Sunday, July 8, 2012

Sentences Within Sentences

With the publication of Issue #17, the OTP blog returns!

A writing technique I've seen few authors get away with is the placement of complete sentences within other complete sentences. Below is what I consider a particularly telling example of what can go wrong when you try to put too much information into one sentence.

The quote comes from Statistical and Machine-Learning Data Mining: Techniques for Better Predictive Modeling and Analysis of Big Data, Second Edition, by Bruce Ratner.

(Don't worry if you don't fully understand the technical content. Just examine it as one huge sentence. And it's a real shame this sentence had the same effect on my reading that a speed bump has on my driving, because for the most part, I thought Ratner's book was well done. In fact, his explanation of CHAID and related techniques is the best I've read to date. All that aside, I wish someone had revised this next bit before it went to print:)

CHAID is a popular technique, especially among wannabe regression modelers with no significant statistical training because (1) CHAID regression tree models are easy to build, understand, and implement; and (2) CHAID underpinnings are quite attractive: CHAID is an assumption-free method, meaning there are no formal theoretical assumptions to meet (traditional regression models are assumption full, which makes them susceptible to risky results). 

That was published as one sentence! I count five distinct sentences in all that text. Here's how I'd rewrite it. For clarity, I put a number in brackets in front of each separate sentence.

[1] CHAID is a popular technique, especially among wannabe regression modelers with no significant statistical training. [2] First, CHAID regression tree models are easy to build, understand, and implement. [3] Second, CHAID underpinnings are quite attractive. [4] It's an assumption-free method, meaning there are no formal theoretical assumptions to meet. [5] (Traditional regression models are full of assumptions, which makes them susceptible to risky results.)

I see such beyond-run-on-sentences quite frequently, especially in technical and academic writing aimed at audiences with a specific educational background. I know it's easy to get carried away with long sentences; I do it myself. When revising, examine those long sentences carefully and ask yourself if splitting them into shorter sentences would work better.

(As a real example, I considered removing the semi-colon in the "I know it's easy..." sentence. After some thought, I decided "I do it myself" is attached enough to the prior idea that the semi-colon works. However, the prior two sentences were originally one long one. Yes, that's right, I violated the very rule I'm talking about while drafting an example of not violating the rule. That's how ingrained long sentences are in me! Sad, isn't it?)

No comments:

Post a Comment