Is newspaper reporter really the worst job? A lesson in the sniff test
If you were on social media yesterday you probably saw that story that was going around about how newspaper reporter is the worst job of 2013.
You probably saw it because, as a sports fan, you likely follow some newspaper sports reporters. That make you think? It made Washington Post columnist Gene Weingarten think:
I bet this Best-Worst jobs ranking was manipulated to put newspaper reporter last. Assures wide coverage.online.wsj.com/article/SB1000…
— geneweingarten (@geneweingarten) April 23, 2013
Did you retweet, comment, pass the story along? Did you ever get the feeling you’ve been used?
Most of those who wrote about the report did so with little or no questioning of CareerCast’s methodology, which the site, to its credit, explained here. You’ve probably guessed by now that this is another post about verifying things you see online. The main messages this time:
- Consider the source: In this case, the source is a jobs-search website. Is there any reason other than marketing to do and release this study? Might CareerCast have something to gain by—as Weingarten suggests but of course does not prove—skewing the results to maximize publicity?
- When someone is pushing analysis based on data, have they shown that the data says what they’re saying it says? Do their conclusions pass the sniff test? Are they basing their data analysis on flawed assumptions?
“There are some subjective pieces but, frankly, it’s really driven by the data,” CareerCast publisher Tony Lee told Poynter.org about the best/worst jobs report.
This is where your skepticism needs to kick in. “It’s really driven by the data” is the new appeal to authority. We, as reporters, are asked to believe that something is true because, hey, the data says so. This isn’t my opinion or anything, the story goes. It’s the objective data talking.
Don’t you believe it.
Remember the old computer programming saw: Garbage in, garbage out. Figuring out which data to look at or how to count something or how to weigh different data points are all subjective choices that can have a huge effect on what the data ends up telling you.
We as reporters can’t always engage with the data. Sometimes it’s not available to us—which alone should give us pause—and sometimes it’s too complex for us non-mathemeticians to parse. But our first line of defense is the sniff test.
Here’s a little excerpt from CareerCast’s best/worst jobs list:
49. Social Worker
50. Physician Assistant
Does that look right to you? Out of all the thousands of surgeons in the United States, how many do you think would trade jobs with a physician’s assistant? With a social worker? I think the surgeons who’d trade could all sit at one table at Denny’s, don’t you? They’d talk about how nobody gets them.
CareerCast’s methodology page explains that it took into account four factors that are inherent in any job: environment, income, outlook and stress. Based on data from various sources, mostly governmental, each job type was given a score on a number of elements within each of those four. Data!
But look at my little sub-list again, this time with income included.
49. Social Worker – $41,169
50. Physician Assistant – $89,097
51. Surgeon – $311,078
Income isn’t everything, of course. But is there any way that whatever measures of stress, environment and outlook make social worker a “better” job than surgeon, they’re enough to overcome an income that’s seven and a half times greater? Is it likely that the typical surgeon would be happy to take a job, physician’s assistant, that’s far lower in the hospital hierarchy, with less prestige and a 71 percent salary cut? Because that would be the case if physician’s assistant were actually a “better” job than surgeon. It would be an easy transition. Surgeons are already qualified to be physician’s assistants. They’d be tossing aside their scalpels from coast to coast.
Similarly, how many newspaper reporters, with all their stress and poor outlook for advancement, would trade jobs with the average cashier, a job ranked 31 spots higher on the CareerCast list, despite income that’s barely half of what newspaper reporters make—itself so low that it’s counted as one of the reasons newspaper reporter is such a terrible job?
This kind of thing happens in sports all the time, is why I mention it—at such length. An example: In the mid ’00s a trio of economists wrote a book called “The Wages of Wins” that asserted, based on their newly invented metric, that Dennis Rodman was a more productive player on a per-minute basis than Michael Jordan.
I was one of many, including many statistical experts, who objected. I couldn’t debunk the math, but I could sure administer the sniff test. Dennis Rodman better than Michael Jordan? You better prove that one beyond any doubt. As I sorted through the authors’ claims, I concluded that, just as I think CareerCast is undervaluing salary now, I thought the economists were overvaluing rebounds then. Some of the stat folks agreed with me, and others had other criticisms, including that the authors weren’t weighing shot attempts correctly.
When data purports to show something surprising, it may well be because there is something surprising going on. Heck, maybe newspaper reporter really is a worse job than cashier—and every other job. And maybe Rodman was a more productive player than Jordan. But a heavy burden of proof is on the person making a statement like that. Are they backing it up, or are they waving around the word “data” and hoping you’ll take them at their word?
Too much of the coverage of CareerCast’s study took the conclusion as a fact. Too much of it said, “Newspaper reporter is the worst job and actuary is the best” and then pretty much repeated the article on the CareerCast site. Not enough of it said, ”Job-search site says newspaper reporter is the worst job” and then questioned the conclusion.
Keep your sniffer close by and in good working order. It comes in handy. It’ll even come in handy if you land a job as a newspaper reporter, which, many newspaper reporters say, is, in spite of it all, a pretty good gig.
Simple observation: If newspaper reporter really were the worst occupation in America, it would be easy to find a job online.wsj.com/article/SB1000…
— Justin Wolfers (@justinwolfers) April 23, 2013