A few days ago I wrote about how the ban on exit polls being published before 10pm on polling day is worded broadly, and so could catch out people reporting a party breakdown for how many comments there have been on social media from postal voters saying they’ve voted.
It’s also pollsters who have to be careful about this ban. With so many people voting by post well in advance of polling day, it might seem natural for pollsters to ask people whether or not they have voted, separating out those who have (who can simply be asked how they voted) from those who haven’t yet, with all the follow up questions and adjustments needed for dealing with the ‘don’t knows’ in this latter category.
However, pollsters who did this would immediately run into a problem to which there is no good answer. On the one hand, the British Polling Council’s rules require transparency and therefore for any such published polls the results for those who have voted already would need to be published. But then on the other hand, the rules about exit polls would make such publication illegal.
The legal problem would apply too to the outlets which commissioned and published such poll results – as The Times found out when it nearly got prosecuted for exactly this problem. In the end, the CPS decided not to prosecute, but it’s why suggestions such as Michael Crick’s are misplaced: