There’s something odd happening with the Savanta ComRes polls for the Scotsman/SoS. Since the beginning of April, consistently coming below all the others and now reporting on the same date as YouGov get 52% for the SNP in the constituency vote and Survation get 49%, they get an astonishing 42%.
25% for Douglas Ross? Now I know they’re fiddling something.
Given that they all ask the same question, it must be the sampling. Surely they wouldn’t construct their sample to deliberately get a low SNP score?