raise a wall that the mla() option can't cross.

[Phil Bromiley]

It is also possible to use the graph edit capability after the graph is created. This provides a great deal of flexibility. You can record and rerun your edits in case you need them on another graph.