MS&E 107 Lecture Notes - Lecture 10: Confidence Interval, O-Ring, Statistical Hypothesis Testing

• School: Stanford University
• Department: MS&E - Mgmt Science & Engineering
• Course: MS&E 107
• Professor: Sam Savage

Document Summary

Utilization rate goes up, the queue goes crazy. This week, forecasting software: xlforecast. xlam, sipmathtools (on probability management, statistical forecasting. Temperature vs. degree of o ring damage on spaceships graph. Postpone the launch or not postpone the launch. Below 65 always o ring damage, above 75 always no damage. Coin tosses : use the data to simulate new data. Devil"s advocate position: temperature has nothing to do with o ring damage. Red word for devil"s advocate"s position: null hypothesis. Probability that it did happen by chance: red word is p-value. How to test: scramble the order in which these appear. Ex: generate column of random numbers, lay out 24 cards that are numbered, rank the numbers. Scramble and sort - f9 key: the original data is scrambled around in a random order. Multiplication: the only way to get four bad o rings at the top and seven perfect at the bottom. But these are the only ones that came back!