Post from October, 2010

Watch Party Nov. 2nd!

Sunday, 31. October 2010 10:05

Just a reminder, the political scientists of the ATU History & Political Science department will be out at Ruby Tuesday’s in Russellville for a watch party as the election returns come in on November 2nd. We’re going to get started around 6pm and will go as long as we can (Close it down? Definate…maybe!). For those of you who can’t make it out, I will be live-blogging the election returns from the watch party so tune in on Nov. 2nd for up-to-the-minute election analysis. D.GOOCH

Category:PoliSciPundit | Comments Off on Watch Party Nov. 2nd! | Author:

Bad Analysis of Early Voting

Thursday, 28. October 2010 12:22

Here’s a recent article suggesting that the conventional wisdom — Republican wave on Nov. 2nd — might be overstated. While the premise is OK (there are alot of decent arguments the wave may be exaggerated in the MSM), the analysis of the data upon which the argument is based is particularly bad. Let’s take a look at this argument (a.k.a. wishful thinking):

There’s abundant evidence – often ignored or discounted by mainstream media outlets – that the organized base of the Democratic Party is stirring in ways that pollsters and pundits are likely to miss.

There’s nothing wrong with this point on its own. It is extremely difficult to identify the population of voters in any given election and there are all sorts of sampling and likely voter screen problems that can crop up to give us a distorted view of the electorate. Polling in the 2002 and 2004 elections, for example, suggested a larger bastion of Democratic support than actually showed up at the polls. But let’s turn to the evidence Mr. Lewis uses in his NY Daily article to posture against the apparent polling trends in 2010.

Republicans may be more fired up, but enthusiasm doesn’t win races. Early voting numbers show Democrats have actually cast more ballots than Republicans in Louisiana, Iowa, Maryland, North Carolina and West Virginia, according to Reuters. Although GOP early voters have the advantage in Colorado and Florida, the early voting doesn’t suggest a Republican blowout is in the works.

In West Virginia, for instance, Democratic turnout is 30 points higher than GOP numbers among early voters. That’s welcome news to Democratic Gov. Joe Manchin, (above) who is running against Republican businessman John Raese for the Senate seat held by the late Sen. Robert Byrd.

Enthusiasm doesn’t win races? Of course it does. A more enthusiastic base turns out more voters. A less enthusiastic base turns out fewer. Enthusiasm can be overrstated, especially when it comes to actual turnout, but let’s not get crazy here. As for the data, the partisan breakdown on early voting is a bad predictor if you’re not going to take account the registration advantage one party can have over the other in a state. For example, touting West Virginia early voting Democratic turnout over Republican turnout doesn’t account for the 2 to 1 Democratic advantage in registration in West Virginia. It also doesn’t tell us who those [mostly] conservative Democrats are voting for. Bad indicator.

Another big story that hasn’t drawn much notice is the role black voters will play. “There are more than a dozen Senate races and more than a dozen governor’s races where the black vote could make a difference,” says David Bositis, senior research associate at the Joint Center for Political and Economic Studies, a Washington-based think tank.

This isn’t a “big story.” It *would* be a big story if it happened, but the reason few people are talking about it is a good one – 2010 black turnout is going to lag significantly behind the 2008 black turnout. In 2008 we had a Democratic year, a presidential year (black turnout is always higher in presidential races than midterms), an historic black candidacy for president, an extremely unpopular Republican president, and a bad Republican presidential candidate. None of that is true in this election cycle. The notion that black turnout could meet or exceed 2008 levels is simply ridiculous. And note, no evidence is offered to suggest we should expect a large black turnout in 2010. Bositis just points to races where a big black turnout could matter. Well, a few extra touchdowns could turn the tide for the Bears on any given Sunday…but that hardly is reason to expect or predict them (much more likely outcome – Cutler throws pick 6).

Bositis says this year could end up looking like 1998, when Dems reversed half a century of history by picking up seats in the sixth year of a President’s term rather than losing any.

And I could win the lottery this year. Well, no I couldn’t, since I don’t play the lottery. Which means my chances of winning the lottery are slightly worse than the chances 2010 turns out like 1998. This simply illustrates extreme ignorance of the politics of midterms and the evidence aligned in favor of a Republican wave in 2010. You have to ignore current demographic characteristics of in-play Democrat-held seats, the sheer number of seats they are defending, and you have to turn a blind eye to every single poll, the polling trends, and historical trends to throw water on the Republican wave thesis…let alone the extreme wishful thinking of Democrats gaining seats in 2010. That’s not just being unconventional…it’s being silly. On par with predicting the end of the world in 2012 because that’s when the Mayan calander ends. Yeah, that’s a reason for the prediction…a really, really bad one.

