Just to clarify: That 6.2M projection is as of TODAY. It could change tomorrow.
I very much appreciate the shout-out from the Washington Post earlier today, but I need to clarify something: My projection of 6.2 million exchange-based QHP enrollments as of 3/31 will likely change as new data comes in.
Example: If California experiences a massive 5-day outage like they did in February (which dropped their daily average, which had been 7,200/day the first half of the month, down to only 5,000/month overall), that will drop the needle dramatically. This isn't just idle speculation--yesterday Healthcare.gov apparently was offline for half an hour for some reason or another. Things happen, though obviously I hope everything will run smoothly throughout the next two weeks. Conversely, if CA announces an astonishing 30K/day day like they did back on December 23rd, that will move the projection up higher.
While I would never compare myself to Nate Silver, I do need to point out that his estimates of various candidates & parties chances changed from day to day as new reliable polling data came in. That's how this stuff works; as more real data comes in, you add it to your existing data, see how it impacts your calculations, and alter your projections accordingly.
My job is far easier than his, because I don't have to worry about whether one of the exchanges is "lying" or "skewing their numbers" (at least I certainly hope not). Aside from the "paid/unpaid" issue, which is hardly worth worrying about just yet, I'm not too concerned about whether the enrollment numbers released are legitimate or not. He had to constantly figure out which polling outfits had good methodology and which ones were full of hooey. I just have to worry about the occasional clerical error, misunderstanding by a journalist or the like, which has been pretty rare (although there have been a few cases, most notably the Minnesota "households vs. individuals" incident in December and the New York Child Health Plus confusion a week or so ago).
I'm no expert on Gov. Romney's "Unskewed Polls" debacle, but I would imagine that part of the problem is that his team probably latched onto one or two decent-looking polls (whether from legitimate sources or not) and ignored any data which interfered with that perception. That's not how it works. Right now, it looks like QHPs will hit around 6.2M. Tomorrow I could get new data which changes that to 6.1M, or to 6.5M. The next day it might jump around some more, and perhaps some existing data will turn out to be a typo and have to be adjusted accordingly. That would stink, but that's the nature of these things.