There are various software tools available that will let you determine the "grade level" and/or complexity of a particular piece of text. The Sunlight Foundation ran a couple of decades' worth of the Congressional Record through one of them, and the results showed a fairly dramatic drop in the effective grade level of speeches given by members of both parties since 2005:
That's one way to interpret the results, anyway — that Congress has dropped from high-school-junior-equivalents to high-school sophomores over the last six years or so. This sort of interpretation does provide plenty of amusement, such as (if you're a Democrat) being able to point out that your party is apparently half a grade smarter on average, or (if you're a Republican) pointing to other numbers suggesting that the more liberal someone is, the lower the grade level.
The problem with those particular arguments, at least, is that they're probably backwards.
Writing (or speaking) at a higher grade level is not a good thing, or at least not necessarily. What these particular numbers really measure (at least the Flesch-Kincaid test) is the complexity and length of sentences. It says nothing about how accurate or intelligent the sentences are, and all else being equal, the shorter and simpler something is then the more thought was probably put into it. (This is why people who use legalese because they think it makes them sound smarter are actually proving the opposite.)
Shorter and simpler almost always means better communication, if nothing else. As the foundation points out, the Constitution (a great document, but not the most readable) measures 17.8 on this scale, the Gettysburg Address measures 11.2, and the "I Have a Dream" speech is down at 9.4. President Obama's recent State of the Union was delivered at an eighth-grade level (8.4), well below the average SOTU score of 10.7. FOX News illustrated this news with a picture of a kid in a dunce cap, possibly not knowing that simpler often means smarter or that 8.4 is almost exactly the same as the average American's reading level.
I just remembered that this test is built into most word processors, so I just ran it on my last several posts and the score came out to 10.0. I really have no idea what that means, if anything, but I thought I should disclose it.
Anyway, regardless of whether you think Congress's apparent decline is good or bad, there are so many possible reasons for it that it's hard to draw any conclusions about what's causing it. Since most incumbents get re-elected, it seems doubtful that any particular election year(s) would make a big difference. Maybe existing members just become better communicators as time goes on. Or maybe they really are getting dumber. They do seem to be doing stupid things more frequently, but that's a lot harder to graph.