I recall anti-wave arguments advanced by Republicans and conservatives prior to the 2006 midterm elections contra all the evidence suggesting a big Democratic win…and we saw how that turned out. Could the wave be overstated? Sure. Is it completely ephemeral? No, no, a thousand times no. Frankly, this doesn’t even qualify as wishful thinking. This isn’t thinking at all…it’s just wishing. Part of the problem is that the author is operating on bad assumptions:

– Republican rhetoric in 1998 was extreme and “grossly unfair”
– The public percieved Republicans as having “gone too far’ in 1998
– The public punished Republicans at the polls for this

None of that is true. Remember, the House didn’t impeach Clinton until December, i.e. after the election. The ‘attacks’ on Clinton were just as pointed in 1994 and especially in 1996, an election year. Clinton’s morality was an issue in 1992 when he ran and remained an omnipresent source of discontent with his presidency among conservatives. A majority of the public didn’t want Clinton impeached, what with the economy booming, but the notion these ‘attacks’ were seen as ‘grossly unfair’ is nothing more than the author of the article inserting his own feelings about the Clinton impeachment into the article under the guise of ‘many say.’ The reason Republicans didn’t pick up big gains in 1998 (the traditional ‘bad’ midterm for two term presidents) is that they’d already picked up most of those seats in 1994. The Republicans were a firm majority in 1998…and defending alot of centrist districts in that election. The “public rejects Republicans for extreme anti-Clinton attacks” meme was post-hoc rationalization of the election result. When, in fact, there were clear structural reasons for that election having been a bit of a Republican fizzle (from a historical perspective). Namely, Republicans had ‘maxed out’ on their possible seat wins in 1994 and there’s only one way to go from your theoretical maximum (that’s ‘down’ for those of you in Rio Linda).

Now, were Republicans at that maximum in 1998? No way to know for sure…but they were likely very close to it. Look at this chart from Cook Political Report. Republicans in 1996 held 57 Democrat-leaning seats (in other words, they held seats that voted for Clinton in 1992 & 1996). IOW, over 25% of their caucus were representing seats that had voted for the Democratic presidential candidate. Contrast that with 2008, where they are holding only 17 seats in Democratic territory, which represents only 9.5% of their caucus. To put it simply, Republicans had a much tougher task of defending their seats in 1998 than in 2010…mostly because they had a majority in 1998 and they’re a minority now.

There is no way the election on November 2nd is going to look anything like that of November of 1998. None.

Could something that dramatic be in the works this year – when, as in 1998, a Democratic President is under heavy fire from Republicans in ways that many see as grossly unfair? Don’t rule it out.

Rule it out. D.GOOCH

Category:PoliSciPundit | Comments Off on Bad Analysis of Early Voting | Author:

Vote for Me!…I have a puppy!

Thursday, 28. October 2010 11:01

Compelling argument, Joe Sestak! You win again, cuteness!

Category:PoliSciPundit | Comments Off on Vote for Me!…I have a puppy! | Author:

So you want a Ph.D. in Political Science…

Thursday, 28. October 2010 10:46

Oh, waaaaaay too much truth in this one:

Category:PoliSciPundit | Comments Off on So you want a Ph.D. in Political Science… | Author:

So you want to go to Law School…

Wednesday, 27. October 2010 18:42

Some racy language. You are forewarned.

Category:PoliSciPundit | Comments Off on So you want to go to Law School… | Author:

So you want to get a Ph.D….

Wednesday, 27. October 2010 18:26

Category:PoliSciPundit | Comments Off on So you want to get a Ph.D…. | Author:

Halloween-Themed Political Ad

Wednesday, 27. October 2010 18:20

Heh. Well, politics is often scary with a number of monsters and back-from-the-dead zombies wandering around. Why not?

