So who came closest?Analysing Poll Accuracy in the 2024 Mid Sussex Election

PollLiberal DemocratConservativeLabourReform UKGreenTOTAL DIFFERENCE
ELECTION RESULT39.627.117.611.13.8N/A
Yougov4324.517.411.33.76.5
Electoral Calculus45.325.1188.33.810.9
New Statesman31.524.518.714.85.417.1
Electionmapsuk33.22421.615519.3
More in Common30282410620.2
Financial Times3124.224.316.23.124
IPSOS25302811529.2
Survation25.129.329.810.34.830.7

At Mid Sussex Election Watch we carefully monitored the pre-election opinion polls to get a sense of how the election might turn out in our constituency. Now that the election has happened we thought it might be interesting to go back and look at just how accurate these polls were. We looked at the difference between the actual result and the prediction for each party in each poll, and then calculated the total difference across all the parties for each poll. This proved to be very interesting.

Among the various polls analyzed, YouGov emerged as the most accurate, with a total difference of just 6.5 percentage points across all parties. This poll was particularly spot-on with its predictions for Labour, Reform UK, and the Green Party, with minimal deviations from the actual results. Such accuracy reinforces YouGov’s reputation as a reliable source for gauging public sentiment.

In second place was Electoral Calculus, which had a total difference of 10.9 percentage points. While slightly less precise than YouGov, Electoral Calculus still provided relatively close estimates, especially for the Liberal Democrats and the Green Party.

The other polls showed more significant discrepancies. For instance, More in Common and Financial Times had larger total differences of 20.2 and 24 points, respectively. These polls overestimated support for Labour and underestimated the Liberal Democrat vote share, highlighting the challenges pollsters face in capturing the shifting dynamics of voter behavior.

Interestingly, Survation and IPSOS were the least accurate, with total differences of 30.7 and 29.2 points, respectively. These polls significantly underestimated the Liberal Democrat performance while overestimating Labour’s share. Such variances underscore the inherent uncertainty in polling, where methodological differences and sampling errors can lead to varied results. Funnily enough, it was the Survation poll that Labour consistently chose to use in their election material. (One of their leaflets was headlined, “Poll after poll shows Lib Dems a distant third in Mid Sussex,” without mentioning that actually it was several versions of the same poll that made this prediction…!).

Election Maps UK provided a new perspective with a total difference of 19.3 points. While not the closest, it offered a fairly accurate prediction compared to other polls, highlighting the evolving landscape of polling methodologies.

Conclusion

Yougov has a reputation for accuracy, and in future elections it is certainly worth reflecting their data. Electoral Calculus is also worth a look. As for the others… well, perhaps take their data with a large pinch of salt!

Leave a Reply