58% of the group didn't respond. That creates a huge selection bias that doesn't make it worth getting into the study's other flaws.