Category:PoliSciPundit | Comments Off on Halloween-Themed Political Ad | Author:

Carter Debunked – Anderson Irrelevant in ’80

Tuesday, 26. October 2010 15:07

Jimmy Carter, the 39th president of the United States and loser to Ronald Reagan in 1980 recently claimed in an interview on Hardball with Chris Mathews that a “third party candidate” came into the election late and picked up “liberal democrats” and thus, given the fact Reagan only won with 51% of the vote, Reagan won “because” of the third party candidate.

Carter claims that Anderson split the Democratic party’s vote, leading directly to Reagan’s victory. But this claim is risible on its face given that Reagan won 489 electoral votes (to Carter’s 49 electoral votes). Even if we give every single vote of Anderson’s to Carter, Reagan still wins the popular vote by 2 million votes. But let’s consider the claim that Anderson’s vote was composed mostly of liberal Democrats (who we could have expected to go to Carter in a 2-man race) and then look at the vote and see what effect it could have had.

Using the 1980 ANES (American National Election Study), we can assess the relative approval of the candidates using feeling thermometers, which ask respondents to rate the person or organization on a scale of 0 – 100 (0 being ‘hate the guy’ and 100 being ‘love the guy’). Now, just because a respondent rates one candidate over the other on a feeling thermometer doesn’t necessarily mean that’s how they would have voted…but it is a reasonable proxy. Anderson got 6.6% of the vote in the actional election, while his vote is 8.8% of the sample vote from the ANES (could be random variation or some reflection of Duverger’s Law). Those percentages are fairly comprable, so we should be able to use the 1980 ANES sample vote as a good proxy for the actual election results.

First lets look at how the partisans viewed the candidates. Here is average thermometer score for the three presidential candidates among Republicans. Remember, it is on a scale of 0-100 with a higher score meaning they like that candidate better:


Unsurprisingly, Republicans liked Reagan a lot more than they liked Carter. But they also liked Anderson more than they liked Carter, which at least suggests they didn’t view Anderson as a further Left choice than Carter. Let’s look at Democrats.


So Democrats tended to like Reagan a bit more than Republicans did Carter…and they liked Carter a little less than Republicans liked Reagan. The opinion of Anderson, however, is pretty much the same. That suggests that Carter’s problem in 1980 was less about Anderson and more about his support among his own partisans relative to that enjoyed by Reagan and his fellow Republicans. Of course, this is among the full sample of partisans and does not screen based on voting. Furthermore, the Anderson feeling thermometer is a bit suspect since respondents tend to pick ‘50’ when they don’t recognize a person they’re asked to rate rather than admitting they don’t know or have no opinion. So there will be a strong central tendency in the relatively unknown Anderson’s feeling thermometer due to voter ignorance.

So let’s, instead, look at those in the ANES who actually voted for Anderson and who they might have preferred if it had just been a two man race between Reagan and Carter.


As we can see, there is some support here for Carter’s supposition that Anderson voters were more predisposed towards him than Reagan. On average, Anderson voters liked Carter nearly 8 thermometer points better than Reagan. Let’s look at the individual scores. Remember, these tables report the difference between the Reagan thermometer and the Carter thermometer, so a positive number indicates they liked Reagan better than Carter and vice versa for a negative number. A zero means they liked both candidates equally.


So we can see at the tails t here are just as many Anderson voters who hated Carter as Anderson voters who hated Reagan. But as we move inward we find significantly more Anderson voters who seem to have liked Carter a little bit to a substantial bit more than Reagan. The largest frequency of Anderson voters fall in the -20 range, which is a fairly large percentage who liked Carter 20 total points (1/5th of the scale) more than they liked Reagan (1.3 in 10 Anderson voters). However, the next largest group is the ‘0’ group, where the Anderson voters liked each candidate equally well. If we assume that Anderson voters, forced to make a choice between Reagan and Carter, would have chosen the candidate they liked better, then the vote distribution between Reagan and Carter among Anderson voters looks like this:


So Carter certainly wins the Anderson vote by a wide margin, lending credence to his argument that Anderson’s third party bid did damage to his own vote totals, but was it enough to be decisive in the 1980 election, as Carter argued on Hardball? Hardly. If we distribute Anderson votes accordingly we get this popular vote result for the candidates:

