General Election 2019 Stat Checks: Week Five

This is my final stat check for the campaign. Including the article before official campaigning started, this is the sixth post on suspicious stats and fallacious figures in the 2019 General Election.

The BBC’s More or Less programme has Tim Harford cutting his scalpel of truth into claims about nurses, hospitals and the economy.

Polls open at 7am on Thursday 12th December.

The poverty conundrum

Claim: There are 400,000 fewer children in poverty than in 2010.
Rating: False.
Reasoning: The Prime Minister claimed:

Actually there are 400,000 fewer children in poverty than there were in 2010.

There are multiple ways to measure poverty. Such claims generally require parties to be more clear about what they mean.

However, no main poverty measure shows a reduction of 400,000.

Image for post
Image for post
No main poverty measures shows a reduction of this scale. (Image: Full Fact)

Labour’s counter-claim that the number of children in poverty has increased by 500,000 refers only to relative income poverty after housing costs. In 2010/11, that measure estimated 2.3m children were in poverty. By 2017/18, that figure was 2.8m. Alternately, you could choose to measure against the estimate in 2009/10 (approximately 2.5m children).

The student vote

Claim: 72% of students are voting Labour.
Rating: Needs clarification.
Reasoning: Students are a small population, and their vote intentions cannot be effectively measured by a national survey of all adults.

These figures come from a sub-sample of 40 respondents (43 unweighted), in an ICM Unlimited poll of 2,029 GB adults. The survey was conducted via an internet panel on 29th November to 2nd December 2019.

Image for post
Image for post
There is a very wide margin of sampling error for such a small set of respondents. (Image: ICM Unlimited)

Sub-samples are not internally weighted, and extreme care should be taken with such tiny sub-samples.

Quitting?

Claim: 11,000 EU academics have quit the UK since the EU referendum.
Rating: False.
Reasoning: In Freedom of Information requests, the Liberal Democrats asked various universities how many of their non-UK EU staff had left in each of the past five academic years.

The requests did not collect any information about destination: this figure represents EU staff members with non-UK nationalities leaving their jobs in the years — not necessarily the country.

In 2015/16, HESA staff statistics suggest there were 33,735 non-UK EU nationals employed at British universities. This statistic represents all staff excluding ‘atypical’ employment. In 2017/18, that figure was 37,255.

Mark Wallace wrote about this figure, and uploaded the request spreadsheet.

Inconsistent costs

Claim: The average household has paid almost £6,000 extra a year since 2010.
Rating: False.
Reasoning: This is the principal claim of the Labour ‘Cost of Living Crisis’ document. The ‘average’ household used for the calculation has some unusual characteristics:

  • Holding two season rail tickets, even though an estimated 8% of adults in England in 2018 use surface rail at least once a week;
  • Needing childcare costs for 25 hours a week — without government support — for one children aged two and the other aged over two;
  • Having a child of primary school age who is not eligible for a free school meal.

This is not an average household. Costings are presented in cash terms, but wage increases are described in real terms (accounting for inflation). BBC Reality Check and Full Fact have articles on this matter.

The great gambit

Claim: Scotland has subsidised the rest of the UK for most of the last 40 years.
Rating: False.
Reasoning: Ian Blackford, the Westminster leader of the Scottish National Party, said:

UK government figures make it absolutely clear that Scotland has subsidised the rest of the UK in most of the last 40-year period.

In response to Channel 4 Fact Check, the SNP spokesperson states tax revenues per person in Scotland were higher than the rest of the United Kingdom from 1980/81 to 2011/12 — which covers 32 years, not 40. Tax revenues are only one side of the equation: government expenditure matters too.

Historical tax receipts for Scotland were produced by the Scottish government, as part of the Scottish National Accounts Project (SNAP). This project was followed by Government Expenditure and Revenue Scotland (GERS), which are National Statistics produced in line with the UK Code of Practice for Statistics.

At the crux, the issue is about North Sea oil revenues. With a geographic share of North Sea revenue, Scotland had a relatively larger fiscal deficit than the United Kingdom as a whole in 15 out of those 32 years. This is slightly less than half.

Looking at GERS in the past 21 financial years, Scotland has generally had a wider fiscal deficit (as a share of GDP) throughout.

Image for post
Image for post
North Sea oil revenue has dropped sharply in recent years. (Image: GERS)

A mistaken peak

Claim: Homelessness reached its peak in 2008 under the last Labour government.
Rating: False.
Reasoning: The Chancellor Sajid Javid claimed:

Homelessness reached its peak in 2008 under the last Labour government. Since then it’s down by almost a half.

No measure of homelessness shows this. Using statutory homelessness, where local councils accept a duty is owed to people with ‘priority need’, this peaked in 2003. It rose between 2010 and 2017.

Image for post
Image for post
Statutory homelessness peaked in 2003, and rose since 2010. (Image: Full Fact)

The Chancellor told Channel 4 Fact Check he did not recall the figure correctly, and the peak was in 2003 — which is right.

Consistent underestimation

Claim: YouGov have consistently overestimated Conservative support.
Rating: False.
Reasoning: Marcus Chown, a science writer and broadcaster, claimed:

The reason YouGov consistently overestimates Tory support and underestimates Labour support is to demoralise you — to make you think there is no hope.

The premise is false. We can look at every national election (General Elections and European Parliament elections) since 2010. Results are drawn from the House of Commons Library electoral statistics.

  • General Election 2010: YouGov’s final poll for The Sun estimated the Conservative share was 35%. The result in Great Britain was 36.9%.
  • European Parliament 2014: YouGov’s last poll on behalf of The Sun estimated the Conservatives on 22%. The GB result was 23.9%.
  • General Election 2015: YouGov’s poll for The Sun estimated Conservative support at 34%. As part of an industry-wide error, this was an under-estimate. The Conservative vote share in Great Britain was 37.7%.
  • General Election 2017: YouGov’s poll for The Times had a Conservative vote share estimate of 42%. The actual result in Great Britain was 43.4%.
  • European Parliament 2019: The Conservative vote share estimate in YouGov’s final poll for The Times was 7%. The GB result was 9.1%.

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