The data prove it: Congress is officially “sophomoric.”

There was a nice little segment on Morning Edition last month discussing the results of an analysis done on the text of the Congressional Record since 1996. The study, performed by the Sunlight Foundation, looked primarily at the readability of speeches on the floor of both the House and the Senate. Specifically, they measured the grade level of speeches by Senators and Congressmen, using the Flesch-Kincaid scale.

The take-away from the study boils down to this: over the past 7 years, the average grade level of speech in Congress has dropped drastically, a full grade level, from 11.5 to 10.6 – hence the “sophomoric” label (for my many non-American visitors, 10th graders are typically known as “sophomores” in the US). The drop was very sudden, as the first decade of the analyzed data held pretty steady (between 11. 3 and 11.5).  While very liberal members of congress and all new members as a group rate poorly in the analysis, the association of low grade level speech and extreme political ideology is “particularly pronounced on the right,” according to Lee Drutman, the author of the study.

In their defense, several congressmen interviewed for the piece noted that this isn’t a good measure of intellectual weight, but instead shows that they’re presenting issues clearly and communicating effectively to their constituents by avoiding difficult vocabulary.

I’m all for clear and simple communication — it’s an idea I pound on all semester long when I teach (sometime in the next month, I’ll do a post on “Engfish,” which is an important interconnected idea). But except for the most general topics, tailored to the most general audiences, there needs to be an acceptable lower bound of both vocabulary use and sentence complexity.

I’m not saying this as simple opinion: the reality is that when you’re writing about any topic with some degree of specialization or expertise, you cannot avoid using specialized language (technical vocabulary, or specialized jargon) and you will probably fail in your effort if you don’t use adult sentence structures. It’s unacceptable, for example, and perhaps impossible, to present neuroscience to interested adults at a fourth grade level, or to offer a description of quantum computing theory for electrical engineers using a third grade vocabulary; for that matter, even if you’re discussing baseball, if you’re doing it in the company of experts you’re unlikely to dip below the 8th grade level, even if you make the effort. (Seriously: I recently analyzed five consecutive columns by Tim Kurkjian of ESPN  – one of the reporters I respect the most over there – and the grade level averaged exactly 8.0, with a high of 9.8 and a low of 6.3.)

As the NPR story notes, the Flesch-Kincaid rating is based primarily on sentence length and the average number of syllables per word. As such, it’s only a rough indicator of the actual reading grade level (or ‘difficulty’ level) of any text. The system can be gamed, which is why travesties like this are possible. (Here’s an explanation of the previous link; but I’ll sum up if you don’t care to follow links today: Les Perelman of MIT has shown that virtually any old nonsense can get the highest possible score on the SAT’s writing component, as long as that nonsense meets certain criteria when it comes to overall structure, average sentence length, and average word length).

The news story includes a quote from Frank Luntz. If you’re not familiar with the name, you should be: he’s the individual most directly responsible for the toxic turn of American political speech during the past decade or two (run some searches on the web and you’ll have no trouble finding some of the memos he’s produced over the years). While I find both Luntz’s politics and his manipulative application of language revolting, I found myself agreeing with the sentiment he expressed, that grade level isn’t the point: clarity of communication is (I didn’t agree with Luntz’ wording, however: he seems constitutionally incapable of speaking without introducing some sort of misdirection).

On the other hand, if our legislators need to dumb-down their speech in order to be understood by the typical voter, it’s a potentially ominous sign that there are fundamental problems with the expectations of the typical voter. Politics and democracy should require some thought, and in the long run it’s not going to serve either the politicians or the voters if this trend continues.

I don’t want to be overly critical, or jump to any too-easy conclusions about how weak vocabulary tends to indicate weakness in other intellectual areas, but I’ll highlight a quotation from the article. Rep. Rick Mulvaney, with the lowest score in the study, said this:

“I hope people don’t take it as a substitute for lack of intellect, but small words can be just as powerful as big words sometimes.”

Re-read it, if you didn’t catch the problem: “don’t take it as a substitute for lack of intellect.” Not only do we have a reading level issue, but in this case we’ve got a double negative problem: he probably meant “don’t take it as a substitute for intellect.” But, then again, the phrase itself isn’t completely clear in context. Did he instead mean “don’t take it as an indicator of a lack of intellect”? Or does the phrase refer to “small words” as the substitute for intellect? I really just don’t know what he actually meant here — which is either an indication of his skill at deflecting questions, or a disheartening sign that the grade level issue is only the tip of the iceberg when it comes to the quality of our legislators.

It might interest you to know that according to Microsoft Word, this blog post had a Flesch-Kincaid grade level of 12.6 in the first draft (with 29.3 words per sentence). By the time I’d finished with it, that had shifted up to a 15.1 grade level (with 31.2 words per sentence). I’m not sure exactly what that means…but I thought you might like to know.

= = =

Note: I’ll be away for most of the next two weeks, so there will be no new posts until at least July 13th. Inquiries about editing jobs will be responded to as soon as I’m back in a place where I can reliably access email.


About thebettereditor

Chris holds a BA degree in history from the University of Virginia and a Master of Fine Arts (MFA) Degree in writing from the University of Southern Maine (Stonecoast). He has worked extensively with professional and semi-professional writers and enthusiastic amateurs for about 20 years. He has several years experience in scientific publishing, but has also worked in information technology, insurance, health care, and education (he taught writing at the university level for a number of years). Since 2011, he's also specialized in helping small business meet their writing and editing needs on a budget.
This entry was posted in Culture, Language, Things you should know, Writing and tagged , , , , , , , , , , , , , , , . Bookmark the permalink.

2 Responses to The data prove it: Congress is officially “sophomoric.”

  1. Carol says:

    The articles that reported on the Sunlight findings were pretty thoroughly debunked by experts as sensationalist and misinformed, for instance by Chad Nilep at the Society for Linguistic Anthropology and Mark Liberman at Language Log.

  2. I haven’t read the original Nilep rebuttal, but I thought Liberman’s response a little strident. Yes, yes: F-K has all sorts of problems, but it’s still a useful (if flawed) indicator. His dismissal of the analysis as nothing more than a headline grab seemed as simplistic a tactic as what he accused Sunshine of.
    Maybe I buried it too deeply (…see the paragraph that mentions Perelman…) but I recognize that an analysis like this shouldn’t be taken uncritically.
    On the other hand, it highlights things that deserve to be looked at more seriously. If I had the tools, there are some follow-up analyses that I’d love to see done with Congressional Record data that would be a lot more interesting. I think, for instance, that historical snapshots of vocabulary at 5 or 10 year intervals over several decades would reveal a lot; but I’d want to see that data compared to other sources from the same times to show if Congressional speech is leading, following, or simply mirroring the complexity of language in society at large.
    That would tell us a lot more about the real state of things.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.