Case Not Proven: The suggestion that there’s been a LAB>CON shift amongst women

February 13th, 2018

This has only been picked up by YouGov

With the apparent sharpish shifts that we’ve seen in recent days in the polls to the Conservatives there has, inevitably, been a lot of examination of the detailed data.

One thing that’s featured is there’s said to have been a noticeable move by women voters from LAB to CON who, in recent times, have mostly leaned more to Labour. This has led to a lot of speculation as to why this has happened including one suggestion that Jeremy Corbyn had alienated the Mumsnet audience over transgender issues!

The discussion come out of detailed analysis of the latest YouGov/Times poll that had CON overall 4 points ahead.

In order to establish whether others pollsters have found the same I’ve produced the above chart based on the datasets which have just come out.

This is based on the three polls so far in February:

Opinium/Observer; fieldwork Feb 6-8; sample 2002
YouGov, fieldwork Feb 5-6; sample 2000
ICM, fieldwork Feb 2-4; sample 2021

The chart is self-explanatory. YouGov is showing very different gender splits from ICM and Opinium. It might be that YouGov is supported by other polling but at the moment there’s nothing to reinforce the analysis based on one poll.

There’s an aspect of male-female splits in polling that is quite important. Women have a tendency to be less certain about their likelihood to vote which means their responses are discounted in the certainty computations. So the numbers we see have a male response bias which is not supported by actual voting behaviour.

Has there been a move as described? Maybe. Maybe not. We need to see it in more polls.

Mike Smithson