> Getting Hugh quality data That's why we need large interventional studies instead of observational ones. Observational studies are interesting and sometimes

> Getting Hugh quality data That's why we need large interventional studies instead of observational ones. Observational studies are interesting and sometimes necessary (e.g. when there's ethical concerns about running an intervention such as with smoking), but it's just way too tricky to separate out the impact of confounders especially when the treatment effects are rather weak (e.g. with coffee or meat consumption).