iOS app Android app

Simon Jackman

Entries by Simon Jackman

Pollster Predictive Performance, 51 out of 51

(30) Comments | Posted November 7, 2012 | 9:07 AM

The morning after a good night for the predictions we made with my poll-averaging model...

The headline: If Florida stays in the Obama column, we will have called each state correctly: 51 out of 51, and the Obama Electoral College count bang on 332.

Read Post

Pollster Predictions: 91.4% Chance Obama Wins, 303 or 332 EVs

(926) Comments | Posted November 6, 2012 | 7:30 AM

We might pick up a few very late polls this election morning, but even so, it is well and truly time to list some forecasts.

National popular vote. The last run of the model tonight produced 50.8% two party vote share for Obama. Since this is a two party...

Read Post

Chart Dump!: Swings, Probabilities, Electoral College Counts, Raw Data

(44) Comments | Posted November 5, 2012 | 1:31 PM

An election eve chart wrap... data is still coming in, and the models are still updating, but here's where things look this morning (Pacific time).

Anticipated swing from 2008, two-party terms. The left column of data shows Obama's two-party vote share by state in 2008; the right column shows the...

Read Post

Obama Ticks Up: Popular Vote, Swing States and Electoral College Forecast

(70) Comments | Posted November 3, 2012 | 7:35 PM

Polls released in the second half of the week generally suggested a slight uptick in Obama support, in the national popular vote, battleground states and the Electoral College forecast.

Recent runs of my poll-averaging model show Obama faring slightly better in the national popular vote. Obama has a 0.5 percentage...

Read Post

Never Mind the Models, Here's Some Data

(91) Comments | Posted October 30, 2012 | 10:03 PM

So, at the end of a day when Twitter was alive with the sound of pundits railing against poll averages (well, one poll average or even one poll averager in particular), I thought I'd try this.

I've simply taken every poll in nine battleground states over the last...

Read Post

Converting a Poll Average to a Forecast

(116) Comments | Posted October 30, 2012 | 1:44 PM

The poll-averaging model I've developed for HuffPost generates estimates of Obama and Romney vote share in each state, and for the national level popular vote. The model also produces probabilities of an Obama lead in each state and in the national popular vote.

As we enter the final...

Read Post

House Effects, Back By Popular Demand

(32) Comments | Posted October 24, 2012 | 11:08 AM

The graph below shows point estimates of the survey house effects produced by my model (as of the evening of October 23, 2012). Negative house effects means that the pollster skews in a pro-Romney direction; positive house effects means that the pollster tends to overstate support for Obama.


Read Post

Close Votes and the Electoral College

(40) Comments | Posted October 21, 2012 | 3:22 PM

My model-based poll average points to a close result in the national popular vote. The national popular vote has been indistinguishable from 50-50 since the first debate, with hints of a modest shift back towards Obama when recent polls are added to the modeling.

At the same time, the model...

Read Post

All About the Swing: 2008, the First Debate and Now

(0) Comments | Posted October 15, 2012 | 8:59 PM

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

Romney's debate bounce -- or rather, Obama's debate slump -- appears to be over. National-level polling appears to have bottomed out for Obama. Further, once we integrate...

Read Post

Romney's Uphill Battle

(32) Comments | Posted October 5, 2012 | 12:23 PM

I thought it would be helpful to give a big picture review of the state of play before we're starting to see polls with post-debate field periods.

The graph below compares two-party vote share for Obama in 2008, by state, with last night's estimates from our model-based poll averaging. Pre-debate...

Read Post

Skewed Polls Nonsense

(130) Comments | Posted September 28, 2012 | 2:22 PM

It has been largely one-way traffic in national and battleground state polling in the last several weeks. My model-based poll average puts Obama's lead over Romney at just over four percentage points, nationally. Indiana looks like the only 2008 Obama state that is an...

Read Post

Model-Based Poll Averaging: How Do We Do It?

(106) Comments | Posted September 14, 2012 | 8:58 AM

Over the remainder of the election campaign, HuffPost will be presenting graphical summaries of the modeling I've been doing behind-the-scenes for them over the last few months. Some additional background on how it all works seems warranted.

The goal is to summarize the "state of play" in states where we...

Read Post

Obama's Biggest Lead Over Romney Since March

(10342) Comments | Posted September 9, 2012 | 4:23 PM

We're now 60 hours or so past the Democratic convention. It is now clear that the Democrats got the better of the convention bounces. Indeed, recent national polls suggest that the Democrats' bounce not only erased some tenuous gains by Romney, but is propelling Obama into a commanding lead. In...

Read Post

Ryan Plus One Week: No Bounce

(430) Comments | Posted August 20, 2012 | 8:26 PM

Since presumptive Republican nominee Mitt Romney announced his choice of Paul Ryan as a vice-presidential running mate, HuffPost Pollster has logged five national polls and 11 state-level polls, all fielded since the news broke. The bottom line? No real change in Obama-Romney vote preference, at least the national level.


Read Post