Any free-text survey like this that doesn’t have clearly defined standard categories should be treated, at best, as weak evidence. The person doing the categorisation decides the shape of the results, you can split into more precise labels to downplay and merge to add emphasis https://twitter.com/jamesjohnson252/status/1390729101732982788

I can say that with confidence because I’ve done it many times, shaping taxonomies to give clear, interesting results from messy data. Even if you try to approach it honestly, your ad hoc post-collection categorisation WILL be shaped by your biases/skills/mood/whatever

Maybe that polling firm does have a set of standard categories it then combines with more topical issues/policies that arise, but the fact that each response was forced into only one option seems dubious & there’s nothing about the manual coding system in the full report

Would you put this under “No chance of winning”, “Keir Starmer/leadership”, “Do not agree with policies/policies not clear” (massive vague category), “Policies not radical enough”, or “Weak”? No idea what they picked as the one option

(not making any kind of political point whatsoever in this thread, just get annoyed when stuff like this has less diligence put into it than daft marketing things I’ve worked on)