Last weekend I flew to Phoenix for the IRE boot camp in statistics hosted at Arizona State University. Three days and 52 cups of coffee later, I can spot statistical significance. I can run a linear regression on a dependent and independent variable, and I might even be able to tell you what an R-Square value actually means – not only to a statistician, but to a journalist.
Statistics is not one of those things you master in three days of training, but here are the three most valuable takeaways from the weekend: A sense of what can be done, an inkling of how it might be done, and a desire to do it.
That last one is the tough one. Journalists hate math, right? Why do this?
The glory factor, for one. Statistics can produce great investigative stories (see the examples at right). And, sure, Pulitzer Prizes. And, don’t forget, you will look very smart to your superiors.
Statistical analyses can provide hard evidence to back up (or discredit) a theory: “do young women get fewer speeding tickets?” “Are minorities less likely than whites to be approved for a home loan, given equal income?” A well-designed analysis isn’t easily brushed off by authorities.
More importantly, there’s the reality factor. Statistics are everywhere, inescapable. I’d wager that you’ve already come across a statistic-based story today (unless reading the IRE blog is the first thing you do in the morning).
Most of us can’t turn linear regression into a story; although after last weekend’s training I know I could (with help) attempt it. But statistics are relevant if you cover:
- Education (standardized test scores)
- Business (median household income, consumer price index)
- Politics (election polls, exit polls)
- Crime (crime statistics)
- Health (birth rates, death rates)
If I haven’t convinced you of anything yet, listen to this: Steve Doig, Knight Chair in Journalism at ASU, recently helped a student analyze provisional ballots in Arizona based on the minority percentage of the 724 precincts. Their findings aren’t surprising: the more minority voters, the more provisional ballots were given. When confronted with this statistical reality, what did the state officials say? In effect, “leave the statistical analysis to the government, OK?”
Now, if that doesn’t make you want to run for the nearest copy of “Statistics for Dummies”, I don’t know what will.