Results that cannot be explained with some valid theoretical argument might be a statistical artifact, and thus are not reported?