I began to do the paired t-tests and realized that when I converted my categorical data into numerical data, SPSS put the number 1 in places where subjects did not respond to certain questions as subjects were not instructed to answer each section. Thus, I have mostly 1s as responses for each section. However, these 1s are being accounted for in the means for my t-tests, and are interfering with my averages. I have been searching for ways to fix this, but none seem to match my situation. I know how to do tests with missing data, but the data has already been recoded and I do not want to retype all of the responses into another sheet. Is there a way for SPSS to recode categorical data into numerical data without putting numbers in for missing variables?