Actually, the phrase you're looking for is correlation, not coincidence. While I agree that, on the individual level, there needs to be more than just correlation (accounting for pre- and post-spanking IQ development with a large sample size), the comparison between nations is more difficult. It's harder to think of alternative explanations that account for entire nations rather than individuals.
In addition, that's just the epidemiological study. The non-statistical study of actual individuals that determined that they had some form of PTSD doesn't have an alternative explanation and PTSD is associated with reduced IQ.