Anderson Voters Distributed to Reagan & Carter Per ANES Percentages

Reagan wins the popular vote by a little more than 8 points, and I don’t believe any presidential candidate in history has lost the Electoral College when they had that large a spread in the popular vote. That said, let’s look at the close states that Reagan won and see if that incremental improvement for Carter might have made the difference in those close states. Anderson got 6.6 percent of the vote. If we distribute his vote evenly (assuming that his vote percentage would have been distributed evenly across the states), then Carter would pick up 1.55% points across the board. Note, the assumption may be faulty as there may be fewer Anderson voters in the states that Reagan won. So this assumption will tend to err on the side of Carter’s thesis if Anderson ran behind is mean vote in the state..and work against it if Anderson ran ahead in that state. If we looked at Anderson’s real percentages in each of these states rather than his overall percentage…some of these states might stay with Reagan rather than flip. I’ll look at that next. For now, we’ll stick with that assumption for the ‘close’ states.

Electoral College Results w/ Anderson Adjustment to State % – Close States

So adding in the Anderson adjustment does flip a few states in Carter’s direction. Assuming the Anderson vote was distributed evenly, Carter would pick up 7 more states. But those 7 states account for only 63 electoral votes. That would still have resulted in a Reagan landslide: Reagan – 426 electoral votes / Carter – 112 electoral votes. Indeed, if we just give Carter all the close Reagan states, the maximum number of electoral votes he could have picked up was 135, which still is a 354 / 184 crushing defeat for Carter.

It’s even worse for Carter if we use the actual Anderson per-state percentages. Below is a table that shows the states that Reagan won where the Anderson percentage of the vote was larger than the difference between Reagan and Carter’s percentage of the vote. I’ve included the actual Reagan, Carter, and Anderson percentages from 1980. I then apply the percentage of the Anderson vote we would give to Reagan & Carter based on the 1980 ANES feeling thermometers (the Anderson adjustment) to that percentage. The next two columns are the Reagan and Carter percentages with the Anderson adjustment added in. If it flips the vote in favor of Carter, I award him that states electoral votes (last column).

Electoral College Results w/ Anderson Adjustment to Per-State Percentages

As you can see, applying the Anderson adjustment, Carter only picks up two of the states that Reagan won in 1980 –- Massachusetts and Tennessee — for a total of 24 additional electoral votes. However you do it, Reagan wins in a landslide every time.

In sum: Carter is wrong. Anderson had no effect on the result of the 1980 election. We don’t even have to bother considering the countervailing effect Ed Carter, the Libertarian candidate, might have had in pulling votes away from Reagan. While Anderson undoubtedly pulled some votes away from Carter and might have cost him a few states (making Carter’s loss a little less embarrassing), it would have in no way changed who was president in January of 1981. That was going to be Ronald Reagan under every conceivable scenario.

Category:PoliSciPundit | Comments Off on Carter Debunked – Anderson Irrelevant in ’80 | Author:

Courier Coverage of the H&PS Department

Saturday, 23. October 2010 17:20

The full article for the Goldberg talk is cached here entitled “Top Author to Talk at Tech” (quite alliterative!).

And an article mentioning my and Dr. Housenick’s contributions in helping the Courier conduct a scientific poll of the Russellville mayoral race can be found here and here.

Category:PoliSciPundit | Comments Off on Courier Coverage of the H&PS Department | Author:

Goldberg Lecture

Tuesday, 19. October 2010 10:36

Well, the Goldberg lecture was last evening, and I think it went quite well and so did Jonah. While the audience was a bit muted, everyone seemed to be paying attention and there were some very good and insightful questions asked during the Q&A session. Goldberg’s argument on the progressive movement’s flirtation with fascim in the early part of the 20th century was both intriguing and provocative. I thought his best point came towards the end of his talk, noting the distinct intellectual and philosophical differences between the individualistic Lockian vision of man, the state of nature, society, etc. and that of the Rousseauian vision of the pursuit of the greater good and the subserviance of the individual to the general will. His point that this division “cuts through the human heart” and that we both want to establish individual identities and belong to something greater than us was a good one and quite illuminating of the political differences that exist in society today. I further thought his Hartizan (though he didn’t mention Hartz explicitly) assertion that America is culturally predisposed towards a Lockian view of government was both astute and correct. While Hartz meant it as a criticism of our culture and Goldberg finds it praiseworthy, that both acknowledge this essential and ‘exceptional’ character of the American identity underscores it as a fundamental truth about American politics. All in all, a very good night. D.GOOCH

Category:PoliSciPundit | Comments Off on Goldberg Lecture | Author:

%d bloggers like this: