rowmiss

[Sunniva]

Dear Stata user,
I will create mean scores for a selection of variables, an I need to take into account missing values. The first command works fine, but for the second one “t1_miss_ boundaries” I got the following error message” too many variables specified r(103)”. I can’t understand this since there are the same amount of variables in “t1_miss_control” as in “t1_miss_ boundaries”.

First command ( works fine)
egen t1_miss_control=rowmiss (Q10a Q10b Q10c Q10d Q10e_r Q10f)
egen t1_control=rowmean (Q10a Q10b Q10c Q10d Q10e_r Q10f)if t1_miss_control<4

Second command ( error message “” too many variables specified r(103)”.)
egen t1_miss_ boundaries=rowmiss (Q11a Q11b Q11c Q11d Q11e Q11f)
egen t1_ boundaries=rowmean (Q11a Q11b Q11c Q11d Q11e Q11f)if t1_miss_ boundaries<4
Can someone see what’s wrong in the second command?
Many thanks

[Nick Cox]

In practice spaces before the function name are indulged as is the lack of a space before an if qualifier. I can't see a problem that would be fatal.

But occasionally strange characters can be problematic particularly if you copied and pasted from somewhere else in the first instance.

Recently I copied and pasted code from one of my own documents and was frustrated at its not working until close scrutiny revealed en-dashes in Stata where minus signs were needed. Oddly enough in the original document minus signs were explicitly chosen....

If I am right then naturally the error message is misleading. But as you say on the face of it the error message makes no sense.

[Sunniva]

Many thanks. I will look more into this.