I mean... there are basically no robustness checks.They only had a pre- and post-intervention comparisons with no random comparison group for a counterfactual.

I mean... there are basically no robustness checks.They only had a pre- and post-intervention comparisons with no random comparison group for a counterfactual. They asked retrospective questions. There are multiple poorly documented interventions. The analysis of the non-positive responses is extremely limited and what does exist draws conclusions that reek of confirmation bias. Their response rate for surveys appears to be crap which brings up questions of sampling bias. There is a strong appearance of cherry-picking. The descriptive statistics are extremely lacking - no confidence intervals, standard deviation, etc. - giving an appearance of p-hacking.