#223 - Sir David Spiegelhalter - "I Immediately Distrust Every Number In The News"

— Statistical manipulation is ubiquitous in media and politics; identical numbers can be framed as frightening or reassuring depending on storytelling and comparators used.

— Trustworthiness in communication requires stating not only what data means but also what it does not mean, and preempting likely misunderstandings.

— COVID messaging failures included oversimplified claims ("safe and effective") that decreased trust among vaccine-skeptics; balanced, nuanced messaging performed better in randomized trials.

— Lockdown timing and school closures warrant scrutiny; infections were already dropping before compulsory measures, and children faced negligible direct risk yet bore educational and mental health costs.

— AI and machine learning amplify the need for statisticians rather than eliminating them; black-box models require human supervision for interpretation, uncertainty quantification, and ethical constraint-setting.

— "Hedgehogs vs. Foxes" framework: hedgehogs (ideologically fixed forecasters) make poor predictions; foxes (evidence-responsive, humble, open-minded) are superior forecasters and advisors.

— When encountering any statistic in news or social media, first ask who is telling you, why they are telling you, and what is in it for them—almost always there is an attempt to manipulate—before examining the number itself.

— Become more neutral and evidence-responsive in your own reasoning and public communication; pre-judging conclusions and cramming evidence into preset ideology undermines trustworthiness and closes off genuine learning.

— Seek out "fox" advisors and colleagues—those who change their minds with new evidence, acknowledge uncertainty, and resist ideological capture—rather than relying solely on visionary "hedgehog" leaders.