Learning statistics is often characterized by tedium and frustration. To make matters worse, pervasive misunderstandings of conditional probabilities often impede learning. We present an interactive spreadsheet designed to elucidate such misconceptions through a comparison three statistical approaches: Bayesian parameter estimation, Bayesian model comparison and null hypothesis significance testing. Learning is facilitated through the systematic exploration of each method and the use of graphical displays of the distributions. The conceptual underpinnings of each approach are described as well as their implementation in the spreadsheet. We conclude with some suggested pedagogical questions designed to elucidate the commonalities and differences between each approach.

4 Fisher Wolfe.xlsx (1080 kB)
Final Fisher Wolfe.xlsx