It's fascinating data, and I'm glad it's being taken, but...I am a bit skeptical that we can draw any reasonable conclusions from it without limiting the data to long-term users.

It seems implausible that there has not been a significant change in the health-consciousness of the typical purchaser in the last month or two, vs. prior to that. I can't tell you what the change would be, but I find it highly implausible that it would be not at all different, given the biggest spike in public awareness of a health issue in several decades.

It's nice to see that Dalziel has thought of the issue, but the article doesn't tell us much about _how_ he accounted for that. I could think of ways, but I could also think of ways to screw it up, and it doesn't say in this article how he "ensured" that the effect is real.