10 August 2008

Words, More Words

In the midst of everything I do on a daily basis, it is sometimes easy to forget that, ultimately, words are my business: ideas and concepts, clearly expressed through words, are the engines upon which my life depends. If this sounds rather obvious, I will not argue the point! Still, it is important to remind myself periodically that the clarity of language can be subjective, and to look for opportunities to improve my writing—even while recognizing that subjectivity.

Therefore, I was thrilled to stumble across the “Simple Measure of Gobbledygook” or “SMOG Calculator.” The approach was developed in 1969, by clinical neuropsychologist Harry McLaughlin, who described his formula in plain terms: “...count the words of 3 or more syllables in 3 10-sentence samples, estimate the count’s square root, and add 3.” The online version of the tool makes it possible to submit texts for an instant analysis of the writing level, resulting in a score on a scale of 1 to 19+. The score correlates both to reading level (e.g., junior high school or university degree) and to a sample publication that fits the reading level (e.g., Reader’s Digest or Atlantic Monthly).

The SMOG calculator then reminded me of a similar site I came across a few years ago, the so-called “Gender Genie.” This tool uses a word analysis algorithm (developed by Moshe Koppel of Bar-Ilan University in Israel and Shlomo Argamon, from Illinois Institute of Technology) to determine the gender of the author based on the presence and repetition of certain words. Again, a submission box generates an immediate analysis, a pair of scores showing the female / male weighting of the text, a list of the critical words that were evaluated, and the tool’s final conclusion about the author’s gender.

Taken together, these two tools can be addictive. To start, I tested eight items I have written, with interesting results: my average SMOG grade for these five items was 15.17, which places me between the “Some college, New York Times” range (a SMOG score of 13-15) and the “University degree, Atlantic Monthly” mark (a score of 16). This sounds right to me: I would say that these texts should be generally accessible to a reasonably educated audience, without being as obtuse as “IRS code,” the highest (or, worst) SMOG score available. (See below for a summary of the specific writing samples I tested, with links to those pieces, and the cumulative and average scores from each tool.)

The Gender Genie guessed I was male 88% of the time, though the difference between the male and female scores on certain texts was in one case as low as 39 (in favor of a male author) and in another as high as 1381. If one takes the science behind the Genie as meaningful, these results suggests there is great variability in the gendered language I use in my writing. I’ll leave aside broader implications about my personality, but for fun, I did test a more personal piece of writing: the Gender Genie pegged the author as female by a lead of 46 points.


Over time, I have come to two different, but complementary, conclusions about writing. The first conclusion is that good writers tend to be confident that they know what is readable, and that they have a good handle on the clarity and calibration of their writing to specific audiences. At the same time, good writers are also aware of when their writing needs the work of an editor (even if they do not always take advantage of one). From my own experience, I have developed various processes to evaluate my written work—from different approaches to re-reading, to knowing to whom I can turn for an edit—and each of these steps help identify problems and catch inconsistencies. I also use other tools periodically, and have even been known to enjoy writing-style brain teasers like this one, which help keep me alert to mistakes I may be making.

So it is interesting to me to think about how the SMOG Calculator could be used to evaluate something I have written before I share or publish it, to evaluate very basic, but important, questions: Is this piece of writing as readable as I think it is? Is the language effectively calibrated for the intended audience? Simple formulas can have their drawbacks—but may also reveal very different elements than the more contextually driven feedback provided by a human reader. And while (generally speaking) my gender is irrelevant to much of what I write, there have certainly been moments when (perhaps in recognition of the different styles of language men and women tend to use) feedback from female friends or colleagues has helped me write more clearly and effectively—suggesting that even the Gender Genie could provide useful information.

The late comedian George Carlin once said “Words are all we have, really,” and he had a point. All the more reason to take care with the words we use, and to make sure that we continually evaluate how we use them, and that we are writing (and speaking, too, for that matter) in a manner that most effectively conveys what we mean.

And for anyone interested: the SMOG score for this op-ed? A grade of 15.25. The Gender Genie is convinced the author is male, by a score of 1369 to 667.


Writing evaluated for this article, analyzed by both the SMOG Calculator and the Gender Genie:

Chinese Torture, Olympic Style
SMOG Score: 12.24
Female Score: 859
Male Score: 1015

In Pursuit of Happiness
SMOG Score: 14.6
Female Score: 1069
Male Score: 1532

The Jobs and Education Con Game
SMOG Score: 16.78
Female Score: 2799
Male Score: 3132

Women Aren’t Commodities
SMOG Score: 17.11
Female Score: 2438
Male Score: 3581

V for Dissociate
SMOG Score: 16.42
Female Score: 935
Male Score: 1413

R.I.P. Elinor
SMOG Score: 13.49
Female Score: 695
Male Score: 649

Arts & Public Policy - A Book Review
SMOG Score: 16.2
Female Score: 1388
Male Score: 1427

On Trends & Statistics
SMOG Score: 14.48
Female Score: 835
Male Score: 2216

Average of all the above scores:
SMOG average score: 15.17
Female average score: 1377
Male average score: 1871


Post a Comment

Links to this post:

Create a Link

<< Home