Stat 418 – Day 2 Binomial Inference (Ch. 1)
Syllabus How is JMP downloading going? Optional subscription
Last Time – AFI films Chelsea and Eric Katie and Julia Victor and Justin
Last Time Definition of categorical data, types of categorical data Recall: Be careful in defining variables and observational units in the first place Recall: Segmented (stacked) bar graphs and conditional proportions as a way to examine association between variables Recall: explanatory vs. response variables
Unusual episode? What observations did you make?
Lessons Many different explanations are reasonable Be explicit in your assumptions What is known vs. speculation Get in the habit of really backing up your statements with specific numbers Counts vs. proportions Distinguish between “more likely to be exposed” and “if exposed, more likely to die”
Binomial random variables Fixed number (n) of trials Each trial has two possible outcomes (S, F) Trials are independent Probability of success on one trial does not depend on outcomes of other trials Constant probability of success ( )
Binomial random variables
Rock-paper-scissors Let Y represent the number of times you throw scissors Binomial? 10 trials Scissors or not scissors Independent trials? Constant probability of success? If assume you choose equally among the three each time
Exact binomial test Define the parameter of interest State null and alternative hypotheses about the parameter Calculate a p-value See online instructions for JMP or applet Make a conclusion (in context)
For Wednesday Continue reading Ch. 1 (mid Pvalue method) Continue working on HW 1