4. Simulating

Simulating scenarios, simulating datasets, simulating random variables.

Artwork by @allison_horst.

Agenda

September 23, 2024

1. Why simulate?
2. What makes a good simulation?
3. Examples

September 25, 2024

1. Bias in modeling
2. Simulating statistical inference

Readings

• Class notes: Simulating

• Baumer, Horton, and Kaplan (2021), Iteration (Chp 7) in Modern Data Science for R.

Reflection questions

• What are some reasons to simulate?
• What makes a good simulation?
• What does “90%” mean when describing a 90% confidence interval? (That is, 90% of what?)
• What does a type I error rate mean when doing hypothesis testing? (That is, rate of what?)
• The p-value is the probability of obtaining the observed data or more extreme if the null hypothesis is true. If probability is interpreted in the long-run-frequency definition, what is happening repeatedly so as to conceptualize the probability of something happening?

Ethics considerations

• How can we use simulations to understand algorithmic bias or other types of problems in modeling?

• What does it mean for two different populations to have different feature distributions?

Slides

• In class slides for both 9/23/24 and 9/25/24.

• WS7 - Simulating

Additional Resources

• Posit cheatsheets – there is one on purrr!

• Simulating who would be in the first GOP debate (NYT 7/29/15)

• Blog by Aaron Roth, Algorithmic Unfairness Without Any Bias Baked In.

• Simulating linear models from the ISCAM applets.

:::