We have examined the YouGov opinion poll results since the start of 2015 to assess whether any discernible movement in party support can be determined.
In the analysis below we have undertaken two pieces of statistical modelling on the opinion poll values for each of the five main UK parties. First we have performed a Wald-Wolfowitz test to explore whether the poll results are randomly distributed or alternatively follow a particular pattern (i.e. rise or fall over time). After this we build a simple OLS regression model on the polling data to see if any trends are significant.
The first set of results are below.
Click to access random-test-unmitigated.pdf
Download – Feb 2015 – Testing Party Support Movement – Unmitigated
The table below summarises the data.
| Party | Random Data? | Linear Trend at 90% Significance Level | Linear Trend at 95% Significance Level |
| Conservative | No | Yes – Upwards | No |
| Labour | No | Yes – Upwards | No |
| Liberal Democrat | No | No | No |
| UKIP | No | Yes – Downwards | No |
| Greens | No | Yes – Downwards | No |
When we calculate Cook’s D for each of the observations we can see that there are some clear outliers in the polling data for each party. If we remove these outliers (we expect at least one in twenty poll results to be beyond the margin of error expected for sampling the full population) we can repeat the exercise above, calculating whether the data is random and if there is a discernible linear trend.
Click to access random-test-mitigated.pdf
Download – Feb 2015 – Testing Party Support Movement – Mitigated for outliers
Again, we summarise the results.
| Party | Random Data? | Linear Trend at 90% Significance Level | Linear Trend at 95% Significance Level |
| Conservative | No | No | No |
| Labour | No | No | No |
| Liberal Democrat | No | No | No |
| UKIP | No | Yes – Downwards | Yes – Downwards |
| Greens | No | No | No |
Once we remove outliers we see that the trends in the support for the Conservatives, Labour and the Greens that we saw previously disappear. As a contrast, it becomes more than clear that the movement in support for UKIP is clearly downward and statistically significant, moving from 15.5% at the start of the year to almost 14% at the time of writing.
Conclusion
As we have previously indicated, there is little evidence that support for the main UK parties is moving ahead of the May 2015 vote in a similar manner to previous elections (broadly, we would expect Conservative and Liberal Democrat support to rise and Labour support to fall). There is however clear evidence that support for UKIP is declining, and that if the decline continues at it’s present level then UKIP will poll around 12%-12.5% come May the 7th (assuming that the YouGov poll rating for them is normally distributed around the actual level of support in the whole population).
Improvements
There are a number of improvements that could be implemented on this analysis.
- We have only assessed the fit of a linear trend. It is possible that a polynomial or other trend is present in the polling data for each of the main UK parties.
- There were not enough data items in the set of YouGov polling data since the start of 2015 to conduct a Kolmogorov-Smirnov test to assess randomness. It might also be appropriate to conduct an Augmented Dickey-Fuller test on the data (more appropriate for a time series set), but again sample size limitations make this difficult.