Re: st: How can I count the number of variables that meet a certain condition per observation?

[Nick Cox <njcoxstata@gmail.com>]

In addition to Daniel's reply

gen count = 0
gen sum = 0

forval j = 101/110 {
        replace count = count + ( var'j' > 1 )
         replace sum = sum + var'j' if var'j'  > 1


}

where ' ' are to be understood as left single quote ... right single  
quote.

Nick

On 3 Dec 2011, at 11:49, "Schreck, Philipp" <schreck@bwl.lmu.de> wrote:

> Dear list members,
>
> Say my dataset includes 10 observations and 10 variables, var101- 
> var110. Entries can take values 0, 1, or 2 (I use Stata/SE 11).
>
> I would like to count for each observation the number of variables  
> that meet a certain condition, say var*==2. Although I searched the  
> list, I cannot find the appropriate command and would be very  
> grateful for any suggestions. I could loop over var101-var110 and  
> (for each var) generate a new variable with value "1" if the  
> respective variable is greater than 1 (and then use -rsum- with the  
> new variables). But that seems overly complicated and I was  
> wondering whether there's an easy command I've been missing.
>
> Maybe it could be something like:
>
> egen varcount = rowtotal (var101-var110)
>
> but then I get the sum of all values (across vars) per observation  
> which is not what I want.
>
> Is there a simple way of computing the sum across variables (for  
> each observation) ONLY for those variables with values greater than  
> 1 (or, alternatively, the number of variables with values greater  
> than 1 per observation)?

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/