This analysis, passed along by GWU's John Sides, confirms what we've been hearing anecdotally: Early voting in Texas does not bode well for Hillary Clinton.
When you had only two media (TV and radio) capable of instantly broadcasting information, like an exit poll, you could install gatekeepers. And in fact, once the networks realized they could be impacting elections, they made it a policy not to release the information until voting was concluded. Now we have stuff like this, somebody's speculation about some data a week before the actual vote, and -- thanks to the internet -- it's out there just as if Walter Cronkite had announced it on the evening news. Maybe not to as many people, but that's coming, too. How do you install gatekeepers among the entire population of computer users? You can't. Obviously, you're going to have to tighten controls over the gatherers of this information, if you decide to do something about it at all -- which I'm in favor of. Let people vote, then count the votes. This stuff -- predicting an electoral outcome using someone's rough analysis of a rough collection of data -- is absurd.
I don't understand why nobody in the print world had thought about this and formulated some kind of policy. Or maybe they did, and this was the result of their decision. Either way, it's the sort of thing that deserves public debate, and I ain't seen any public debate about it.
*Addendum. Now that I think about it, I think it was people in West Coast states complaining about exit polls cheapening the value of their votes that made the networks adopt this policy. So another way of looking at it is a market-based