What We Can Learn From Our Latest MRP (and what we can’t)

Survation’s latest MRP analysis of a poll of 15,029 residents in Great Britain conducted on behalf of Best for Britain, and published last night in The Times, paints a worrying picture for the Conservatives, putting them on 98 seats, with Labour on 468 seats. Here are three things that we can learn from it, and three things that we can’t.




What Does This Tell Us?


1. The Conservatives are in Deep Trouble


If this result were to be repeated in a general election scenario, it would be the worst historical Tory defeat that the country has ever seen. Their lowest previous result in terms of seats was in 1997, where the party won 165 seats - 67 more than under this scenario. 

This would also mean that a large proportion of the current cabinet are at risk of losing their seats. Meaning that some of those rumoured to have the leadership of the party in their sights such as Penny Mordaunt, James Cleverly, and Grant Shapps are unlikely to hold on to their seats if this scenario were to be repeated in a general election.

Out of the remaining cabinet ministers, Kemi Badenoch remains with the largest lead over the second-placed party in their seat, currently estimated to win with a small majority of 8%, where Labour are on track to overtake the Lib Dems as the second party in the North West Essex constituency.


2. Who are the Conservatives Losing To?


What MRP analyses are useful for is to tell us how the national vote shares are distributed across constituencies. Considering Starmer’s low personal popularity compared to Tony Blair’s in 1997, much of the discussion in political circles has been whether people are voting against the Conservatives, or towards Labour

From the plot below, we can see that Labour is underperforming in terms of gaining lost Conservative votes in the majority of seats. The black dashed line shows how we would expect the Labour vote to rise if the Conservative losses and Labour’s gains in each constituency were evenly distributed. However, we find that this is the case in very few constituencies. 

Instead, we see that even in constituencies which Labour is set to gain from the Conservatives, the swing towards Labour is lower than the swing away from the Conservatives. Thus, while Labour are set to gain a significant number of seats directly from the Tories, the Conservative decline in vote share is not a sole result of Labour's rise.



3. Reform UK Splitting the Conservative Vote


What we do find is that the Reform UK vote share is the highest in places where the Conservatives are set to lose the most votes. 

Thus, even though Labour are not gaining all the lost Conservative votes, they are likely to gain seats where the Conservatives held majorities close to 50% in 2019 as a result of dissatisfied 2019 Conservative voters being split between Labour, Reform UK and the Liberal Democrats in places such as Brigg and Immingham, Hinckley and Bosworth, and Gosport. 



To see what would happen if Reform were to stand down their candidates to reduce Conservative losses as UKIP did in 2017 and the Brexit Party did in 2019, we asked those who are currently intending to vote for the party which party they would vote for if there weren’t a Reform UK candidate standing in their seat.



As it stands currently, we find that under this extreme scenario, the Conservatives would be able to up their current seat count by over 50% to 150, while Labour’s seat count would go down to 422. This would mean that the Conservatives would win back 46 of the seats lost to Labour, and 7 of the seats projected to be won by the Lib Dems if Reform were to stand in all constituencies.


What MRP Doesn't Tell Us

The advantage of MRP analyses compared to standard polls is that we can find out how the national level results are spread across individual constituencies. As with any mid-term poll outside of an election context, we are not trying to predict something that would happen in an election setting. Instead, we ask people how they would vote if there were to be a general election tomorrow, and if there were candidates from all political parties standing in their constituency. This has several implications:


1. If/How much the Polls Will Narrow


We know that people respond to events differently when they know that an election will be taking place and are certain what the stakes are. What we can’t foresee is whether, or by how much the polls will narrow between now and the election. What we do know and show is that this is the current feeling of the electorate across different constituencies.


2. The Extent of Tactical Voting


Relatedly, while we make the effort to, we cannot estimate precisely the extent to which tactical voting will play out in an election scenario. With the new constituency boundaries in play, many voters are likely to turn to tactical voting recommendations from campaign groups which are likely to swing the vote in marginal seats one way or another. 

For example, we know that in 2019, tactical voting websites such as and had a significant impact on the Lib Dem vote share in certain seats. While it would be unwise for us to try to guess how exactly this would play out in an election scenario this far out, it would be reasonable to assume that tactical vote recommendations are likely to sway the vote in certain seats, especially where small swings could be deciding for the winning party. This is also an indication that the currently predicted Lib Dem seats are likely at the lower end of what we might expect to see in the election where they tend to benefit significantly from tactical voting.


3. No Method is Faultless


As with any statistical analysis, there is a certain level of uncertainty in our MRP estimates. We report this through the low and high estimates alongside the mean values for each party in each constituency which represent 95% credible intervals. These can be interpreted as confidence intervals which show our certainty in the mean values that we provide. 

When we report an estimated number of seats won by each party, we take as the winner the party with the highest mean value from our estimates. This does not mean that we are always 100% certain that the reported 'winning' party would actually win in the seat. It is important to interpret the results together with our uncertainty about the estimates, particularly in marginal and three-way marginal seats.

One such seat is Woking, where the predicted mean estimates for each party show a close win for Labour with 31.4% of the vote, followed by the Conservatives with 31.2%, and the Lib Dems in close third with 27%. However, when we take into account the low and high intervals, we see that it is possible that any of those three parties could actually win the seat. The plot below shows how certain we are about the vote shares of each party in the Woking constituency.



As an additional aid for such scenarios, when we publish our MRP estimates, we also provide the probabilities of each party winning a particular seat from the many simulations of the model that we run. In this case, they help us understand that despite winning by a very small margin, the chances of Labour winning the seat given the data are 43%, the chances for the Conservatives are 35%, and for the Lib Dems - 22%.


Taking the above into account, there is much to learn from MRP analyses, even in the mid-term. What the current analysis shows is that the Conservatives have a long way to go in order to minimise their loses in key battleground seats, and while Labour are on track to win a majority at the next election, a lot rests on the impact that Reform will have in traditional Conservative constituencies, as well as the potential impact of tactical voting. 


We have provided a guide to interpreting the results from our MRP here, and you can find out more about how MRP works here.


Survation used MRP during the 2019 General Election to correctly predict a large Conservative majority and call 94.3% of seats correctly. For more information, you can read this blog post by our partner Professor Chris Hanretty here.

Get the data

Survation conducted MRP analysis of 15,029 adults aged 18+ in Great Britain on their voting intentions. Fieldwork was conducted between 8th - 22nd March 2024. Tables for the standard voting intention results are available here, and tables for the Reform stand aside constituency estimates are here.


If you are interested in commissioning MRP or to learn more about Survation’s research capabilities, please contact John Gibb on 020 3818 9661, email or visit our services page.

For press enquiries, please call 0203 818 9661 or email

Survation. is an MRS company partner, a member of the British Polling Council and abides by their rules. To find out more about Survation’s services, and how you can conduct a telephone or online poll for your research needs, please visit our services page.


< Back