Polling Clarity

The Tory growth shows itself

2008-06-05T06:21:00.000-07:00

Some more polls are in and the Tory growth that the media love to talk about now seems real. I still think its not as dramatic as the media like to make out, but there does seem to be a clear upturn in their fortunes over the last few months. I'd say about four months or so looking at the most recent trend.

What the media are saying about Labour does seem to be true. They took a real battering in the last month. Just look at that dip at the start of May!

The Lib Dems are continuing their trend. It looks to me like seven straight months of growth. Slow growth, but definite growth. No sharp increases or decreases, just slow, consistent, growth.

Here is the same chart again, with the recent growth and decline marked clearly:

STILL no Tory progress. Have they maxed out?

2008-05-02T05:49:00.000-07:00

By all accounts, last night was a bad night for Labour. But while the Tories picked up seats, we shouldn't be fooled in to thinking that this represents a surge in Tory support. National trends show that they have in fact remained at about 40% in the polls since the Brown Bounce came down to land in October.

Labour support has taken a slide recently, that's for sure. The slide below clearly shows the brown bounce, the return to Labour support at the level it was during the final Blair months and a recent dip BELOW that level. A new low for Brown. But what is next?

You can also see where the Labour supporters went: Lib Dems and 'other'. Lib Dems are just back to their long-term average of 15-20%, though. Nothing too exciting there. And nothing too exciting in the resurgence of 'other', either. Just a return to historic levels of around 12%.

So all in all the most interesting thing about the current situation seems to be that the Tories just can't get above that 40% barrier. Even when Labour are struggling in the press, in policy and with voters, the Tories just aren't gaining.

Still no Tory progress

2007-12-13T08:57:00.000-08:00

Another week and another batch of polling. Despite a number of catchy speeches by Cameron in the last few weeks, he has failed to lift the Tories' standing in the polls since mid-October.

That isn't to say that Labour are doing any better. Their slide looks to have slowed a little in the last week, but there isn't any sign of improvement. Perhaps both Labour and Tories have found their natural extreme levels and that it was only undecideds that gave Brown his bump, and then took it away?

What goes up must come down

2007-11-28T14:45:00.001-08:00

On the face of it, the most recent batch of polls show quite a slide for Labour. They were at 42% a couple of months ago and now find themselves down at 30%. Not good. Interestingly, the recipients of those points (they must go somewhere) isn't the Tories. They have been on about 40% for just over a month now. In the same period Labour have shed seven points. The Lib Dems and 'other' have done nicely in recent weeks though.
If we look at the recent Labour slide another way, it just seems like the pattern you get when you throw a ball in the air. What goes up must come down.

The question is, will it go down below the level Brown inherited it at, or are we just back to the pre-Brown levels for a while?

Interestingly, you can also see a similar pattern in the Tories' numbers. A period of stability at the end of Blair's time followed by a dip as Brown rode his bounce, and then a return to the pre-Brown levels.

2007's third phase

2007-11-05T12:40:00.000-08:00

Recent polls and a slight change in methodology bring us a new picture to look at today: a clear third phase of the year. Labour's decline is more than a blip. It is persisting.

Labour have now experienced a consistent slide in the polls for almost the last six weeks straight. This translates to a slide of just over a percentage point per week!

The Tories however, seem to have halted their rise about three weeks ago. The last three weeks seems to show a flattening and if anything a slight decline in support. They capitalised on Brown's election fumble to rise by about eight points in just three weeks, but now they seem to have lost momentum. Perhaps it was more about Brown's fumble than about a real increase in Tory support?

The question is: what will the future look like? Will we call this third period a 'Cameron bump', or

That leaves us with a gap in the last three weeks. Labour are still sliding and the Tories are stable, so what is going on? The reason for the seeming contradiction is that the Lib Dems have picked up a few points in the last couple of weeks. After a steady gradual slide for pretty much the whole year so far, and a five point nose-dive over the last month, they seem to be on the up just a little in the last couple of weeks.

Here is the same chart again, with a labels to help tell the story:

[FYI, here is the chart as it would have looked using the old methodology:

Crossing paths. An update

2007-10-15T14:28:00.001-07:00

I added some recent polls and filled in a few more from the last weeks that weren't already included. Looks like Labour and the Tories are crossing paths, and not just in policy terms.
And it looks like the poor old Lib Dems are accelerating their long term path of decline. Will Sir Menzies Campbell's resignation help or hinder them?

The raw data I gathered is here. A lot of my new data came from Googling recent polls, but I also grabbed some from Polling Report, who seem to update their polling data more frequently and more comprehensibly than my original source. Thanks!

Remember that my chart lags real polls by a little bit or tries to pick up the underlying trend without reacting too much to noise, whichever way you want to describe it :). Look at the detail of the chart if you want to work out what you think is really going on. My line is just a suggestion!

Lets step back. What a view!

2007-10-10T00:32:00.000-07:00

Since we are no longer preparing for an election, I took the chance to step back a little and look at the trend as far back as my data went: 1992. (Thanks again to IpsosMori for the data!).

I was more than a little surprised at quite how much of a lead Labour had over such a long period! (Which makes me want to talk about spending / leaking political capital, but I will refrain. This isn't a political blog!)There are a couple of solid trends that stand out to me. First, Labour have been on a pretty consistant and fast-paced decline since '95. There were a couple of spikes and dips along the way, but thats a pretty consistent eroding of public support. The 'Brown Bounce' really shakes that decline, but will that turn out to be a bounce or just another bump?

The other consistent trend is that the Tories have been on a constant, but slow rise in popularity over the same period. Again, there have been some bumps along the road, but thats a pretty solid, if slow, trend.

[I did a lot of date cleaning and data sorting to be able to generate the trend line and chart, above. The raw data is available from the link to IpsosMori, but the cleaned version (with the points in my trend line) is available here:
http://spreadsheets.google.com/pub?key=pOx1WcdWcFQWWwtF_xV8CHA&output=html]

Lets get started

2007-10-05T12:37:00.000-07:00

Being both frustrated by the lack of decent presentation of polling data, and a person that doesn't like to complain about something without doing something about it, may I present my chart of all the polls I could get my hands on.

I use exponential smoothing to reduce the noisy pattern you would get if you drew a line between every poll that came out, and try to pick up just the underlying pattern. I am inspired by pollster.com who has been using this technique for some time in the US and whose website you should definitely go and visit!

It does of course make a difference how reactive I make the trend line. I can set it to zig zag around and follow all of the latest trends, but this would be a mess and would miss the point. Instead I have found what I think is a good reactivity mix of taking some of the new information that an additional poll provides, and yet having regard for the other recent polls, and so trying to represent the real underlying trend. What do you think?

[Note: re-posted the chart on 6th Oct with just a little smarter maths, removing some of the kinks caused by two polls on the same day]