Statistics Power Half-Hour - Recordings
Statistics Power Half-Hour is open to anyone interested in gaining a better understanding of statistics. These recordings provide short introductions to important statistical concepts.
Power and Sample Size
Video Description: Why do we need power calculations, and how do they work? We’ll discuss the logic behind power and sample size calculations.
Instructor: Dr. Emily Griffith, Statistics Research Associate Professor
P-values
Video Description: Is it time to ditch the concept of statistical significance? What are p-values, and when and how should you use them? Join us to learn about the controversy around p-values and figure out where you stand.
Instructor: Dr. Emily Griffith, Statistics Research Associate Professor
Experimental Design
Video Description: How can you be sure you're finding what you want? We'll talk about general principles of experimental design and how to ensure that your design will help you answer your research question.
Instructor: Dr. Emily Griffith, Statistics Research Associate Professor
Misinterpreted Statistics in the News
Video Description: Calculating statistics is fairly straightforward, especially with the availability of free statistical software and online calculators. Correctly using and interpreting statistics is often the harder task. We will discuss a few relevant examples of misinterpreted statistics from recent headlines, focusing on the assumptions that underlie these misinterpretations. Assumptions are the foundation of statistical inference, and understanding why certain mistakes are made will help you become an informed user and consumer of statistics.
Instructor: Dr. Emily Griffith, Statistics Research Associate Professor
The Bootstrap
Video Description: Bootstrap methods can be a fun and easy way of obtaining confidence intervals for estimates and predictions, especially in machine learning. This talk will give a brief overview of what bootstrap methods are and when to consider using them. Additionally, we’ll spend some time discussing the assumptions behind the bootstrap and the limitations that those assumptions put on statistical inference.
Instructor: Dr. Emily Griffith, Statistics Research Associate Professor
Statistical Reasoning 101
Video Description: Crafting a convincing scientific argument requires careful consideration of the problem at hand, the form of the evidence available, and the strength of the conclusions it supports. Of course, the same care should go into the statistical part of the argument, yet the recent debates tend to focus on questions like "p-values: good or bad?" rather than on the statistical reasoning behind such considerations. The truth is there are no shortcuts, each problem is different and there is no magic statistical argument that is right for every problem. Being able to navigate this requires an understanding of statistical reasoning, and this presentation covers the basics, including a comparison of classical and Bayesian perspectives.
Instructor: Dr. Ryan Martin, Statistics Professor