Six polling firms published final calls for party vote shares before the 2026 Scottish Parliament election. Survation’s final call was the most accurate of any pollster. For the Constituency vote, Survation’s mean absolute error of 0.64 was the lowest of any polling firm, with the only MAE below one percentage point. 

Mean absolute error measures the average deviation between a pollster’s published figures and the actual result across all parties. Survation called SNP at 39% (result: 38.18%), Labour at 19% (result: 19.18%), and Conservative at 12% (result: 11.83%), all within one percentage point.

On the regional list vote, Survation were again close to the actual result, with an MAE of 1.26. The list ballot proved harder for the industry to call overall, with some firms reporting large errors predicting Labour, Reform or Green support. Combined, Survation had the lowest error across the Constituency and Regional ballots.

Survation’s MRP projection of 59 SNP seats came closest to the actual result of 58

Several pollsters also published MRP seat projections ahead of the election. Survation’s projection of 59 SNP seats came closest to the actual result of 58, and our overall seat projection error was the lowest of any firm producing an MRP model.

The 2026 Scottish Parliament election was a good test of polling methodology in a multi-party, dual-ballot system. Survation’s results across the three measures were the strongest of any firm that published a final call. 

Why we remain accurate

Holyrood elections are hard to call. A multi-party system, two separate ballots, candidate turnover and a significant share of voters deciding late all mean that a small number of methodological choices can move the final numbers materially.

Survation has been polling Scotland for over 14 years. We were the most accurate pollster for the 2014 referendum, and have built on that foundation continuously since. Our approach tries to capture the underlying structure of Scottish public opinion in ways that demographic weighting alone cannot, and we avoid making strong assumptions about voters who have not yet committed. Our MRP and analytical framework are built specifically for Scotland, that means being trained on Scottish polling, calibrated against local dynamics, not treated as a regional subset of a GB-wide model.

Our final poll included a combination of online and telephone fieldwork, rather than relying on either method alone. That matters because online-only samples systematically underrepresent certain types of voters, and weighting can only do so much to correct for that. By recruiting through both channels, we capture parts of the electorate that pure-online polling tends to miss, and our weighting starts from a more representative base.

We are proud of our performance in the 2026 Holyrood election and will carry the same approach going forward.
This cycle we were pleased to be trusted polling partners for Scotland in Union, the University of Strathclyde, Diffley Partnership and Ballot Box Scotland.

________________________________________

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

For press enquiries, please call 0203 818 9661 or email media@survation.com

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.