Twitter poll results don’t show ‘wide support’ for Thailand’s travel policy

UK attitudes to Thailand’s foreign travel policy are not normally high up the list of misinformation concerns.

Nevertheless, an article on the subject in The Independent does neatly encapsulate almost everything that is bad about doing polls on social media. Though, to be fair, it does seem to have some self-awareness of this.

The article claims that most people are in favour of Thailand’s plans to open up to vaccinated travellers from 1 July. 

It cites as evidence the results of a “Twitter poll conducted for The Independent”. This turns out to be a Twitter poll posted by the journalist who wrote the article. 

Forget “We’re all going on a summer holiday”
This year it’s “Us oldies are going on a jab jaunt” as Thailand opens up a tourist island for vaccinated visitors from July.
Should older holidaymaker happily leave younger travellers behind and say “Phuket, I’m going”?

— Simon Calder (@SimonCalder) March 27, 2021

In general, polls are only meaningful if you’ve got some reason to believe that the people surveyed reflect a wider population of interest. For example, if you want to know about UK voting intention, you’d survey a group of UK adults who broadly match the population profile in terms of age, location, gender and other factors. 

But because you never really know who is answering a social media poll, they can never be said to actually represent a wider population. So when the headline claims the policy has “wide support”, that doesn’t really mean anything.  

In this case, the results, at best, could be argued to reflect the wider views of the journalist’s social media followers. But even that would be biased towards those followers who were on Twitter between 9am and 11am on 27 March when the poll was open, who saw the tweet and who actually cared enough to vote.

Twitter polls cannot avoid these issues of “self-selection”, which The Independent acknowledges in its piece.  

Then there is the problem of question wording. 

Good surveys ask neutral non-leading questions to get a fair read on public opinion. In this case, the poll framed the question almost as a matter of intergenerational inequality.

This might have, for example, inflated the under-50 “no” vote among people who might be quite happy for older vaccinated people to be allowed to travel, but not if they were going to lord it over the unvaccinated younger travellers. Or it might have made the idea of a “jab jaunt” sound more appealing to the over-50s, prompting them to vote “yes”.  

And finally, it's important that survey results are reported accurately. In this case, the poll asked: “Should older holidaymaker [sic] happily leave younger travellers behind and say ‘Phuket, I’m going’?”

But in the article write-up, the results were reported as the number in favour of Thailand’s travel policy in general, not the number in favour of older holidaymakers leaving younger travellers behind, apparently in an impulsive act of self-satisfaction. 

A subtle difference but a real one.

For more on this subject, you can read our guide on how to spot misleading polling figures.