I have submitted an abstract to a conference that has been accepted and will be published for the conference. There is a minor error that changes the data I have included in the abstract. What should I do?
There is probably nothing to be done at the moment, though publishing a correction on your personal web site is appropriate. Even withdrawing the abstract is probably impossible now.
Make the correction in the final paper and, when you cite the abstract (as you should), note that it contains an error corrected here.
Lots of incorrect stuff gets published for various reasons. But we hope that statistically we wiggle toward the truth.