STATISTICS 135,  FALL 2006

INSTRUCTOR:    Ani  Adhikari

Course grades are ready available on Bearfacts.  I am very pleased with the performance on the class.  More details, including how to see your final exam score, are here.

Course description

Banish your fear of covariance. Do these!

 Probability prerequisites (READ!!)

The final exam is on Friday 12/15, 5 p.m. to 8 p.m., in 60 Evans.  I will have office hours Wed 12/13 and Thurs 12/14, 10-2.

You may bring to the final a calculator and two 8.5x11 pages (= 4 sides) of notes.

Summary of material in the second half of the term
Not on the final: moment generating functions, delta method, sufficiency, partial F-tests.  As in the midterm, you may use without proof any result proved or used in class, section, or homework.  I will not ask you to prove them again, though I may ask you to prove or derive other things.  R output in the questions will be largely self-explanatory, and I will explain anything unusual.  You are expected to recognize the standard inference commands that you have used in homework (e.g. oneway.test, lm, etc) and to know exactly what the assumptions are for each one as well as what the output means.
The handout on inference in simple linear regression
(This was handed out in class so most of you have it already.)

Course contents

Summary of midterm results: Scores range from 7 to 49; six people got more than 40.  There's plenty of symmetry in the distribution. The median is 23.5 and the mean is 23.9.  Partial credit was given when you made silly mistakes, so don't get all anxious when you look at the answers here.

Summary of material for the midterm

What you are expected to know, half-way into the semester
Homework (now includes a page on R and regression)

R stuff
If you don't know how to get R to do something:  browse An Introduction to R and look at the reference card.  Those cover a lot of ground.  
E.g. for HW 4: how to generate random numbers, how to find percentiles (quantiles) of a dataset, and how to get  maximum likelihood estimates (there's even a worked example of this).  If the command contains an option you don't understand, you probably don't need it.


What can you do with statistics?
Information about the VIGRE seminar, Wed 12-1 in 1011 Evans, where you can hear about "applications of statistics, about graduate study, and about opportunities for employment as statisticians.''   Highly recommended - "This semester there will be speakers from LBL, Yahoo, the National Security Agency, eBay, a political polling firm, UCSF, and Berkeley.''