jackman.stanford.edu/blog
• 114th U.S. Senate
• ideal point estimates svg pdf csv 2/27/15
• scatterplot against 2012 Obama vote share svg pdf
• roll call object: RData
• 114th U.S. House
• ideal point estimates svg pdf csv 2/27/15
• scatterplot vs Obama vote share svg pdf
• roll call object: RData
• Bayesian Analysis for the Social Sciences Wiley; Amazon; errata as of 5/23/13

## Monday October 15, 2012

Filed under: politics,statistics — jackman @ 6:52 pm

Re-posting what just appeared on HuffPost:

On the eve of the 2nd presidential debate, it is helpful to take stock of what current polling is indicating about the campaign.

Romneyâ€™s debate bounce – or rather, Obamaâ€™s debate slump – appears to be over. National-level polling appears to have bottomed out for Obama. Further, once we integrate the information about the national trend contained in the state-level polls, it appears that Obamaâ€™s post-debate slump was not as large as it might have seemed at the time (and Iâ€™m talking to you, Andrew Sullivan). Indeed, with the benefit of only a little hindsight (and only a little modeling) it appears that only about 1.5 percentage points separate Obamaâ€™s late September, national peak from his early October trough, with a recovery of a few tenths of a percentage point over the last 7-10 days.

Indeed, at the national level, it appears that even in the immediate aftermath of the 1st debate, Obama never trailed Romney, once the information in the state level polls is integrated with the national-level polling.

All the same, the national race is unquestionably tighter now than it was. Obamaâ€™s lead in the national track produced by my model is about half a percentage point. In late September this lead was four percentage points.

The swing state picture has generally remained better for Obama. In the picture below I look at the the national-level estimate of Obamaâ€™s two-party share, along with results from 9 battleground states, all won by Obama in 2008. A useful reference point is Obamaâ€™s two-party share in 2008, appearing on the left of the graph. In the middle of the graph I plot the estimates from my model using polls fielded prior to the 1st debate.

The general fall in Obama support between 2008 and where the campaign stood on the eve of the 1st debate is clear (the pattern of lines sloping down, left to right). Only in Ohio was Obama generally outperforming his 2008 result.

Since the debate, Obamaâ€™s national vote share has slipped by roughly the same amount it had fallen between the 2008 election and where it stood on the eve of the debate: from 53.7% to 52.2% (pre-debate) to 50.4% (now). Post-debate slumps of a similar magnitude appear across the nine battleground states on the graph. North Carolina now lies on 48.5% (Obamaâ€™s share of two-party vote intentions). Florida had been resisting the national trend away from Obama prior to the debate, but has also fallen below 50%. Virginia is at 49.9%. At this stage, only Indiana, North Carolina and Florida look like they will be picked up by Romney, which wonâ€™t win Romney the election.

Obama won Ohio with 52.3% of the two-party vote in 2008. Pre-debate, I estimated Obama two-party support there at 53.4%. Today that figure is 51.5%, with the falls relative to both the 2008 result and the pre-debate estimate the lowest among the swing states. Relative to 2008, Obama is off 0.8 percentage points in Ohio; the corresponding numbers are 1.7 in North Carolina, 2.1 in Florida, 2.9 in Iowa, 3.3 in Virginia (the same as the national-level swing), 3.9 in Colorado, 5.1 in Wisconsin, and 5.3 in Nevada.

So long as Obamaâ€™s Ohio numbers hold up, so too will his chances of winning the Electoral College, and thus far, they have. Music to the ears of owners of Ohioâ€™s television stations…

## Tuesday October 9, 2012

Filed under: Australian Politics — jackman @ 9:28 pm

I’ve been amazed by the number of friends around the world asking me if I’d seen Gillard responding to Abbott in the Australian parliament.

From one friend (little doubt about her political leanings, but you get the idea):

“Simon, wish you were here, would love to here your commentary. Abbott is a disgrace, and Julie Bishop is a fool, her/the liberals response this morning was one of the unprofessional demonstrations in the history of politics.”

Yes I have seen it, even before everyone started emailing me…! And yes, my computer programs watching the betting markets alerted me that something was up back in Oz (I’ve been paying a bit more attention to the US polling situation of late).

Gillard did a great job, in several respects. She pivoted the debate about Slipper (and what a wretched mess that is) right back onto Abbott. It was a speech that will resonate with many who have watched Australian politics of late and wondered “WTF”. Finally Gillard addressed head-on the bile that she’s had to deal with, speaking plainly, powerfully and from the heart. I think she spoke in a way that will resonate with many women in particular, undoing any damage control Abbott might have engineered in the wake of the Jones matter. It could well be the most “Prime Ministerial” moment I’ve seen from her, in no small measure because it was incredibly personal as well.

Australian politics has been pretty unstable of late, but I’d predict that this shores up any lingering doubts that Gillard will lead the ALP to the next election; she might have even bought herself a point or two or three in the polls.

Slipper did Gillard a great favor by letting the showdown with Abbott dominate the 6pm news, delaying the news of his resignation until after 7pm (as best as I understand it from this distance).

Abbott of course showed incredibly poor taste and a lack of discipline in the “government dying of shame” line. Little surprise that this morning in Australia the Opposition was keen to re-frame the matter as about Slipper.

Finally… My poll tracking in the United States right now has the national story all tied up at 50-50. Obama’s had some bad polling in the wake of a poor debate against Romney last week. In the space of a week a 4.5 point lead in the polls evaporated. It does make you wonder what damage Abbott and Co. have done to themselves over-reaching on this Slipper matter, creating some space for Gillard to perform an amazing piece of political jujitsu.

### meanwhile, in Australian politics…too much intrigue for the bookies

Filed under: Australian Politics — jackman @ 4:19 pm

Centrebet and sportingbet moved in lock step around mid-day yesterday, on the back of a 54-46 Newspoll.

Both sites took their Australian Federal betting markets down last night just after 7pm, when the news re Peter Slipper’s resignation broke. They’re yet to come back up. You’ve got to conclude that they’re worried about being clobbered by insiders when things are swirling in Canberra. Thumbnail below is clickable.

Sportsbet’s Labor price eased out to $3.25 from$3.00 (Coalition $1.33 in to$1.28) between 9 and 10 this morning, but they stayed up overnight.

Implied probability of a Labor win is about 25% at the last Centrebet prices ($3.72 to$1.27).