Where's the Liberal Bias?

Remove all ads!

Now that all the votes have been counted, we can actually judge how well the polling firms did with the election. An analysis has been performed to gauge polling bias during this election cycle. In this sense, bias does not represent a deliberate action by the polling company to skew the results one way or another, but merely the margin by which their prediction came to the actual results.

So Obama won the popular vote by 2.6%. Any polling firm that had that number on the nose (none did, but some came within half a percent), would have no bias in their poll. A poll that had Obama favored to win by 4.2% would have a 1.6% Democratic bias, while a poll that showed a dead tie would have a 2.6% Republican bias.

If there is a liberal bias in media, it apparently doesn't extend to the polling firms. Of the 24 polling firms that released results in the week preceeding the election, 21 of them had results more favorable to Mr. Romney than what he actually received. Even Public Policy Polling, which tends to have Democratic leaning results, had Obama as the winner, but by 1.5% instead of his eventual 2.6% win.

However, the polling firm that had the worst results was one that is generally regarded as one of the most prestigous polling firms around - Gallup. Gallup missed the election by 7.2% in the Republican direction, not only missing the winner, but actually calling for a Romney win that would have been similar to what Obama had in 2008. (If Romney actually had won the popular vote by nearly 5%, not only would he have likely won all the close states like Ohio, Iowa, Colorado, Virginia, and Florida, but states like Pennsylvania and New Hampshire may well have wound up in his column as well.)

One of the more surprising results was that the online polls actually faired better, when taken as a whole, as compared to the call based polls. Prior to this year, the firms using live callers had the best results. The worst polls were those performed by automated polling firms. That's nothing new - automated polls have had among the worst results for several years. There are laws in many states that prohibit automated calling to cellphones, and now that about 1/3 of Americans use a cellphone as their only phone, automated polling firms miss a huge subset of the population, meaning their polls are no longer random samples. (They tend to miss on the Republican side exclusively - cellphone only people tend to by young, urban, and more likely to be African American or Latino, all characteristics that correlate with Democratic voting.)

The Best and Worst Polls

 

Taluntain is clearly a plant, to lead you off the scent of the bias! Didn't you know that Taluntain is an anagram of "A Nil Taunt"? It's all falling into place! You can't fool me.

 

I just remember in the weeks leading up to the election that even though the polls were close, everyone said that there was a Democratic bias in the polls, so Romney was actually ahead by a few points, and that the race wasn't really tied. Nate Silver actually did the talk show circuit in the week leading up to the election, and his prediction of Obama as at least a 75% favorite (that went all the way up to 90% by the day of the election) was roundly criticized. He actually stated that the election wasn't even a tossup, owing to Obama's advantage in the swing states, and that a Romney win, while possible, was not likely. Nate ended up going 50 for 50 in his state predictions. (And it's not the first time he was very accurate - he hit 49 out of 50 in the 2008 election. He missed Indiana if you're wondering.)

So I admit to getting a chuckle watching the Republican talking heads sputter as the election results started coming in, especially Karl Rove. Would you like another serving of crow Mr. Rove?

 

Well, it's certainly helped Karl Roves profile. Nobody in this part of the world had heard of him until last week, but now his Election Day clowning around have made him a YouTube hit to rival Gangnam Style.

 

Really? I could believe that you didn't hear about him in a long time, but Karl Rove was referred to as the "central nervous system" of the Bush presidency. He came to fame (or perhaps infamy) in Bush's second term when a special prosecutor started investigating use of waterboarding in the military, and who knew what, and when. While I'm not aware of Rove ever holding public office, I'm pretty sure most people who follow US politics even a little bit have heard of him.

 

I'm not sure what polls you and Nate were seeing -- the numbers he's putting up don't make any sense at all to me (are you sure he doesn't just pull them out of his ... well nevermind). Most polls I saw had Obama ahead by 2-5% with taking 300-330 electoral votes. The Gallup poll was interesting and I even posted a link to it -- Gallup showed Obama ahead by ~1% of the total polled but in the category of "voters most like to vote" Romney held a lead by ~2% (they obviously were not counting on the 'get out and vote' campaign by Obama to work). When the Gallup poll came out it surprised everyone in that it was the lone (unbiased) poll that gave Romney a chance and they were criticized widely for being so divergent from the other respected polls.

If Nate wants to pick and choose the polls he puts in his column just so he can look "prophetic" more power to him. I think his article and data are crap (and not because of the election result -- again, this election was a choice between two marginal candidates in my eye and I voted for the one I disliked least). As a voter, I'm republican leaning but when I'm unhappy with current leadership I generally vote against incumbents -- including my current republican congressman (he's a jerk). I was very happy Duckworth beat that slime Walsh.

 

Most of the predictions I saw had the electoral votes a lot closer than that. But granted, I'm in Canada, and I didn't really follow actual polls that closely.

Well, that could be said about any pollitical system. Anyone who follows New Zealand poltics even a little bit has probably heard of the United Future party.
(Edit: For the record, while I had heard of Karl Rove before last week, and knew he was somehow tied to the Republicans, I didn't really know who he was.)

 

T2B,

The numbers in my link (as I thought I explained, but apparently not well enough) were not the numbers that the polls released prior to the election, but rather the amount their prediction diverged from the actual result. So any number that is Romney +2.5 or less would have correctly picked Obama as the winner, but by a smaller margin than the actual result.

Releasing the results of all 24 national polls doesn't sound like cherry picking, btw, unless you mean he was picking all the cherries from every tree.

 

Nah, I'm just dense. But I still don't see where he got his numbers. As I mentioned, the polls I saw nearly all had Obama with a few point lead. There are always some "sub-polls" which show other factors to consider and I would believe the worst of those may be want Nate is quoting.

 

Nate states in the article that he took all the polls from the week prior to the election, so anything from October 30th - November 5th. In the case where a pollster released more than one poll on the presidential race prior to the election (as was the case for all the companies that released a daily tracking poll), he used the average of their results for the preceeding week.

Once again, the results shown in the link were NOT the numbers the pollster released. They were a comparison of the numbers of what the pollsters released as compared to Obama's margin of victory of 2.6%. So any poll listed as Romney +2.5 or less (and the vast majority of them were) were polls that were indicative a small lead for Obama, but smaller than his actual margin of victory.

I really don't see where the confusion is here. Obama outperformed his polling numbers, so it should come as no surprise that most of the misses by the pollsters were in Romney's direction. It doesn't require any deep thought or analysis to conclude that - it's pretty self-evident. If you outperform your polling numbers most polls had the other guy in a more favorable position, and if you underperform your polling numbers, most polls had you in a more favorable position.

 

Good old Peter Dunne, the leader of a one man party in government. That's what you get with proportional representation.

And, I just asked my wife if she knew who Karl Rove, and she's still never heard of him. The only Rove she knows is Rove McManus.

 

All too true Harbs. That's why I've always been a fan of it. There are no fewer than two different "types" of Democrats in the US, and they easily could be a separate party. With the rise of the Tea Partiers in the Republican ranks, there are no fewer than three different types of Republicans. Ironically one of the two groups of Democrats and one the three groups of Republicans could be broadly described as Libertarians, and actually hold decently close social views. Throw in a Green Party, that would have no chance of winning an election, but could probably grab 3%-5% of the popular vote, and I think the US could reasonably split into as many as five different political parties, and as a result you'd get a much more broad based view of ideas.

I think one of the reasons that politics is so divided in the US is because of the two party system. It's an A or B vote, and necessarily invites the attitude of "if you aren't with us, you're against us". Granted there are always other candidates on the ballot that you could, theoretically, vote for president. But under the current system, voting for someone who isn't in one of the two major parties feels like throwing your vote away - if they guy has a cieling of 5% of the popular vote, then he's never getting elected, while in a proprtional system, such a party would get some voice.

Of all the different laws in different states, my favorite is in Nevada, where "none of the above" is a choice. During the Republican primaries for Senate in 2010, "None of the above" finished second.

 

Yes, with proportional representation, you get wider, well, representation, and less bipartisanism which is why I support it. But the downside is that you get people like Peter Dunne, and more famously, Winston Peters, who end up with a bit too much power, as they sit between the bigger parties and do deals to help the largest party achieve enough of a majority to govern.