Headline figures of many polls feature views of many more men than women and this could be skewing results

December 15th, 2014

Fewer women say they are certain to vote and they are more likely to be undecided

Just look at the chart which is based on the Populus November aggregate data with a very large overall sample from 9 separate polls.

As can be seen male voters account for more than 50% of each party’s support with, not unexpectedly, UKIP showing the biggest divide. What is striking is that although the overall sample is weighted properly for gender balance when it comes to the output figures there are 11 men for every 9 women.

The happens because women’s certainty to vote figures are lower and they are more like to be undecided about their party choice.

This is a common feature with online polls generally while the phone surveys tend to be more balanced. YouGov doesn’t weight for certainty but does have significantly more female don’t knows.

    Given that we know that in general elections male and female turnout levels are roughly the same isn’t there a case for gender weighting of the final figures?

Polls that are male heavy might not be as good snapshots of opinion as those that have a better balance.

Mike Smithson

2004-2014: The view from OUTSIDE the Westminster bubble