They said there’d be no math.
And that’s why most of us went into journalism. But don’t be afraid: spreadsheets are our friends.
One of the messages constantly drilled into the minds of eager content producers at the 2013 Online News Association conference in Atlanta earlier this month was: newsrooms must embrace data journalism.
“You should be reading statistics, looking up charts every day,” said Emma Carew Grovum, a data journalist at The Chronicle of Philanthropy. She ran a seminar called Big Data, Little Newsroom at #ONA13. You can check out her slides.
Carew Grovum says if people in your newsroom don’t have the skills to analyze data, they need to learn. “Good documentation, outlines, deployment schedules – arm reporters with as much as they can handle.”
And yes, she says, “sometimes data means math.”
Nate Silver agrees. This master of statistics was the rock star who made an appearance at the ONA13 conference.
Silver, a number cruncher and journalist, is the creator of fivethirtyeight.com. (By the way, 538 is the number of electors in the U.S. electoral college system.) Silver correctly predicted the outcome of the 2012 U.S. presidential election in 50 states. Collecting data and analyzing it is his thing. That’s why ESPN bought his brand and is now setting up a new team to focus on political issues, sports and economics.
As he builds that new team, Silver said he’s looking for critical thinkers; journalists who know how to look at data sets. At ONA, Silver listed his eight cool things journalists need to know about statistics.
1) Statistics aren’t just numbers. It’s important to properly interpret the data. Know where the data has come from. Silver says the questions that need to be asked about stats are no different than any good journalist would ask about any story.
2) Data requires context. You can’t just throw out a statistic without putting it in the proper context. (He used the example that China is quite often held up as the 2nd biggest economy in the world, without recognizing it’s also a huge population. If this country’s economy was gauged on a per capita basis, it would rank differently, because it’s actually a poor country despite it’s growth.)
3) Correlation is not Causation: Silver pointed to the idea that in the summer, both murder rates and ice cream sales go up. There’s no direct connection, so we shouldn’t create one. This is another example of journalists needing to simply use critical thinking skills when making comparisons or linking issues.
4) Take the Average – Stupid: For instance, he suggests the outlier polls should likely be ignored. If a politician loses one poll out of 20, it likely means that one poll is less critical. But he says journalists like to turn that outlier poll into the story. He says reporters need to acknowledge the grey areas.
5) Intuition is a poor judge of probability.
6) Know thy priors: This is where he gets into Baye’s Theorem (He also does this in his book: The Signal and the Noise: Why so many Predictions Fail, but some Don’t.
Baye’s is a 200 year old formula to test and refine beliefs. He says when your theory seems to be wrong, you need to acknowledge it.
7) Insiderism is the opposite of objectivity: He goes against the grain here by saying access to inside/secret information is sometimes not as valuable or accurate as you think. Silver says it leads to elite and insider opinion, which isn’t the same as accurate opinion.
8) Making predictions improves accountability: He says, “Journalists, like statisticians and like scientists, ought to be concerned with truth rather than appearances.”
The move towards “open data” is creating an explosion in the kind of information and statistics now available to reporters. But it takes a lot of time, energy, tech-smarts and patience to actually analyze and discover the stories hiding in all that data. Yes, we may need to practice a bit of math in order to deliver some good, enterprise journalism. Just don’t forget the art of story-telling along the way.