So Donald Trump has beaten the polls and is to be the next president of the United States of America.
Or has he? Beaten the polls, that is. Or have the polls beaten themselves?
Once again, there is a feeling that the polls have been found wanting. And, as followed the UK general election of 2015, there is bound to be extensive and intensive effort to understand what they did and did not get right.
In the meantime, a lot of people will still be using polls and surveys to inform strategies and campaigns. Public relations professionals are among them.
My advice, for the time being, is don’t abandon your surveys and polling. But do check you understand their limitations.
Here are just a few things to have in mind:
- The responses give a snapshot of opinion as it was at the time of asking the questions. Caution is needed in treating them as a guide to future views and actions.
- The views expressed may not be the views held by the respondent, especially on sensitive issues.
- The wording of questions, and their order, can influence how a person answers. Indeed, just doing a survey may have a biasing effect on their responses.
- It is increasingly difficult to get representative samples. Various methodological ‘corrections’ may lead to a few individuals having a highly distorting effect on the overall result. And, worse still, some important groups may get missed out entirely.
- When an inference is made about a population from a sample, it comes with uncertainty – make sure you understand what it means whether it’s called a margin of error, a confidence interval, a credibility interval, or whatever. And understand to what level you can be confident. (The typical range is plus-or-minus three percentage points on a figure, to 95 percent confidence.)
If in doubt, get advice! Plug – I can help.