National Polls Mean Next To Nothing In April
Last week, I passed on some of Nate Silver’s tips for how to look at General Election polling over the next six months. Silver is out with another piece, though, that points out the very simple fact that national polling in April tells us next to nothing about how things might turn out in November:
Since Rick Santorum suspended his presidential campaign, essentially ending the Republican primary contest, the starting positions for the general election have taken shape. About 10 national polls have been released: President Obama leads Mitt Romney 47.7 percent to 43.7 percent, according to the RealClearPolitics average of recent surveys.
Already, partisans and pundits are criticizing individual polling samples and pollster methodologies. Entire “-gates” – Rosen-gate and dog-gate, for example – have erupted, been fought over and then forgotten in mere days, and each poll is scrutinized to see how the latest controversy has affected the race.
It is easy to get lost in the weeds. But, of course, the election is still more than six months away, and in the past 10 presidential campaigns, the national polling leader in late April has won the election only half of the time.
Silver produces this chart of April polling vs. final results for every election going back to 1972:
So in the past ten Presidential elections, the April polling leader has gone on to win half the time, and lose half the time. Perhaps more significantly, in elections involving re-election of an incumbent, the April leader has gone on to win three out of six times. In other words, when you’re looking at these April polls, you may as well just a flip a coin if you want to know who might win in November. Or, you know, tune back in sometime around the conventions.
Yeah, but without meaningless polling data over what would political junkies obsess and fester? It’s not as if there are any other pressing issues out there. Besides, if people didn’t pay attention to meaningless polls then the likes of Scott Rasmussen would have fewer millions with which to play. That would be tragic.
@Doug Mataconis:
That would be winning the popular vote. Which doesn’t matter.
So in six of the past ten Presidential elections, and in the four latest, the leader in April has ended up winning the election.
Two other points:
The sample is really to small to make any assumptions.
And since there isn’t a popular vote, It’s all about a small number of states, a candidate can win California or Texas 95-5% and still lose the election.
I take from that, that people lied to the pollsters half the time. That seems about right.
I agree such early polling is useless which makes me wonder why OTB discuss such polls on a near daily basis.
Come on folks, we’ve got a blonde female missing with the added bonus of her being a current member of the military! ;0)
If we didn’t discuss the polls every day, we’d have to discuss solutions to national problems. If we discussed solutions to national problems, we’d eventually run into an inextricable difference between value systems. If we eventually ran into an inextricable difference between value systems, we’d have to meticulously sort through data in academia and publicly-available information to buttress our value systems. If we meticulously sorted through data in academia and publicly-available information to buttress our value systems…
…we wouldn’t be here, having this conversation.
I have, overusing syllogism, arrived at why we are here discussing the polls every day. Doing otherwise would make us both not lazy and boring.