When correlation is causation

  • Event: 2016 CAR Conference
  • Speaker: Stijn Debrouwere of independent journalist
  • Date/Time: Saturday, Mar. 12 at 4:45pm
  • Location: Penrose
  • Audio file: Only members can listen to conference audio

Is red meat bad for your health? What's a triple-blind clinical trial? Are the Finnish really that great at mathematics? What's so significant about statistical significance? During the first half of this workshop, you'll get acquainted with concepts like confounding, selection bias, Simpson's paradox, randomization and statistical power. In the second half, it's up to you: judge your way through a series of small case studies and practice what you've learned. It's not enough to say that correlation does not imply causation. Learn exactly what stands in the way of drawing truthful conclusions from data, how to recognize great science and how to spot the flaws that invalidate statistical conclusions.

Speaker Bios

  • Stijn is a freelance data scientist, specializing in analytics. Previously at Fusion and The Guardian. He writes about statistics, metrics and the news industry at debrouwere.org.

Related Tipsheets

  • Correlation is causation: Questions to ask
    Is red meat bad for your health? What's a triple-blind clinical trial? Are the Finnish really that great at mathematics? What's so significant about statistical significance? During the first half of this workshop, you'll get acquainted with concepts like confounding, selection bias, Simpson's paradox, randomization and statistical power. In the second half, it's up to you: judge your way through a series of small case studies and practice what you've learned. It's not enough to say that correlation does not imply causation. Learn exactly what stands in the way of drawing truthful conclusions from data, how to recognize great science and how to spot the flaws that invalidate statistical conclusions.

  • When correlation is causation: presentation
    Is red meat bad for your health? What's a triple-blind clinical trial? Are the Finnish really that great at mathematics? What's so significant about statistical significance? During the first half of this workshop, you'll get acquainted with concepts like confounding, selection bias, Simpson's paradox, randomization and statistical power. In the second half, it's up to you: judge your way through a series of small case studies and practice what you've learned. It's not enough to say that correlation does not imply causation. Learn exactly what stands in the way of drawing truthful conclusions from data, how to recognize great science and how to spot the flaws that invalidate statistical conclusions.