> Editors, reporters and producers at GBH News sent nearly 100 letters from different places in the metro area at various hours on the same day to correspondent

> Editors, reporters and producers at GBH News sent nearly 100 letters from different places in the metro area at various hours on the same day to correspondents of their own choosing in 38 states, creating a random sample. That's not a representative random sample of the mail the post office delivers. It likely oversamples from long tail requests: reporters are likely to choose zanier locations, and even the fact that they're sending the mail (instead of some kind of an institution) makes for a weird sample. Now, I don't particularly love that the modal mailing is a piece of junk, or that I have to rely on mail to receive certain kinds of institutional communications, but there's no particular reason to think the experiment revealed a failure to meet their SLO.