To Harry Enten, who teaches WaPo's Richard Cohen a bit about Republican primary and caucus electorates. Uh, yeah, New Hampshire Republicans are sharply less, not more, conservative than typical GOP primary-goers.
I do wonder about Iowa, however. As Enten details it, caucus attendees are quite a bit more likely to call themselves "very conservative" than primary voters in Ohio. I should know more about this than I do, but presumably some of that is about caucuses compared with primaries (caucuses get only the most intense voters, who are most likely not only to be strongly ideological but also more likely to think of themselves that way). The part I'm curious about is whether Iowa's first-in-the-nation status also plays in, with Iowa's voters getting the most partisan "education" of any.
At any rate: the more important point here is that everyone tends to overestimate the effects of going first. Yes, Iowa and New Hampshire do still perform some winnowing, but for the most part they simply implement choices made by the party as a whole. Candidates who are rejected by party actors nationally but take advantage of favorable local circumstances to run well in these early contests -- and Rick Santorum is a fine example -- generally tend to fizzle.
This also is the legitimate basis for the much derided "expectations" game. Some of that derision is deserved, but some of it is simply noting that context matters: a conservative candidate doing well in moderate New Hampshire should be treated as having accomplished more (all else equal) than that same candidate doing well in Iowa or South Carolina. And smart party actors react accordingly.
In other words, the whole complaint about Iowa/New Hampshire has been massively overblown, at least since the mid-1980s when the parties learned how to deal with the reformed process.
Also: nice catch!