art photo

DSDR KB

Q:  

I am working with NELS data (base year, first follow-up and second follow-up). When I tried to extract the variables from raw data using the SAS setup cards provided by ICPSR, I do not get the right number of cases – way too few.

A:  

The problem is in the setup files; these are sometimes dated and no longer work. The specific issue in your case is the setup file does not have a an LRECL statement.

SAS has a default of 256 columns of data that it will read. If the record is longer than that and there is no LRECL statement, SAS will go to the next record to continue reading in the variables. You end up with garbage for your variables and as you have noticed, you end up with too few cases.

The following is an example of where to put the LRECL statement:

http://www.psc.isr.umich.edu/dis/data/prgmlib/sas/lrecl.html

Annotated Resources:

Direct Links:

Related Question Groups: