Wednesday, February 18, 2004

Post Auditor-General's Report Polls

The Ottawa media are salivating. After a decade of boredom the collapse of Liberal support in recent polls (two Ipsos-Reid, a National Post/Compas and a CROP poll in Quebec) has built expectations of a competitive election and, who knows, maybe even a change in government.

However, a key factor one would expect to find in this dramatic narrative is missing. The thing that strikes me about the latest polls is the weakness of Conservative support. Nationally, compared to the 2000 election the Conservatives are down from the combined Alliance-PC vote of 37.7% to 27% in the latest Ipsos Reid poll even though the Liberals themselves are down 5 points from 2000. Some analysts are obtusely refusing to aknowledge the reality that the Conservatives are just not as strong as some think they should be.

The Conservatives are weak for a number of reasons. First, they are dead in Quebec and not likely to revive. Quebec, however, was key to the last PC majority won in 1988. It also appears that, as the idea that the new party is a Reform-Alliance takeover sinks in, support in Atlantic Canada is weakening. The Conservatives registered 37% in the January Reid Poll but just 32% in the most recent poll. NDP support in 2000 in B.C. was unusually depressed by the unpopularity of the Glen Clark/ Ujjal Dosanjh government. The NDP now appears to have taken back a more traditional level of support there, probably mostly from the Conservatives.

The two parties that have gained support are the NDP and the Bloc. The NDP, in particular, are being overlooked partly because the media has spent years consigning the party to oblivion, before its phoenix-like rise from the ashes after electing Jack Layton as leader. It has risen from 8.5% in the 2000 election to 17% in the latest Reid poll, which I actually think understates the real level of NDP support. Reid started adding the Green Party to its ballot question about 18 months ago, because so many people were volunteering support for it. Since then Layton took over as NDP leader, started recruiting from the Green Party leadership, and adopting an aggressively Green agenda. I suspect many of the Green Party's adherents are younger voters who will probably gravitate to the NDP once an election is called. Add say 4 out of the 5 Green points to the NDP and suddenly its poll number is 21%, a level I think is more realistic and relativly high compared to recent history.

All this points at the moment to a Liberal minority with exceptionally strong showings by the NDP and the Bloc. But this muddies the media narrative and complicates the story.

Sunday, February 01, 2004

Polls and Media in American Politics

I don't normally think much of New York Times columnist David Brooks, but I like this column from Saturday January 31. I think he is absolutely right about the self-reinforcing tautological character of modern American politics. Here is an excerpt:

"Let us review the Democratic presidential primaries so far:

In the beginning, John Kerry surged to a big lead in the New Hampshire polls because he seemed so electable. He had plenty of experience, lots of money and big hair, and, as somebody said, he looks like an animatronic version of Abraham Lincoln. But then Howard Dean raised a lot of money, and New Hampshire voters figured that he was bringing so many new people into the process that he must be electable — and if he was electable, then they should probably support him because they wanted somebody who could beat George Bush.

So Dean's poll numbers rose, and the news media noticed his momentum, and other voters noticed how much great press he was getting. And that led to a self-reinforcing upward spiral of electability as more people concluded that he was electable because so many other people were concluding he was electable. People around the country saw that Dean was doing so well in New Hampshire they, too, concluded that he must be electable, a perception that led to an impressive rise in the national polls, which only enhanced his electability.

All this time, Kerry had not changed his views particularly, and he had not changed his campaign style, though he might have changed the bags under his eyes, depending on whom you ask. But savvy Democratic voters wanted to vote for somebody who could win the most votes in November, and they decided that since Dean was ahead of Kerry, therefore Kerry must be less electable, so voters moved away from Kerry. So Kerry's support plummeted, and the more his support plummeted the more he looked pathetically unelectable."

And so on. You get the drift. We see polls for candidates soar on news that they are doing well in the polls or have just won a primary.

The issue for Democrats this year is supposed to be the ability of a candidate to beat Bush, but this self-reinforcing process has happened before.

It well explains Howard Dean’s rise and fall. The one missing element in Brook's analysis is that Dean’s most serious problem was being the subject of numerous free media and ad attacks in Iowa (the ads were from Gephardt). This drove his negatives way up. Kerry, who was behind in the polls, was overlooked. He did have effective positive ads, (You can see them on John Kerry's web site ) but in effect he backed into his success because the two leading candidates hacked each other to death.

The process of success begetting success means that the lazy, superficial news reporting, combined with inattention on the part of voters, creates an extraordinarily unhealthy political process where success is built on a foundation of nothing more than prior success.

The media speaks frequently of “momentum”. What is momentum exactly? Is it not the self-reinforcing tautological process described by Brooks? Simply defining momentum reveals shallowness of the process.

More from Brooks:

“Suddenly Kerry, who had not changed his views particularly, nor his campaign style, began to see his poll numbers rise in Iowa because Dean seemed a little less electable. Then other Iowa voters began to notice the momentum behind Kerry, which made him look still more electable, so more voters decided that maybe Kerry was the man to support after all.

And, what do you know, Kerry won the Iowa caucuses, and from that moment on the election turned into a postmodernist literary critic's idea of heaven. It became an election about itself, with voters voting on the basis of who could win votes later on.

It's the tautology, stupid.”

There is more evidence of this process in this post I found on the blog Daily Kos

"As many of you know, through the use of Google News, I have been monitoring total media coverage of every candidate since August. During that time, three campaign stories received by far the most coverage. In order, those stories were:
1. John Kerry's victory in the Iowa Caucuses (1/19-1/24)
2. Al Gore's endorsement of Howard Dean (12/07-12/14)
3. Wesley Clark's entry into the race (9/16-9/26)

All three of these stories were very positive to the candidate in question. With the date of these three stories in mind, examine schwa's remarkable chart that tracks the five-poll moving average of national polls on the Democratic field. What you will see in schwa's chart is that Clark moved up around ten points at the same time positive stories about the General were swamping the media. Further, you will notice that Dean also saw a roughly ten-point upswing during the news cycle when he was endorsed by Gore. Now consider Kerry's national poll movement immediately after the Iowa caucuses:

Fox News
Kerry before Iowa: 7
Kerry After Iowa: 29

Newsweek
Kerry before Iowa: 11
Kerry after Iowa: 30

Q-poll
Before: 8
After: 30

Overall, Kerry saw a national bounce of around 20 points immediately after he won Iowa. Considering what appears to be a clear correlation between massive, positive news stories and huge swings in nationwide Democratic preference in all three of these cases, I draw the following, depressing conclusion: Democrats are Dittoheads who will do whatever the Political Opinion Complex tells them to do. "

It provides empirical evidence for Brooks’ analysis perfectly.

I think the most striking example of this process historically was the race for the Republican nomination in 2000. Then Texas Governor George Bush had the same name as his ex-President father. The name recognition drove up his polls which led to money, endorsements and even better polls. But the early numbers were probably based on simple confusion with Bush the elder. Would the U.S. and the Republican Party be better off today if the nominee in 2000 had been John McCain? We can never know, but a fairer, more reasonable process might have produced exactly that outcome.