A reason to be sceptical of what the public tells opinion pollsters

By Mark Pack | Thu 14th April 2011 - 8:55 am

Much can be learnt from opinion polls, but a reminder of why not all results should be taken at face value is this:

If there were local council elections in your area on May 5th, how likely would you be to vote in them, where 0 means you will definitely not vote, and 10 means you definitely will vote?

10 – will definitely vote: 52%

This poll is not unusual in showing more people saying they will certainly vote than seems credible – and polls before previous elections (i.e. where we know the actual subsequent turnout) have often shown the number of people saying they’re certain to vote exceed the number who do. Where research has compared such results against the marked register, there is a relationship – the more likely people are to say they will vote, the more likely they actually are to vote, but even so a fair number of those “certain” to vote in reality aren’t.

This poll, by the way, was an online survey so the possible peer pressure from talking to a fellow human being and giving the answer you feel you should does not apply.

Read more by Mark Pack or more about turnout or yougov.
This entry was posted in Polls.

3 Comments

Mick Spreader 14th Apr '11 - 9:44am

One of the local newspaper web sites was running a local election poll. It was clearly unrepresentative of actualy voter intention so I made a few adjustments.

They have just taken it down and replaced it with an AV poll ;0)
paul barker 14th Apr '11 - 2:59pm

Theres a general problem that we often give the anwers we think we “should” or those we think others would give. This gets worse when theres an “everybody knows” issue like Immigration or The Libdems. Its a sort of Echo Chamber where the genuine hate for the Libdems of a minority becomes amplified.
The question is whether that effect will change Voting in actual Elections, we will know in 3 weeks. My feeling is that our losses will be very small.
Matthew Lambert 15th Apr '11 - 11:59am

If it’s an online poll then maybe it is representative of the sort of person who is visiting that site/are happy to spend their time filling in questions when asked to online.

I fancy if a similar simple poll was ran on here you wouldn’t find many not offering a “10” response.