Methodology - "Weighted Swing"
A Uniform Swing calculates the national support change and applies this to each constituency, you guessed it, uniformly.
A Weighted Swing also calculates the national support change, but instead of using this "as is" applies a weighting to it per constituency. This is based on several factors explored below.
We currently hold the results of each constituency from the 2017 General Election. We use these results as a baseline to which we apply our methodology.
ChUK (Change UK) - The baseline for the ChUK is currently set as the projected values of the average swing from each party. For example, if the support for the ChUK stands at 7%, with 2% from LDem, 3% Lab, 1% Con and 1% Oth, a projection will be made with these numbers to get the predicited support per constituency. This will be used as the baseline.
BRX (Brexit Party) - This Baseline is calculated in much the same way as ChUK, but as you would expect a lot more support comes from Conservative and UKIP voters, with less support coming from other parties.
As the average is constantly changing with each new poll that is released, our baseline changes also. The more polls released, the less volatile this baseline will be.
The Weighting is calculated using various different constituency specific statistics. I am not going to give out the exact way I calculate the weighting, but the statistics and amount each statistic effects are as follows. Note that weight effect will differ for each party.
|Statistic||Effect on Weight|
|Age Group Density||High|
|Relative Strength of Seat||Medium|
|Swing Volatility (2015, 2017)||Medium|
The equation for each Constituency will then look something like this:
CV = Base +/- (NS * Weight)
CV = Constituency Vote
Base = Baseline
NS = National Swing
Weight = *See Above