Lecture 8

Question 1

What is the goal of data analysis?

Answer 1

To understand underlying processes and find meaningful associations between variables.

Question 2

What are some reasons for misleading associations in data analysis?

Answer 2

Not robust, Reverse causation, Common cause, Indirect effect, Common consequence.

Question 3

What is the multiple testing problem?

Answer 3

Running many tests increases false positives.

Example: 100 million tests can lead to 5 million false positives at p ≤ 0.05.

Question 4

When is the binomial test used?

Answer 4

When testing proportions for a binary outcome (e.g., coin tosses).

Question 5

What is the null hypothesis in a coin toss experiment?

Answer 5

The coin is fair: H0:p=0.5.

Question 6

How do you perform a binomial test in R?

Answer 6

binom.test(10, 12, p = 0.5, alternative = “two.sided”).

Question 7

What is a contingency table used for?

Answer 7

To summarize frequencies of two categorical (binary) variables.

Question 8

What is Fisher’s Exact Test?

Answer 8

A test for small sample sizes in contingency tables to determine association.

Question 9

How do you run Fisher’s Exact Test in R?

Answer 9

cont_tab <- matrix(c(10, 20, 10, 70), nrow=2)
fisher.test(cont_tab, alternative = “greater”)