I have done linear regression in SPSS several times before and never run into this problem. I recoded some variables into dummies and was careful to ensure there was no "overlap". An example of a variable I recoded is "marital status". It was: married = 1, cohabit = 2, single = 3. I recoded this into 2 vars. Now I get the following warning when I try to enter them as i.d. in linear regression: "variables were deleted from analysis due to being constants or having missing correlations." I have recoded these variables at least 5 times to ensure I didn't mess up in coding. I still get the warning. What could be causing this? HELP!!!