I hope my statistics textbooks will change the world. They have been my passion for 17 years since I ‘retired’ (gave up the salary) from La Trobe University. The intro textbook (Routledge 2017), co-author Robert Calin-Jageman, was the first to have Open Science and estimation (confidence intervals and meta-analysis) introduced in Chapter 1 then used throughout.

The second edition (Routledge 2024) introduces the wonderful new Calin-Jageman free esci software, open source in R, with jamovi or JASP as super-friendly environments for anyone, from beginner to seasoned researcher, to use esci. Exactly the software researchers need as they adopt Open Science, and focus on effect sizes, precision, and cumulative science. Exactly what’s needed for how research should now and into the future be conducted.

---

ABSTRACT:

Statistical significance and p values: The researcher’s heroin

For many decades distinguished scholars across science have explained numerous problems of p values and statistical significance, ranging from logic to researchers’ misconceptions and common practices. But cogent arguments have been ineffective: significance testing remains entrenched as the basis for research conclusions.

Might visceral demos of dramatic problems with these addictive practices be more effective? Few researchers understand that the sampling variability of p is astonishingly huge! I made two demos of that: the ‘dance of the p values’ (search YouTube for videos) and p intervals—80% prediction intervals of where p from a replication will fall.

A third, and possibly most dramatic demo is significance roulette (search YouTube for videos). We’ll visit the p Value Casino for the public launch of Bradley Dean’s free demo that runs in any browser.