I built a Federal Reserve System exercise from the Sudrvey of Household Economic Decision Making data set--300 + variables and 11K cases. The SAS (and STATA) varuable names represented in the codebook and in the data set are four characters long, and a only a bit mnemonic. The variable labels are a sentence derived from the question. Importing the table from SAS to JMP using variable laels would render output unprintable because of the length of the variable labels. Using the table with the original namee, however, is a far reach for students--i reqires reopening the codebook or the Column column to see the variable label. Then the variable label must somehow be handled by hand in the final analysis output, perhaps by dumping it to Excel and then to introudce meaningful column names.
Proposals to hand edit the column characteriistic page to paste the variable label into the variable name block, with whatever kind of editing are pushing the limits for undergraduate work expectations. Writing or even copying and executing JSL code is beyond the reasonable expectations for the class.
Is there a cleaner way to represent the output rows and columns with some version of the variable label, without requiring JSL or cutting and pasting. The other option, reimport the file fro the SAS file, using the column labels as column names coming into JMP would still produce a file with 30 character column names.
