ITV and Social Media Surveys

The broadcaster ITV uses Twitter ‘polls’ in their reporting.

In their online news articles, ITV include the results from Twitter surveys.

Due to self-selection bias, these surveys are not proper measures of opinion. We cannot use these surveys to infer the views of the public. Broadcasters should avoid citing social media surveys.

The many problems of social media surveys

On social media platforms, users may set up a survey for other accounts to take part.

There is no control on responses other than one vote per account. The same person can respond with many accounts.

There are no guarantees that responses are from the intended population. A person may start a survey asking only certain types of people to vote (e.g. teachers). Everyone with an account can vote.

Voluntary responses cause unknown error. Those voting are those who see the ‘poll’ and animated enough to take part. When people volunteer their responses, we get self-selection bias.

In a 2018 report by British Future and HOPE not Hate, they compared an self-selecting survey to a proper poll. The question was about perceptions of immigration impacts.

Image for post
Image for post
ICM Unlimited conducted the poll. (Image: British Future/HOPE not Hate)

Extreme responses were over-represented in the open access survey.

Having a large sample does not fix this bias. It is common to see accounts ‘share for a bigger sample’. People are sharing with their friends and followers, who share political views. Self-selection bias can rise, not wane.

Social media users are not representative of the population. People with social media accounts tend to be younger, with higher education levels.

Twitter ‘polls’ on the ITV website

The following examples are from May 2020, for articles on the ITV website.

On 18th May, the ITV website ran the headline:

Will you send your kids back to school in June? Here’s what you really think

The article does not cite any public opinion poll conducted by reputable researchers. The figures for “you really think” are from a Twitter ‘poll’:

In a poll run by us on Twitter, 74.6 per cent of respondents said they would not be sending their child back to school on 1 June.

Image for post
Image for post
The question wording is also imbalanced. (Image: ITV)

On 25th May, the ITV article headline asks:

Should Boris Johnson sack Dominic Cummings?

Rather than citing a representative opinion poll, ITV refer to a self-selecting survey.

Good Morning Britain also ran a Twitter poll asking whether Cummings should sacked, you can see the results below:

There is a danger of acquiescence bias in the question wording too.

Three days later, ITV refers to another self-selecting survey about schools reopening. Twitter users do not represent ITV News viewers either.

We wanted to find out how our viewers felt about the idea of schools reopening.

In a Twitter Poll we asked “Do you think schools should reopen?” After 492 votes (as of 7:45pm on 28 May), almost 64% of these were “No”

On 29th May, ITV News Central again reported on a self-selecting Twitter survey.

In a Twitter Poll we asked “Are you planning on meeting people outside of your household from Monday?” After 395 votes (as of 4:35pm on 29 May), almost 54% said “yes” whilst almost 46% said “no”.

Due to self-selection bias, Twitter surveys are not proper measures of public opinion. We cannot make inferences about the general population from social media surveys.

Broadcasters should avoid giving credence to such surveys. This kind of reporting may cause confusion during elections, with people citing Twitter polls.

Self-selecting surveys are free to run: you get what you pay for.

This blog looks at the use of statistics in Britain and beyond. It is written by RSS Statistical Ambassador and Chartered Statistician @anthonybmasters.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store