Post-election Analysis Done Correctly

Instead of letting your blood pressure rise unnecessarily due to listening to political pundits talk about the recent elections, why not just start with Nate Silver?

Why did Democrats lose in Virginia and New Jersey on Tuesday? Because independent voters moved against them, say the pundits.

This is true, insofar as it goes; Democrats lost independents nearly 2:1 in the gubernatorial race in Virginia, and by a 25-point margin in New Jersey.

But it doesn’t really tell us very much. It’s a lot like saying: the Yankees won the Game 6 last night because they scored more runs than the Phillies. Or: the unemployment rate went up because there were fewer jobs…. [Source]

BTW, I am getting together a listing of resources for bloggers who may not be statistical wizzes. (I include myself in that number—educational credentials notwithstanding.) My idea is that as blogging as a means of social media has matured, we bloggers need to be more intentional and factual about how we interpret statistics. If you have any links to suggest, please drop me a line.

 

 

Cranberry Homophobia

"Cranberries in fudge." villoks, http://www.flickr.com/photos/villoks/34352744/

Good point: Now that marriage equity has been repealed in Maine, will we see the same vitriol against segments of the Maine population as we saw against Black Californians (actually, all Black folks) following Proposition 8? Will Dan Savage write about feeling betrayed by small town and rural voters in Maine, who apparently voted overwhelmingly in support of Question 1?  Will he write,

I do know this, though: I’m done pretending that the handful of racist cranberry- and lobster-hating gay white men out there—and they’re out there, and I think they’re scum—are a bigger problem for African Americans Maine residents, gay and straight [and cranberry- or lobster-loving], than the huge numbers of homophobic African Americans cranberry and lobster lovers are for gay Americans, whatever their color [or juice/seafood preferences]… [Original quote here]

Probably not.

At the time of Prop 8 I read lobster-crates full of incredibly racist, stunningly naive, or just plain ignorant commentary about Blacks and marriage equity and responded around blogland until I was cranberry-red in the face and utterly depressed. Eventually I quit reading and quit commenting, concluding that folks just needed to vent and that they were just (unfortunately) using Black folks as a temporary whipping group.

But I am not so sure. It is my belief that that racism never goes away and never will go away. That may be an unpopular and pessimistic view, but it is one that I hold. I do not think there will ever be a “cure” for racism, resulting in its eradication forever. There may be vaccines for uninfected young people. There may be treatments that put the scourge in remission. But it will always be there—virulent as ever in those who refuse treatment, ready to newly infect others with mutated strains, and powerfully rebounding in those who thought they were over it for good.

The fight, then, to manage racism generally and anti-Black racism specifically is an on-going struggle. As is and will be the fight against homophobia. The building and maintenance of ally relationships will have ups and downs. And probably more of the latter than the former. There will be steps forward* as well as steps—like Question 1—back.

But the way forward when we find our chins deep in the cranberry bog is to pick ourselves up, dry ourselves off, and start all over again. (Without pointing fingers at others absent the benefit of facts and context.)

*And there’s at least this look on the bright side from the world of the Chicago Cubs.

“We’re a complicated country, and we have complicated politics.”

Real data plotted by Nate Silver reveals that

The correlation between the fraction of conservatives and the fraction of liberals agreeing to a given question is essentially zero.

Is this surprising? Perhaps. If conservatives and liberals had fundamental disagreements on most major political questions, you’d expect to see a statistically significant inverse correlation in their responses. But you don’t see that. Conversely, if they agreed on most of these fundamental questions, with the differences being only around the periphery, you’d expect to see a statistically significant positive correlation in their responses. But you don’t really see that either.

What this means basically is that people who identify as conservative and those who identify as liberal agree sometimes and disagree other times, and the (non-)pattern of this agreement/disagreement is not linked to their identification as conservatives or liberals.

Meanwhile, additional (fake) data plotted by PPR_Scribe at GraphJam:

moar funny pictures

moar funny pictures

Nate Does Oscar

"Film reels." mag3737, http://www.flickr.com/photos/mag3737/458226761/

Brilliant statistician Nate Silver of FiveThirtyEight.com has lent his considerable numbers crunching talent to the prediction of this year’s Academy Awards. He has Slumdog Millionaire at 99.7% to win for the Best Director category, and 99.0% to win for the Best Picture category.

If there is an upset here, Silver thinks it will be Milk, but perhaps only winning one of these two big categories.

For Best Supporting Actress, Silver has Taraji P. Henson (The Curious Case of Benjamin Button) at a  51% chance of taking the shiny golden man home.

(He has more comments about his predictions on his website here; He also recently appeared on Countdown with Keith Olbermann.)

This is how he figures it all out, as explained in New York Magazine:

Formally speaking, this required the use of statistical software and a process called logistic regression. Informally, it involved building a huge database of the past 30 years of Oscar history. Categories included genre, MPAA classification, the release date, opening-weekend box office (adjusted for inflation), and whether the film won any other awards. We also looked at whether being nominated in one category predicts success in another. For example, is someone more likely to win Best Actress if her film has also been nominated for Best Picture? (Yes!) But the greatest predictor (80 percent of what you need to know) is other awards earned that year, particularly from peers (the Directors Guild Awards, for instance, reliably foretells Best Picture). Genre matters a lot (the Academy has an aversion to comedy); MPAA and release date don’t at all. A film’s average user rating on IMDb (the Internet Movie Database) is sometimes a predictor of success; box grosses rarely are.

That Nate Silver has gotten into the entertainment prediction business is wonderful news for someone like me who rarely ever sees more than three or four of the nominated films prior to the Oscars megaprogram. And who is not that much into fashion, or much interested in knowing who’s zoomin who. Now if I watch I will have something to do–seeing if logistic regression analysis and Oscar history are worthy predictors of who gets to thank the Academy, their agents and mothers at the podium and who must content themselves with claiming “It was an honor just to be nominated.”

Perhaps I’ll even watch the award ceremony with a big bowl of popcorn and a beer or two–the latter in support of my country’s economy: more of Silver’s numbers crunching suggests that beer may no longer be “recession proof.”