The latest "The Numbers Guy" column from the Wall Street Journal can be found here. In it, Carl Bialik breaks down the numbers in a recently published survey about lost PDAs and cell phones.
He has good advice: "Too much precision in a statistic is a good signal to dig deeper into the methodology and the origin."
But how do we as copy editors learn enough about numbers to realize when a statistic is too precise? I think a good starting point is by reading this column every week. Get used to the questions Bialik is asking; try to understand why he is asking them. Then practice asking similar questions about the numbers you see in copy.
Most of us will admit that we have a lot of learning to do in this area. Don't wait around for the newsroom to have a brown-bag on it.