Topical Review 2
Fall 2023
Spring 2024
Hypothesis Testing, Bootstrap, A/B Testing, and
Confidence Intervals
We will start at BerkeleyTime!
,Hypothesis Testing Basics
,Hypothesis Testing: Terms to Know
● Null hypothesis - a claim about the world--a specified chance model
● Alternative hypothesis - there is some reason (other than chance)
why the data in the observed sample differ from the null hypothesis
○ This comes from the set up of the question! Read carefully to see
whether we care about the direction or not.
● Test statistic - a statistic used to evaluate samples, e.g. TVD or
average. This takes a sample and represents it in a single number.
○ Indicator of whether a sample is consistent with the Null or not
(Alternative).
, Hypothesis Testing: Terms to Know
● Observed test statistic - value of the test statistic for an
observed sample
● P-value - likelihood that the test statistic generated under
your null hypothesis is equal to or further in the direction
of the alternative than your observed test statistic