Maarten L. Buis

• Overview
• Curriculum Vitae
• Publications
• Dissertation
• Working papers
• Presentations
• Workshops
• ⇒ Software
• Stata example FAQ
• Links

postrcspline

Author: Maarten L. Buis

The postrcspline package consists of programs that can help with the interpretation of a model that uses a restricted cubic spline as one of the explanatory variables by displaying a graph of the predicted values against the spline variable adjusted for the other covariates, or the marginal effects of the spline variable.

This package can be installed by typing in Stata: ssc install postrcspline

Supporting materials

• summary on ssc.
• helpfiles:
  • helpfile postrcspline,
  • helpfile adjustrcspline,
  • helpfile mfxrcspline, and
  • helpfile mkspline2
• Presentation at the 2009 German Stata Users' Group meeting on using the postrcspline package
• Announcement of postrcspline on statalist

Examples

The most basic syntax and the resulting graphs:

. sysuse uslifeexp, clear
(U.S. life expectancy, 1900-1999)

. mkspline2 sp_year = year, cubic nknots(3)

. reg le sp_year* if year != 1918

      Source |       SS       df       MS              Number of obs =      99
-------------+------------------------------           F(  2,    96) = 2126.16
       Model |  7466.92497     2  3733.46249           Prob > F      =  0.0000
    Residual |  168.572512    96  1.75596367           R-squared     =  0.9779
-------------+------------------------------           Adj R-squared =  0.9775
       Total |  7635.49749    98  77.9132396           Root MSE      =  1.3251

------------------------------------------------------------------------------
          le |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
    sp_year1 |   .4347183   .0125385    34.67   0.000     .4098296     .459607
    sp_year2 |  -.1845163    .015519   -11.89   0.000    -.2153213   -.1537113
       _cons |  -778.6582   24.14547   -32.25   0.000    -826.5865   -730.7298
------------------------------------------------------------------------------

. adjustrcspline

. mfxrcspline

. 
. 
[do-file]

Adding extra options to make the graph look pretty

. sysuse uslifeexp, clear
(U.S. life expectancy, 1900-1999)

. mkspline2 sp_year = year, cubic nknots(3)

. qui reg le sp_year* if year != 1918

. adjustrcspline,                                            ///
>     addplot(scatter le year if year != 1918, msymbol(Oh)|| ///
>             scatter le year if year == 1918, msymbol(X))   ///
>     name(pred, replace)                                    ///
>     title("expected life expectancy in the US")

. 
. mfxrcspline,                                               ///
>     name(mfx, replace)                                     ///
>     ylab(0(.1).5)                                          ///
>     title("trend in life expectancy in the US")

. 
. graph combine pred mfx,                                    ///
>     ysize(3)                                               ///
>     note("excluding 1918 due to the Spanish flu pandemic")

[do-file]