The UK general election 2005 poll tracker follows public voting intention as gathered by six polling companies from 1 January 2005 to election day.
In addition, key events that may have a bearing on who people say they will vote for are shown in a box below the graph. These include policy pledges by the main parties and news events.
Analysis from BBC Political Research Editor David Cowling also appears in the box.
How do I use the poll tracker?
Each point on the graph represents the results of a particular poll.
Mouse over the points to see the standing of each party at this date. Further details of who commissioned the poll, when the fieldwork was conducted, and the sample size appears in a bar at the top.
Fieldwork dates are especially important as a poll that is published after a key event - such as the Budget - may have been conducted before the event itself.
Use the tabs at the top to switch between polling companies.
To see the key events, drag the slider cursor along the line to the square buttons with blue arrowheads on them. Each button below the slider bar represents a key event or a piece of analysis.
You can also click on the buttons themselves or use the arrows at each end of the slider bar to move the cursor.
Why is one tab already selected when I open poll tracker?
The poll tracker defaults to the latest published poll and so will change each time a new poll is published.
If several polls are published on one day we will rotate alphabetically. So the first time it happens poll tracker will default to ICM (if they have released a poll), the next time it will be MORI etc.
I don't have Flash / I use a screen reader
Poll tracker is designed for use with the Macromedia Flash application, however, a non-Flash version with the same information (in text) is available and should automatically appear if your machine does not have Flash.
How do I know if these polls are representative?
Polling companies use a variety of methods to try to ensure their results are correct to within an acceptable margin of error (+/- 3% for polls of 1,000 people; +/- 2% for polls of 2,000 people).
More details can be found here: Methodology guide
Which sample size do you show?
In line with convention, we show the total sample size - that is the total number people interviewed.
However, because polling companies weight and filter their data, their final percentages are based on fewer people's answers than the total sample (see the methodology guide for more details.)
This has particular relevance for MORI who filter their data so that only those who say they are absolutely certain to vote are included.
Why don't the percentages add up to 100
This does occasionally occur due to rounding of the figures by the polling companies. Sometimes a poll will add up to 99% or 101% but it will never be more than 1% away from 100%.
Why hasn't poll tracker updated?
There is likely to be a new poll most (but not all) days during the election campaign. These will be added to poll tracker the morning they are released.
If you think poll tracker looks out of date on your machine, try pressing CTRL R. This should refresh the Flash file, forcing the latest version to load.
This may be necessary if your machine or network is caching the old version of poll tracker.
What does "Times (tracker)" mean?
Populus is running a daily tracking poll for the Times in which a quarter of the sample is replaced each day. These polls are carried out separately from Populus' other polls for the Times.
Although the newspaper is publishing the results daily, the BBC is only reflecting the results each time there is a totally new sample.
On Sunday 24 April there was a new sample but there was no edition of the Times that day (the Sunday Times is a separate publication).
Therefore poll tracker is showing the results from Monday 25 April, the first time since the tracking poll began (on 20 April) that a totally new sample was published in the newspaper.