Diagnostic testing: what do those statistics actually tell you? Sensitivity, specificity, positive predictive value . . . you’ve probably seen these terms before. Maybe you memorized them for a test. But do you actually know what they mean? In this episode, we take a closer look at how diagnostic tests are evaluated—and how they’re often misinterpreted. From a genetic test for cellulite to a blood test for autism, we explore how “statistically significant” findings can turn into tests that don’t actually help anyone. Along the way we meet the freckle gene, the wanderlust gene, and infidelity gene.


Statistical topics

  • Base Rate
  • Bayes Rule
  • Case-Control Study
  • Matching
  • Conditional Probability
  • Sensitivity
  • Specificity
  • Positive Predictive Value
  • Prevalence
  • Negative Predictive Value
  • False Positives and Negatives
  • True Positives and Negatives


Methodological morals

  • “A biomarker paper is not the same thing as a biomarker test.”
  • “If your sample doesn't match the real world, then for all of your positive predictive value needs, call on Bayes' theorem.”


Detailed Show Notes with calculations


References

Kristin and Regina’s online courses


Find us on:

Kristin -  LinkedIn & Twitter/X

Regina - LinkedIn &ReginaNuzzo.com


  • (00:00) - Introduction
  • (02:24) - The Cellulite Test
  • (06:41) - Understanding Sensitivity and Specificity
  • (12:50) - Enter Positive Predictive Value
  • (18:40) - Why Base Rates Matter
  • (24:06) - More Ridiculous Tests
  • (32:33) - The Wanderlust Gene Deep Dive
  • (40:30) - The NeuroPoint Autism Test
  • (51:37) - Trying to Set the Record Straight
  • (01:00:42) - Personal Stories
  • (01:03:57) - Wrap-up


Podden och tillhörande omslagsbild på den här sidan tillhör Regina Nuzzo and Kristin Sainani. Innehållet i podden är skapat av Regina Nuzzo and Kristin Sainani och inte av, eller tillsammans med, Poddtoppen.