Topics List: Exam 2, Spring 2026
Material covered
- Lectures 8-15, HW 4-6, Labs 4-6, AE 2-4
Coding
pnorm()andqnorm()andpt()andqt()slice()andnames()andgetwd()andsummarize()mutate()to create new variables,select()to select variables, andarrange()to arrangeReview using
ggplot()to create various visualizations, such as boxplots, histograms, scatterplots etc. Alsogeom_sf()for maps.What is the
sfpackage used for? When would you use it?st_read()What are some examples of metadata?
Be able to read/understand code using R functions for ANOVA, t-tests, power:
aov(),t.test(),power.t.test,summary()(used assummary(aov(y~group))).Be able to recognize good and bad coding style in example code, and be able to suggest a correction. Topics include: proper spacing, curly braces, long lines, pipes, layering in
ggplot2()
Conceptual and Applied
What’s the basic idea behind the Central Limit Theorem (CLT)? Why does it matter?
Be able to find approximate distribution (based on CLT) of sample averages given a population mean and standard deviation. (IQ example in Lecture 8)
What is the rule of thumb (sample size) where CLT kicks in?
What is a confidence interval? What is the basic form of a confidence interval?
What is a confidence multiplier? How to find confidence multiplier in R using
qnorm()?What is a margin of error?
How do you correctly interpret a confidence interval? Why can’t we say “There’s a 95% chance that…”? Know how to interpret a confidence interval, (asthma and SBP examples in Lecture 9).
How does the t distribution relate to the normal distribution? How are they different? Why would we use the t-distribution instead of the Z?
How many parameters does the t-distribution have?
Know how to find the cumulative probability and quantiles from a t-distribution using
qt()andpt().What is p-hacking?
Basic steps to hypothesis testing framework
Know how to write null and alternative hypotheses in words and symbols
Definition of a p-value
Interpret
t.testoutput in RDefinitions of Type 1 and Type 2 errors. Name each type of error for a given context (e.g. pregnancy test or HIV test, identify which type of error might be more severe and you would therefore want to minimize)
Dangers of using a one-sided test
Independent sample t-test vs. paired/matched t-test
When would you use a pooled variance estimate vs. a Satterthwaite approximation?
What type of research question can you answer with an ANOVA? State null and alternative hypotheses, and assumptions of ANOVA.
What is the issue of “multiple comparisons”? Why do we want to avoid them? What is a Bonferroni correction?
What is the basic idea of an F-test in ANOVA? (What is it testing?) How do you get the numerator and denominator degrees of freedom? Be able to interpret R output from doing an F-test (pet data example in Lecture 12).
What is a chi-square test for independence of two categorical variables? State null and alternative hypotheses and assumptions. What are the degrees of freedom? What are “expected” vs. “observed” counts?
When would we use a Fisher exact test instead of a chi-square test?
At a high level, what is the advantage of nonparametric hypothesis testing vs. parametric testing?
What are the nonparametric alternatives to paired t-test, independent samples t-test, ANOVA?
What’s the definition of power?
When we vary sample size, variance, alpha, and minimum detectable difference, how is power affected?
How to use
power.t.test()to calculate sample size and power.