One of the features of current polling is that we are getting very different findings for the best Prime Minister question as the chart above shows. While the Starmer figures are in a narrow range of 31%-35% the Johnson numbers have a spread of 30%-48%. Could it be the pollsters’ formats that impact on this.
While YouGov has Starmer ahead and Opinium has it very close the most prolific pollster, Redfield and Wilton, consistently have Johnson with substantial leads. I think that a lot of this is down to the question wording.
YouGov has it this way “Which of the following do you think would make the best Prime Minister?”. Dead simply and straightforward and NOT leading.
Newbie pollster Redfield and Wilton uses this format “At this moment, which of the following individuals do you think would be the better Prime Minister for the United Kingdom?”. This is much more wordy but critically has the prefix “At this moment.. which changes the nature of the question suggesting, I would argue, that this is a factor should affect people’s response.
Redfield and Wilton is new to UK polling and has yet to build up a track record. It did not publish polls ahead of GE2019 which could be compared with the outcome.