The first article isn't valid as it's paywalled and the abstract doesn't mention any studies; it could be someone's well-educated opinion. The second article c

The first article isn't valid as it's paywalled and the abstract doesn't mention any studies; it could be someone's well-educated opinion. The second article covers how bright light can be without being painful, and doesn't mention sleep or melatonin. Or at least the abstract does, again the article is paywalled, but at least it involves 500 participants. The third study only covers 12 participants. I'll grant there may be thousands of these studies, each with a handful of subjects. Is there some statistical way to stitch the results together in any meaningful fashion - or would that just result in a low p-value?