"unless the laboratory's legal and regulatory [...]"
That [...] could hide a lot of shady stuff being done via NSLs (etc.).
EDIT: There is a very interesting issue here, though, namely how the findings by 23andme are presented to their customers. There's good research that shows that presenting relative probabilities[2] (as opposed to just picking a sample size and doing everything in numbers relative to that) is very hard to understand for the general public (and even for statisticians unless they're paying close attention!). The Base Rate Fallacy is basically a consequence of presentation. Hopefully, 23andme are doing this responsibly, but I honestly don't know.
[2]: Example: "Eating X increases risk of cancer by 50%". Well, yeah, that might change my risk of cancer from 0.01% to 0.015%, but that that's not something I should worry about. Yet, we see these headlines because they grab people's attention.
That [...] could hide a lot of shady stuff being done via NSLs (etc.).
EDIT: There is a very interesting issue here, though, namely how the findings by 23andme are presented to their customers. There's good research that shows that presenting relative probabilities[2] (as opposed to just picking a sample size and doing everything in numbers relative to that) is very hard to understand for the general public (and even for statisticians unless they're paying close attention!). The Base Rate Fallacy is basically a consequence of presentation. Hopefully, 23andme are doing this responsibly, but I honestly don't know.
[2]: Example: "Eating X increases risk of cancer by 50%". Well, yeah, that might change my risk of cancer from 0.01% to 0.015%, but that that's not something I should worry about. Yet, we see these headlines because they grab people's attention.