Statistical Ambassador: Year Two
In June 2018, the Royal Statistical Society selected me as a Statistical Ambassador. A misunderstood statistic can have dire consequences. We aim to improve public understanding of statistics. We work with policy-makers, researchers, journalists, and the public.
This article looks at my second year as a Statistical Ambassador.
Statistic of the Year and Decade
I checked the submitted figures.
Survey Research Primer
Patrick English (Exeter) asked me to write about the principles of survey research.
The article for POL/SOC1041 students covered purposes, terms, methods, and errors. It was good to help others learn. I have much to learn about surveys myself.
All a Hubbub
The environmental charity Hubbub asserted “half of all flights by men aged 20–45 were for stags”. The charity miscalculated the figure. This error was inadvertent.
Yet, the survey estimates were still implausible. The Censuswide survey suggested the average man aged 20 to 45 went on 3.3 stag dos in 2019.
Hubbub retracted this claim.
Within the RSS, there were concerns about misunderstandings of radio listening statistics.
I wrote a guide to help journalists understand listener estimates.
Writing for ‘UK in a Changing Europe’
I wrote an article on Medium about the claim: ‘Labour had been just 2,227 votes away from the chance of government.’ This claim rested upon diamond-strong assumptions.
I tend to write in a formal style. I needed to rework my original article, to make it more conversational.
At the end of October 2019, MPs called for a general election.
Throughout the campaign, I wrote weekly fact-check articles. On top of that, I wrote about:
- Multi-level regression and post-stratification estimates
- House effects
- Self-selection bias
- False recall
- Methodology of the political opinion polls
- The broadcasters’ exit poll
Mistaken claims continued after the election too. Two academics wrote an LSE blogpost about the BBC’s coverage, containing a solar system of errors. The authors corrected most of these errors. A big one (the median of polls “was highly likely to be accurate on national vote share”) remains.
I also recently wrote about claims of a “spike” in postal voting. Electoral Commission data shows postal voting fell between 2017 and 2019.
I got thanks from researchers for my voluntary work during the election. I am very grateful for those kind words.
The SARS-CoV-2 pandemic spread throughout the world. One consequence was changes in the kind of statistics that interest people.
I helped write a guide about coronavirus numbers with other Statistical Ambassadors. On this blog, I also wrote about:
- Case fatality rates
- Differences between DHSC and ONS death counts
- Excess deaths (helping with a Full Fact article)
- Reproduction numbers
- A simple cellular automata model for epidemics (and forest fires)
- Problems of international comparability and changing measures
Over the next year, I want to:
- Practice public speaking: through recording videos and taking opportunities.
- Continue my professional development: including survey methods, Bayesian statistics, quality assurance, and R.
It was great to work with the other Ambassadors, and I look forward to sharing their efforts.