Of six polls released so far on Prop 19, the November ballot initiative that would legalize marijuana for adults 21 and older, three have the yeas winning easily, while three have it losing narrowly.
What gives? As usual, if you have a polling conundrum, Nate Silver is your man. And in this case, his hypothesis is fascinating.
Silver is the stat whiz who first created a formula for predicting -- with astounding accuracy -- the performance of baseball players. Then, under the pseudonym Poblano, he began analyzing polls to forecast the many outcomes of the long Democratic presidential primary campaign. Then he set up shop as fivethirtyeight.com, which has since been bought by The New York Times. He predicted with uncanny accuracy the 2008 election.
Anyway, Silver notes that three polls in which Prop 19 is winning are "robopolls," meaning they're automated -- you push a number on your phone to respond. In the three where Prop 19 is losing, the pollsters are actual humans asking the respondents questions.
Silver notices something else: Huge spreads in the numbers among African-Americans.
In the robopolls, blacks favor Prop 19 by 28 points or more. In the only human poll where they break out demographics, African-Americans are opposed by 12. There are large disparities among Latino voters, too.
The hypothesis: The automated polls give the respondent a more secure feeling of anonymity and no social stigma, so he feels free to voice his true opinion. (Silver also notes that it's possible the automated polls are having trouble getting a representative sample of African-Americans because they have a lower response rate to robopolls.)
But if indeed black voters are more honest with the robopolls, it would amount to a mirror of the Bradley effect.
If you like this story, consider signing up for our email newsletters.
SHOW ME HOW
You have successfully signed up for your selected newsletter(s) - please keep an eye on your mailbox, we're movin' in!
The Bradley effect was the notion -- broadly speaking -- that people lie to pollsters and say what they think they want to hear. It refers to Tom Bradley losing the 1982 California governor's race even though he was leading in late polls. The idea is that white voters, out of some vague racial guilt, told pollsters they were voting for Bradley but didn't once in the safe confines of the voting booth.
Silver: Nevertheless, it's possible that we're seeing some sort of Bradley effect in reverse, which I've reluctantly dubbed the "Broadus Effect" after the given name of the rapper Snoop Dogg, himself a frequent consumer of cannabinoid-rich products.
So the idea here is that once voters are in the cozy confines of the ballot box and can register their choice in secret, they may feel free to say yea on Prop 19.
And with that, we have delightful new lingo for the Prop 19 battle: The Broadus Effect.