30th June, 2019

Why should we care about statistics?

  • Uncertainty is integral to both science and life
  • Statistics is the discipline that deals with quantifying uncertainty
  • In contrast to how it is usually taught, statistics is a messy discipline full of opposing viewpoints

  • In the 20th century and beyond, there were 3 big traditions in statistics
    • They continue to influence our thinking profoundly

Historical context

Significance? What significance?

  • Before 1940, significance testing was virtually non-existent
  • By 1955, 80% of empirical articles reported significance tests
  • Today, the figure is close to 100% (see Gigerenzer, 1993)

“[Statisticians] have already overrun every branch of science with a rapidity of conquest rivalled only by Atilla, Mohammed, and the Colorado beetle.”

  • Kendall (1942, p.69)

Ronald Fisher

  • Father of modern statistical inference
  • Randomization, design of experiments
  • \(p\)-value, \(\alpha\) level, null hypothesis, analysis of variance
  • Maximum likelihood estimation (Stigler, 2007)
  • The “greatest of Darwin’s successors” (Edwards, 2011)

Neyman-Pearson

  • Jerzy Neyman and Egon Pearson put Fisher’s ideas on a rigorous mathematical basis
  • Fisher didn’t like it, and a personal feud began that lasted until his death

  • Some concepts Neyman & Pearson introduced were
    • The alternative hypothesis
    • Type I (\(\alpha\)) and Type II (\(\beta\)) errors
    • Statistical power
    • Confidence intervals
    • Decision theoretic foundation of hypothesis testing