Confidence Intervals II
Lecture 24?
Duke University
STA 199 - Fall 2023
2023-11-16
Checklist
– Clone ae-23
– Exam-2 released today (between 4 and 6pm)
– Class next Tuesday will be on Customizing Quarto reports, Communicating results effectively, and and time to work on your project / ask questions during class
Exam-2
– Don’t cheat
– Cumulative with a focus on content post Exam-1
– Extension questions
– Use Slack for clarification questions
– Do not post code / results in Slack
Warm Up
Population vs Sample
– What the difference?
– How does notation change for the following:
— Mean
— Proportion
— Slope
Warm Up
Last time, we set up the bootstrap distribution to create a confidence interval for the following: \(\mu_{setosa} - \mu_{versi}\)
– Why do we make confidence intervals?
– In general, describe the difference in simulation scheme for a bootstrap distribution different versus creating a null distribution? Where are each distribution centered?
Connection
Some of you have asked about how Z and t tests relate to what we are doing…
It’s the same!
These simulation techniques (with some help from CLT) approximate the normal and t-distributions you have used before
An advantage of simulation techniques is that they have less strict assumptions (sample sizes often just need to be > 10)
Simulation techniques were developed once we had the computing power to do so. Theoretical techniques came first
There is no disadvantage of using simulation techniques versus theoretical in this context
We may talk about theoretical procedures and how this relates after break
