I have written two articles (see here and here) that contest the widely reported notion that results in recent congressional primaries provide evidence of a raging anti-incumbent fervor. When incumbents have lost (and not many have faced defeat), other factors -- like ideology -- explain the results. Despite this fact, media commentators continue to describe election results through the problematic lens of anti-incumbent fervor.
Larry Sabato, an esteemed Professor of Political Science at the University of Virginia, has conducted some empirical research on the subject that debunks the anti-incumbent theme. Sabato observes that only four congressional incumbents and one governor have actually lost their election bids this year. By contrast, 200 incumbents have won their primary contests, and many of these individuals ran unopposed or had comfortable leads throughout their campaigns.
Furthermore, over the last 40 years, an average of 6 to 7 incumbents have lost primary contests per election year. This year's numbers are in-line with that trend (or even show a stronger performance by incumbents). Sabato also agrees that ideological differences, along with corruption and party-switching, explain the defeat of incumbent candidates.
This analysis does not mean that incumbents will not face acute vulnerability in November. Instead, it strongly suggests that media reports generally linking election results with anti-incumbent fervor are grossly inaccurate.