Approximating a normal distribution with a binomial distribution
Approximating a normal distribution with a binomial distribution
This page supports an in-class exercise that highlights several key Bayesian concepts. The scenario is as follows: a large paper bag contains pieces of candy with wrappings of different color, and we are interested in learning about the unknown proportion of yellow-wrapped pieces of candy. After completing the exercises, we will be familiar with the following concepts and ideas: probability distributions can quantify degree of belief, prior distribution, posterior distribution, sequential updating, conjugacy, Cromwell’s Rule (http://en.wikipedia.org/wiki/Cromwell's_rule), the data overwhelm the prior, Bayes factors, Savage-Dickey density ratio, sensitivity analysis, coherence.
Find the best linear fit for a given set of data points and residuals (or let this app show you how it is done).
When does a significant p-value indicate a true effect? This app will help with understanding the Positive Predictive Value (PPV) of a p-value.
This app is based on Ioannidis, J. P. A. (2005). Why most published research findings are false. PLoS Medicine, 2(8), e124. http://doi.org/10.1371/journal.pmed.0020124
Can you "see" a group mean difference, just by eyeballing the data? Is your gut feeling aligned to the formal index of evidence, the Bayes factor?
Visualizing the Bayes factor (quantification of evidence supporting a null or altermative hypothesis) using the urn model.
Use presets or change parameter values manually to explore the cost-effectiveness of different research approaches to unearth true scientific discoveries. For detailed explanation and conceptual background, see LeBel, Campbell, & Loving (in press, JPSP), Table 3. This app is an extension of Zehetleitner and Felix Schönbrodt's (2016) positive predictive value app.
This page presents a series of tutorials and interdisciplinary case studies that can be used in a variety of blended as well as brick-and-mortar courses. The materials can be used in introductory level data science courses as well as more advanced data science or statistics courses. These materials assume that students have a basic prior knowledge of R or Rstudio.
The goal of this text is to provide a broad set of topics and methods that will give students a solid foundation in understanding how to make decisions with data. This text presents workbook-style, project-based material that emphasizes real world applications and conceptual understanding. Each chapter contains:
The text is highly adaptable in that the various chapters/parts can be taken out of order or even skipped to customize the course to your audience. Depending on the level of in-class active learning, group work, and discussion that you prefer in your course, some of this work might occur during class time and some outside of class.
The Military Spending lab uses interactive, online graphs to better understand total military spending for each country. We see the limitations of traditional histograms and also consider the importance of using appropriate scales when comparing countries. The emphasisis of this lab is on understanding the impact of appropriate data transformations and data visualizations.
App: http://shiny.grinnell.edu/Military_Spending_Basic/
Handout: http://web.grinnell.edu/individuals/kuipers/stat2labs/Handouts/MilSpendB...