ComRes and the Telegraph come under fierce fire from other pollsters over the latest poll
I can't recall a national opinion poll for a leading newspaper being so criticized as tonight's one from ComRes for the Daily Telegraph.
To think that just 4 years ago after the 2015 general election ComRes was regarded as the top pollster
— Mike Smithson (@MSmithsonPB) August 12, 2019
Important – @Telegraph reporting taking the piss by ignoring “don’t knows” to present an apparent majority – one for @FullFact @RoyalStatSoc @TweetMRS https://t.co/Qyz0hopxZL
— Ben Page, Ipsos MORI (@benatipsosmori) August 13, 2019
How ComRes presented their controversial findings in their press release on the poll pic.twitter.com/9Q21IPnFCs
— Mike Smithson (@MSmithsonPB) August 13, 2019
POLL WARNING: This is a highly misleading use of polling data. Yes it’s 54% who actually have an opinion, but that’s not representative of all of us if you go & leave out the group of undecideds/DKs from the data/calcs. Not nationally representative & therefore false. #FakeNews https://t.co/zaMJBp3J51
— James Endersby (@JamesEnders) August 12, 2019
Two of the Tweets above are from leading figures in the industry – Ben Page is boss of Ipsos-MORI and James Endersby boss of Opinium and it is hard to recall a time when one major national pollster has under fire in public from their peers in the industry in this way.
Anthony Wells of YouGov also joined the critics overnight.
There are two issues. Firstly it is the ComRes approach of asking agree/disagree question because these can be seen as misleading and there’s a tendency for respondents to opt for “Agree” when faced with a choice. This is something it has been doing for years and for which I’ve criticised in the past.
Secondly there is what ComRes did in their presentation of the data to ignore the don’t knows and round the answers up to 100″. This is distorting but it didn’t stop the Telegraph from putting out one of the Tweets above in the form it did.
Given the tone one the attacks this surely won’t be the end of the matter.
ComRes is a pollster that past vote weights that something that is proving problematic at the moment because fewer GE2017 LAB voters are “remembering” what they did meaning those that do remember are having their views weighted up. Anthony Wells had a good critique here.