Dear all, i have a problem which deals with indexing again, i guess. I am a beginner and though i read the S Poetry stuff it is not that easy for me to use the examples for my problems, maybe all this needs some practice... Here?s the question. As i want to perform a factor analysis i need the same number of rows for every column to get a matrix that i use as an argument of factanal. I?d like to find out which of 5 question i asked explain the variance in my dataset. People were able to choose "dont know". Of course someone who didn?t know, could know the answers for the other question. To put in a nutshell, only those results are valid for my factanal in which every question is answered (which is ok, as it is that way most of the time). i have matrices question1, question2 etc which contain the answers to each question. Every person has a result id. There fore my idea is to put all the matrices into one big matrices and then just take those of out the matrices, where the resultid occurs 5 times - as i have 5 questions. And here are my two questions : a) is there some syntax that helps me with: if number of any resultid appearance in column 1 is is smaller than 5, dont put it into my new matrix ? b) is the idea in general wrong ? is there an easier way to get the matrices into one.. prepared for a factanal ? thx in advance! matthias