More or Less Behind the Statistics
More or Less Behind the Statistics

Alzheimers, Psychology science, John Conway, Red cards, Decimate

Alzheimers What's behind the claim that 1 in 3 people born in the UK this year could get Alzheimers? How reliable is the science in psychology? The Reproducibility of Psychological Science project reported recently and it made grim reading. Having replicated 100 psychological studies published

Featured Speakers

BBC HostJohn Conway Guest

Topics Discussed

Episode Summary

Executive Summary: This episode examines three recurring More or Less themes: how shaky headline statistics can mislead, how psychology is grappling with a reproducibility crisis, and how a mathematical genius, John Conway, shaped modern mathematics. It also debunks a football cliché about red cards and closes by correcting language nitpicks, arguing that modern usage of "decimate" and even "less" with count nouns is well established.

Main Topics: Dementia forecast headline scrutiny (Priority: 5/5): The show questions a widely circulated claim that one third of British people born in 2015 will develop dementia, arguing it relies on outdated incidence data and ignores recent evidence of declining dementia prevalence. Dementia prevalence is falling (Priority: 5/5): Using a recent Lancet study, the episode explains that dementia prevalence among people aged 65+ in the UK has dropped by over 20% in the last 20 years, likely linked to improved vascular health and broader health improvements. Psychology’s reproducibility crisis (Priority: 5/5): The Reproducibility Project is discussed as a large-scale effort that replicated nearly 100 psychology studies and found far fewer successful replications than expected, raising concerns about publication bias and analytical flexibility. John Conway’s mathematical legacy (Priority: 4/5): The episode profiles mathematician John Horton Conway, highlighting his work in group theory, the Conway group, and the Game of Life, while also portraying his eccentric personality and prolific intellectual life. Football red-card cliché tested with data (Priority: 4/5): A statistical look at whether it is really "hard to play against 10 men" finds that having a player sent off generally hurts the team, with a measurable average points loss and especially strong effects for the home team. Language and usage: decimate and less (Priority: 2/5): The programme ends by defending the modern use of "decimate" to mean devastate and noting that "less" with count nouns is historically attested, framing both as issues of style rather than grammar.

Key Arguments: The dementia headline is built on outdated incidence data, making the forecast unreliable without newer figures. Recent evidence suggests dementia prevalence has declined substantially, so future incidence may also be lower than older studies imply. A low replication rate in psychology does not mean every original finding is false, but it does indicate serious structural problems in study design, analysis, and publication practices. Publication bias and excessive analytical choices can make apparently significant results fragile and hard to reproduce. John Conway was not just the inventor of the Game of Life; his mathematical work in group theory and beyond was broader and highly influential. Red cards are generally harmful to a team’s chances, and the effect is larger for the home team than for the away team. Modern usage has long accepted "decimate" in the sense of destroy badly, and "less" with count nouns is a usage issue, not a grammatical error.

Data Points: Estimated dementia risk in British people born in 2015: 1 in 3 - Headline being examined; based on older incidence data Dementia prevalence among UK residents aged 65+: down by over 20% - Finding from the Lancet study cited in the episode Time period for dementia trend: last 20 years - Interval over which prevalence declined Psychology studies examined in Reproducibility Project: almost 100 - Replication project across three leading journals Successful replications: 39 - Number of studies in the project that were confirmed by replications Replication attempts with statistically significant results: 36% - Share of replications reaching significance Relative size of replications: half as big - Effect sizes in replications were about half the originals Researchers involved in Reproducibility Project: 270 - Scale of collaboration coordinating replications Premier League games analyzed: 1,520 - Football red-card study, 2009 to 2013 Red cards in sample: 238 - Total red cards across four Premier League seasons Average points difference from 11 vs 10 men: about half a point - Estimated average impact on match points Time delay effect on red-card impact: 18 minutes - Delaying a red card by this amount reduces the expected points loss by one point on average John Conway's dimensions for Conway group: 24 dimensions - Describing his work in group theory Monster group size: nearly 200,000 dimensions - Another major mathematical object associated with Conway Conway’s marriages: 3 - Biographical detail discussed in the profile Children: 7 - Conway’s family size mentioned in the interview

Pivotal Quotes: "the age-specific prevalence of dementia had gone down substantially by over 20%" — Carol Brain: Summarizing the Lancet study's main finding "I think they should publish less and they should publish more at the same time." — Andrew Gelman: Commenting on psychology's publication culture and data sharing "Mathematics is the simple bit. It's the stuff we can understand. It's cats that are complicated." — John Conway: Opening quote used to introduce Conway's personality and wit

Implications: Listeners should treat dramatic headline statistics cautiously, expect more emphasis on replication and data transparency in science, and recognize that both medical forecasts and social-science findings can change as evidence improves.

🔓 Sign Up for Unlimited Episode Search

About More or Less Behind the Statistics

Tim Harford and the More or Less team try to make sense of the statistics which surround us. From BBC Radio 4

View all episodes from More or Less Behind the Statistics