Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Significance level stars in Stata tables

    Hello,
    I am currently exporting some regression tables from Stata to LaTeX with the use of the . estout instruction.

    I need pretty detailed Stata tables, so my code is:
    . estout *, style(tex) cells(b(star fmt(%9.3f)) se(par)) stats(r2_a N, fmt(%9.3f %9.0g) labels(R-squared Observation)) legend label collabels(none) varlabels(_cons Constant)

    I would like to modify the default setting for significance stars of the b coefficients with the instruction . starlevels, possibly adding a fourth possibility (****), however after many attempts I can't make it work.
    - Do you know how I may modify my code in order to make the instruction work?
    - Do you know whether Stata automatically changes not only the p instructions at the end of the table, but also the coefficients' stars within the table, which is made of multiple regressions? (actually, the same regression with an increasing number of control variables).

    Thank you!

  • #2
    Hello Ambra.

    Welcome to the Stata Forum / Statalist

    I would like to modify the default setting for significance stars of the b coefficients
    Yes, I gather you can do it. According to the help files,

    starlevels(levelslist) overrides the default thresholds and symbols for "significance stars". For instance, starlevels(+ 0.10 * 0.05) sets the following thresholds: + for p<.10 and * for p<.05. Note that the thresholds must lie in the (0,1] interval and must be specified in descending order. To, for example, denote insignificant results, type starlevels(* 1 "" 0.05).
    Please read the FAQ, particularly on how to share command and output.

    Finally, estout is a user-written program.
    Last edited by Marcos Almeida; 27 Mar 2017, 04:32.
    Best regards,

    Marcos

    Comment


    • #3
      With regards to the "fourth possibility", I mean, to present p-values under stars with 4 levels, I'm not sure whether it is feasible with estout, but you may give it a try, having the commands shared above.

      Anyway, p-values, you know, won't provide the magnitude of the effect. What is more, with the exception of a few fields (suchs as genetics), the journals tend to demand p-values below 0.001 to be presented as "<0.001".
      Best regards,

      Marcos

      Comment


      • #4
        I think you can do this with outreg2. It will let you specify levels and indicators (like stars).

        Comment

        Working...
        X