Small Data: Do people lie in surveys?

http://www.bbc.co.uk/news/blogs-magazine-monitor-29206289

Version 0 of 1.

I received a press release recently that made my head spin, writes Anthony Reuben.

The release said, completely without irony: "A third of people in the UK will not give truthful answers about themselves when asked questions by pollsters, according to a new survey."

In paradoxical terms, that is well up there with the words of Psalm 116: "I said in my haste, 'all men are liars'."

But while people's honesty when responding to polls or surveys may be a difficult issue to get to the bottom of through the use of polling, it is still an important factor.

The most often cited example of this in political circles is the polling ahead of the 1992 election, when people did not want to admit that they were voting for John Major.

And a note for the picky - while they are used interchangeably, strictly speaking a poll is only meant to have one, multiple choice question. If there are more questions or open answers it's a survey.

Anyway, dishonest respondents are a serious problem for pollsters, especially in a situation such as an election, in which the quality of the sampling will be tested shortly afterwards by the actual result.

There is almost nothing that polling organisations can do about this. Online polling may give some weight to how people voted in previous elections, although they may also be lying or misremembering about that.

When you look at the full results of the ORB International survey it turns out that overall 80% of people say they always provide truthful answers to surveys.

The one-third figure is in answer to a question about whether you would answer questions truthfully about "your intimate life", and is a result of adding up those who answered that they were not likely to answer truthfully together with those who said they didn't know or didn't answer.

The survey also found that only 9% of respondents said they were prepared to trust published polls.

Interestingly, the polling also suggested that respondents were most likely to be truthful about political questions, with 91% saying they were likely to tell the truth.

But there were no questions about whether they were likely to lie about lying.

Subscribe to the BBC News Magazine's email newsletter to get articles sent to your inbox.