Re: st: most requested *simple* features to help SPSS users transition

[Nick Cox <njcoxstata@gmail.com>]

On (1), I think this has been mentioned at recent users' meetings and
I imagine wildly that StataCorp take it seriously.

On (2), see -tabchi- and -tabchii- from -tab_chi- (SSC) for a partial answer.

On (3), see -egen-'s -group()- function with -label()- option and

SJ-7-4  dm0034  . . . Stata tip 52: Generating composite categorical variables
        . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  N. J. Cox
        Q4/07   SJ 7(4):582--583                                 (no commands)
        tip on how to generate categorical variables using
        tostring and egen, group()

It would be easy to code up a friendly wrapper for this. Note that it
works just to clean up a single variable.

. sysuse auto, clear

. program nicecode
  1.         version 8
  2.         syntax varlist [if] [in], Generate(str) [Missing]
  3.         egen `generate' = group(`varlist'), `missing' label
  4. end

.
end of do-file

. nicecode foreign rep78, gen(mycatvar)
(5 missing values generated)

. tab mycatvar

group(forei |
  gn rep78) |      Freq.     Percent        Cum.
------------+-----------------------------------
 Domestic 1 |          2        2.90        2.90
 Domestic 2 |          8       11.59       14.49
 Domestic 3 |         27       39.13       53.62
 Domestic 4 |          9       13.04       66.67
 Domestic 5 |          2        2.90       69.57
  Foreign 3 |          3        4.35       73.91
  Foreign 4 |          9       13.04       86.96
  Foreign 5 |          9       13.04      100.00
------------+-----------------------------------
      Total |         69      100.00




Nick

On Fri, Oct 19, 2012 at 3:43 PM, JVerkuilen (Gmail)
<jvverkuilen@gmail.com> wrote:
> Having decided to use Stata instructionally I run into a few little
> things that would be nice and, to my knowledge, don't exist. Slowly
> but surely the students here are switching over as they realize they
> need to learn things like propensity scoring. In all honesty I don't
> care what they do when they grow up, but it greatly simplifies my life
> not to have to deal with SPSS. However there are some little things
> that turn out to be resistance-causing on the part of folks used to
> it:
>
> (1) A simple way to clear the Results window. -clear- nukes the data
> in memory. An easy workaround is simply to hold the return key down
> for a goodly bit of time, but a way to get a clean results window
> would be very nice.
>
> (2) -tabulate- could use a few more options. For instance, it doesn't
> generate cell residuals or adjusted residuals, or compute odds ratios.
> -cc- runs the latter but other people use odds ratios and
> Mantel-Haenszel tests too and won't necessarily think to look under
> Tables for Epidemiologists. ;)
>
> (3) SPSS has one "killer app": Data management in it is quite nice.
> For instance, "automatic recode" is monumentally convenient. This
> takes a variable, automatically creates a new one with sequential
> numbers and labels it with the values of the old variables. The
> variable can be big 'n nasty, like a string with hundreds of unique
> entries. The ability to do things like paste labels and so on is also
> handy, though it may go against the general philosophy of careful
> tracing of all changes to data.
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/