The same code for coefplot yields different results

[Honorata Bogusz]

Dear Stata users,

I'm new here and I have a problem which I'm, unfortunately, unable to solve. I'm plotting odds ratios from logistic regression using coefplot (Stata 16.1). Below is my code for the plot:

xi: asdoc logit child1 ln_time i.edu_level i.married i.pgemplst i.cohort  num_s, cluster(pid) nolog or
    
coefplot, title({bf:Transition to first birth}) subtitle(Odds Ratio) eform drop(ln_time _cons) xlabel(0(1)5) xscale(range(0(1)5)) headings(2.edu_level="{it:Education Level}" 2.married="{it:Marrital Status}" 2.pgemplst="{it:Employment Status}" 2.cohort="{it:Cohort}" num_s=" ") mlabposition(3) mlabgap(*10) mlabsize(3) mlabel(cond(@pval<.01, string(@b,"%9.3f") + "***", cond(@pval<.05, string(@b,"%9.3f") + "**", cond(@pval<.10, string(@b,"%9.3f") + "*", string(@b,"%9.3f"))))) msize(large) legend(off) note("*** p<0.01, ** p<0.05, * p<0.10")
graph  export first_birth.png, width(1000) replace

All of the variables are categorical (except for the "number of siblings" which I treat as continuous) and specific values are labeled. The problem is that the first time I generated a plot using this code, I got:


which is the outcome that I want. However, I made a spelling mistake, so I wanted to correct it. After opening my syntax the next day and running THE SAME code I got:



So it appears that coefplot reads only the label for "numer of siblings" and not for the other variables and it also drops headlines.
I have no idea why it is so. Please help me if you have a clue.

Regards,
Honorata

[Honorata Bogusz]

Ok, I know the answer after a lot of trial and error but I do not understand it. For some reason, when I run:

    logit child1 int_1-int_8 cohort_1-pgemplst_8 i.edu_level i.married i.pgemplst i.cohort  num_s, cluster(pid) nolog or

without "xi: asdoc" or "xi:", coefplot displays the right labels. Still, if someone knows the answer why it is so, I'll be happy to hear an explanation.

[William Lisowski]

The xi prefix is outdated syntax. If you type help xi and scroll down a few lines in the output, you will see the following.

    +-----------------------------------------------------------------+
    | Most commands in Stata now allow factor variables; see          |
    | fvvarlist. To determine if a command allows factor variables,   |
    | see the information printed below the options table for the     |
    | command.  If the command allows factor variables, it will say   |
    | something like "indepvars may contain factor variables."        |
    |                                                                 |
    | We recommend that you use factor variables instead of xi if a   |
    | command allows factor variables.                                |
    |                                                                 |
    | We include [R] xi in our documentation so that readers can      |
    | consult it when using a Stata command that does not allow       |
    | factor variables.                                               |
    +-----------------------------------------------------------------+

So your version of Stata is new enough that your i. variables were interpreted as factor variables and the xi: was unneeded.

See help fvvarlist for details.

[Honorata Bogusz]

Thank you! That explains it.