Professor Pedantic 教授的考究學問
The professor awaits your query on academic writing, though in all honesty, he doesn’t have a lot of time for you. He is a tenured full professor and working on yet another magnificent academic tome. Even so, he has graciously consented to entertain your question. Submit it and prepare to be edified.

QUESTION: I always include statistics when I can to prove different points of a paper, because the numbers sometimes tell the story better than words can. How do I know when I have overused statistics in a paper?

One cannot address this subject without rolling out the truism about statistics, namely, that there are lies, damned lies, and statistics. The inference is, of course, that while hard numbers might seem straight-forward fact, when misused, numbers can be as misleading as any stated falsehood. For example, an imbalanced sampling can produce numbers skewed one way or the other. Or a false cause-and-effect relationship might be deduced from raw data, either from error or on purpose. Statistics unquestionably do not tell a whole story when they are used mischievously.

So how often should you employ statistics in an academic paper? As often as the numbers can honestly advance or bolster an argument. That is the purpose of every argumentative device employed by a writer in a paper, to persuade the reader of the worth of an argument. Yet you are right to fear that too much of a good thing can detract from, rather than add to, the overall effort to persuade. At some point, numbers piled on numbers begin to look like padding. After all, academic papers are exercises in written communication, not statistical round-ups.

The other misstep in using statistics concerns interpretation. Some data need no interpretation beyond a succinct restatement. Other compiled numbers need restating in some detail to reinforce the import of the numbers or to ensure that the reader grasps their meaning. (If even more explanation is needed, a footnoted table probably is the wiser choice.) Where some writers err is in misinterpreting numbers or in applying them too generally to make a case. A pertinent reminder: The professor reading a paper is apt to be as statistically conversant as the writer.

cron web_use_log