Monday, November 9, 2015

do opinion polls actually mean anything?

Do opinion polls actually mean anything? Here’s an interesting piece from the Pew Research Centre on the declining response rates  to opinion surveys. In some cases the response rate is as low as 9%. If only 9 percent of those approached agree to take part in a poll how likely is that that the poll represents a truly random sample?

The article goes on to claim that this is not a major problem, but since they are a polling company you would of course expect them to say that.

They do admit that low response rates have introduced a certain bias. People who are more engaged in the political process, not surprisingly, are more likely to respond. Alert readers might well see this is as being potentially very significant indeed. People who are obsessively engaged in the political process might well be expected to be a lot more left-wing than the general population. One might reasonably suspect that the sorts of people who are involved in “political activism” and who spend their leisure hours carrying placards in demonstrations would be much more likely to respond to a public opinion survey. This could mean that opinion polls are seriously overstating the level of support for notions like climate change, homosexual “marriage” and open borders.

It’s interesting to speculate on the possible reasons for declining response rates. I would personally suspect that people who hold politically incorrect views are increasingly cautious about expressing such views. This caution is perfectly understandable in a society in which you can lose your job and your future prospects by expressing a forbidden view. Guarantees of anonymity are also increasing regarded with scepticism. We live in a world in which privacy is thing of the past. Even if the risk seems small would you risk your entire future just to participate in a poll?

I also suspect that even those who do respond may be reluctant to be entirely frank.

So do opinion polls increasingly reflect a subtle socially liberal bias? I would suggest that it is entirely possible.

One might also ask the question - if opinion polls are no longer reliable how much sense does it make for governments to go into a panic every time a slightly unfavourable poll comes out? And how much sense does it make to organise a coup to remove an elected leader who won a huge majority in the previous election (as happened recently in Australia) just because of rather questionable opinion polls?

No comments:

Post a Comment