Progressives often respond to conservative stories with data. That’s a problem, because stories are more convincing. (More)

Climate Change, Campus Assaults, and Police Violence: Data vs. Stories

Yesterday we discussed a new study that found Southern white voting patterns in presidential elections from 1968 to 2000 correlated with county-level Ku Klux Klan organizing in the 1960s.

Of course, that’s just a statistical correlation, as a conservative reader argued in response to the Raw Story article:

Really!? These morons have never heard of the first rule of statistics, correlation does not imply causation? If this was a freshman sociology paper, I’d give them an F.

Deadly Ice Cream?

It’s true that correlation does not imply causation. You’ll probably feel a breeze when a nearby windmill is turning, but that correlation does not mean the windmill is causing the breeze. Instead, the breeze is causing the windmill to turn.

A statistical correlation may be caused by a third variable. For example, monthly ice cream sales correlate to homicides, but not because ice cream makes you homicidal or because murders make people want ice cream. Instead, both ice cream sales and homicides increase in the summer months, and for different reasons.

Finally a correlation may be random noise. For example, many people believe you can predict the stock market fall based on that year’s Super Bowl winner. From 1967-1997 the ‘Super Bowl Indicator’ was right 28 of 31 times, an impressive 92%. Since then, it has slipped to 80% and, given a large enough sample, it will likely drop more. As the Wall Street Journal put it:

Looking at history since 1966, there is a 70.2% likelihood that an NFL team wins the Super Bowl and there is a 72.3% chance that the Dow advances in a given year. Given these probabilities, by simple chance the Super Bowl Indicator should be correct 59.1% of the time.

“Lies, damned lies, and statistics”

In 1906 Mark Twain popularized the quip “There are three kinds of lies: lies, damned lies, and statistics,” although he falsely attributed it to former British Prime Minister Benjamin Disraeli. The joke had been used in speeches and letters to British newspapers since at least 1891, so perhaps Twain incorrectly believed that Disraeli had coined it.

Regardless, it’s become common coinage. We progressives can cite plenty of data on climate change, campus sexual assaults, and that police are more likely to arrest and kill black people. And conservatives routinely offer Twain’s quippy response for allthreeissues. They often follow up by demanding and then purporting to ‘debunk’ specific examples, from weather events to the Rolling Stone article campus rape to the deaths of Michael Brown, Eric Garner, and other black victims of police misconduct.

“A discrete incident with its own unique facts”

Conservative Robert Tracinski admitted as much in his Federalist column on Eric Garner’s death:

As I warned, one of the most insidious errors you can make is to turn each case into a symbol of “systemic racism” rather than an individual case to be judged on its own merits.

He clarifies in that embedded link:

People are individuals. Their actions are individual actions. Every case of the use of force is a discrete incident with its own unique facts. It is not an abstract morality tale about racism or poverty or heavy-handed policing.

Set aside the overweening privilege inherent in that claim. Focus instead on the willful, ideological blindness to structural and systemic issues. He argues that statistical data has no place in stories about our justice system. Every “discrete incident” has “its own unique facts.”

No matter how many trees we see around us, Tracinski demands we discuss only one tree at a time. And if he and other conservatives can discuss the tree du jour without reference to a forest … then there is no forest.

“Can’t see the forest for the trees”

As progressives, we can’t rely solely on data. Stories are more convincing, and more memorable, and we must present stories that illustrate our data. But then we must be ready to defend those stories from conservative attacks. And one good rebuttal is the familiar quip: “Can’t see the forest for the trees.”

Don’t get spun into knots defending every detail. Attempting that will get you Gish galloped into a ditch, and no every-detail-100%-true story will be a perfect exemplar of anything. Life is too messy for that, and people who demand we put every branch, twig, and leaf under a microscope will never see the forest.

Not because the forest isn’t there, but because they don’t want to see it. You won’t change their minds. But mentioning the forest might get another listener to notice that … wow, there sure are a lot of trees….


Happy Thursday!