The fact that people can’t accurately account for what has influenced their behavior in experiments doesn’t stop them creating reasons of their own that appear, at least superficially, to account for what they’ve done.