In the run up to the 2019 General Election there was a lot of speculation about the potential impact of cold weather on voter turnout. This general election was the first outside of the spring/summer months since 1974. It was the first December election since 1918. If cold wintry weather put some groups of people off voting more than others, then it could have an impact in some marginal seats in a close election.

This Insight will compare factors like weather and daylight hours on election day to previous general elections. It will also consider the variations in these factors across the UK and the potential impact on overall turnout.

Weather on election day

The forecast for election day, 12 December, was for a frosty start in the North and East; milder, wet and windy in the West. Maximum predicted temperatures varied from 3oC in North East England to 11oC in the South West.

On 12 December across the UK:

  • the highest temperature was 11.3oC in the Isles of Scilly
  • the lowest temperature was -2.7oC in Frittenden (Kent)
  • Kirkwall (Orkney Islands) had the most sunshine with 3.3 hours
  • Port Ellen (Islay) had the most rain at 26.2 mm

During the day there was some snow across the tops of the Peak District, North Pennines and Southern Uplands.

Putting the 2019 weather in context

The Central England Temperature (CET) range on election day 2019 was 0.4‑7.9oC. This was the coldest minimum and lowest maximum for an election since February 1974. Ranges back to 1931 are given below.

This chart shows election day temperature ranges from October 1931 to December 2019.
Source: Met Office Hadley Centre, HadCET datasets

The October 1931 election was the only one in this series where the CET went below freezing.

According to the Met Office daily England and Wales precipitation series, average precipitation on election day 2019 was 14.32 mm. This makes it the wettest general election recorded on this series, which goes back to 1931. More rain fell on 12 December 2019 than on the previous ten general election days combined.

This bar chart shows wet weather on elections days from October 1931 to December 2019.
Source: Met Office Hadley Centre, HadUKP datasets

Daylight hours

The number of daylight hours is entirely predictable, unlike the weather. A December election day has relatively little daylight. This could, potentially, have an impact on turnout, especially alongside any wintry weather.

12 December 2019 was 10 days before the winter solstice – the shortest day of the year. This was in stark contrast with the 2017 General Election which was around two weeks before the summer solstice – the longest day of the year. With polls open for 15 hours from 7am to 10pm, some areas of the country had less than half of their total voting hours in daylight in 2019. The places with the longest days only had just over eight hours of daylight (sunrise to sunset).

In London there was just under eight hours of daylight on election day. This was less than half the length of daylight in the 2017 General Election. Sunrise in London was around 07.56am, 08.08am in Cardiff, 08.34am in Edinburgh and 08.36am in Belfast. The sun set at 3:51pm in London, more than six hours before the polls closed. Sunset was at 3:38pm in Edinburgh, 3.57pm in Belfast and 4.03pm in Cardiff.

Across the country sunrise was as early as 07.48am in Dover and as late as 08.44am in Londonderry. Penzance had one of the latest sunsets in the mainland at 4:19pm. The shortest day was in the Shetland Islands with less than six hours of daylight. The sun rose at just after 9am, two hours after the polls opened, and set just before 3pm or seven hours before the polls closed.

Turnout in the 2019 General Election was 67.3%, down from 68.8% in 2017. It was higher than turnout in the four elections between 2001 and 2015, despite the cold and wet weather and the lack of daylight on election day.


About the author: Paul Bolton is researcher at the House of Commons Library specialising in social and general statistics.

Photo: “Rail Rain” by mpardo.photo is licensed under CC0 1.0.