I broadly agree but I usually trust one of the big ones (e.g. ‘recovery’ in the UK) since they have pretty big sample sizes and seem to be on top of stats.