New on SSC: markobs

[daniel klein]

Thanks to Kit Baum, a new command, markobs, is now available from the SSC.

markobs is intended as a safe wrapper for mark and markout, helping to avoid accidential overwriting of variables. markobs originated from a discussion with Dirk Enzmann, who also reviewed the code and help file.

Here's a toy example to demonstrate the problem

. sysuse auto
(1978 automobile data)

. tabulate foreign

 Car origin |      Freq.     Percent        Cum.
------------+-----------------------------------
   Domestic |         52       70.27       70.27
    Foreign |         22       29.73      100.00
------------+-----------------------------------
      Total |         74      100.00

. tabulate rep78

     Repair |
record 1978 |      Freq.     Percent        Cum.
------------+-----------------------------------
          1 |          2        2.90        2.90
          2 |          8       11.59       14.49
          3 |         30       43.48       57.97
          4 |         18       26.09       84.06
          5 |         11       15.94      100.00
------------+-----------------------------------
      Total |         69      100.00

. tempvar touse

. mark `touse'

. markout `tose' foreign rep78 // <- misspelled `touse'

. count if `touse' // <- want 69
  74

. tabulate foreign

 Car origin |      Freq.     Percent        Cum.
------------+-----------------------------------
   Domestic |         53       71.62       71.62
    Foreign |         21       28.38      100.00
------------+-----------------------------------
      Total |         74      100.00

Above, we load auto.dta and look at the distribution of foreign and rep78. We then create a temporary variable, touse (read: to-use), to indicate the sample we wish to use in subsequent code. We intend to set the marker variable to 0 for observations with missing values in foreign or rep78. Unfortunately, we misspell touse (omitting the u in tose). Even worse, markout doesn't complain. We're left with a sample of all 74 observations, whereas we wanted 69 observations because rep78 has 5 missing values, and -- much worse -- markout has changed the (byte) variable foreign!

Here's how markobs solves this problem:

. sysuse auto
(1978 automobile data)

. tempvar touse

. markobs `touse'

. markobs `tose' foreign rep78 // <- misspelled `touse'
foreign not created by markobs
r(498);

Naturally, the 0/1 to-use variable created by markobs can be used with markin and svymarkout.