First, note that only couples were included, so that if particular enneagram types correlate with being single then they would be under-represented in this study.
Secondly, the page doesn't describe how the couples were selected. So it's also possible that self-selection skewed results as well.
Finally, as the page itself points out, there really aren't good validated enneagram instruments. Therefore, the results rely on self-typing which can also be problematic.
Still food for thought, though. The more/less frequent couple combinations are interesting.