Statistics 102

Announcements

• I will have a review session for questions related to the final exam next Monday evening (May 8), from 8:00 pm -- 9:30 pm in SH-DH 351 (not room 215 as previously announced).

• I will have office hours Tuesday, May 9, from 3-5 and on Wednesday from 2-3:30.

• I am missing scores for some of you for one or two assignments. To make sure there is no error in the recording of grades, please let me know if you have the missing assignment. If you didn't seem to do any assignments, I omitted your name from the list. I have updated the list to include PS #5 , which quite a few of you evidently did not do.

  Boorsma	ps1, ps3
  Folkinshteyn	ps2, ps5
  Haefele	ps4, ps5
  Hejazi	ps1, ps3, project
  Hunter	ps4
  Khan-Acito	ps4, ps5
  Kinderman	ps1
  Levine	project
  Loughrey	ps5
  Massie	ps3, ps4, ps5, project
  Meyers	ps1, ps4
  Munoz	ps5
  O'Rourke	ps5
  Penske	ps3, ps4, project
  Sawyer	ps3, ps4
  Simon	ps1, ps4
  Stipp	ps4
  Tasugi	ps5
  Tolbert	ps5
  Travis	ps5
  Waluch	ps5
  Yu	ps1, ps2

  If there's an error in this list, simply bring me the marked assignment, and I will fix the grades.

• Some review notes on regression are here . Its 10 pages long, so I do not have copies to distribute. Have a look, and print it if you want a paper copy. Evidently, the printed version is a lot more easy to read than when you view the document with Acrobat.

• The final exam is on Wednesday, May 10, in Annenberg 110 (site of two previous exams) from 4-6 pm. Following the procedures of the university, make-ups for this exam will be given in the Fall Semester (sometime in September).

  The final will be dominated by regression (about 3/4), with the remaining material drawn from other methods. This other material will focus on the use and interpretation of these other methods, rather than the associated details of calculation.

  The make-up final exam is scheduled for Friday, September 8, 2000, from 4-6 pm. If you do not take the final exam on May 10, you will be required to take the make-up final next fall.

• The average for the second exam was 85 with an SD of 14. The median was 91 and mode was 92.

• The mean for the first midterm was about 80 with SD = 13. You can figure out your relative standing from these summaries since the grades are close to normal - albeit a bit skewed.

• I have the impression that some of you are not aware of the grading distribution for the material in this class. I have given them when asked, but not published it here. Sorry, but I thought it was in the class syllabus. Here it is:
  • Homework 10%
  • Project 10%
  • First test 25%
  • Second test 25%
  • Final 30%

• For questions related to the project please look at the list of "frequently asked questions" here .

• For JMP questions, you should always see the TAs at the Stat Lab for help. They are there most of the day M-F and can show you how to use JMP for doing the calculations. Their schedule is on the class web page.

For the datasets from the regression casebook, go to this link.

Supplemental Handouts

Extra handouts, including some JMP files.

1. Class survey data
   Analysis of survey data
2. Power curve JMP file
3. ROC curve JMP file
4. Paired testing handout (computer retail advertising)
   Software Ads jmp file
5. Summary of one and two-sample tests
   

   ---

6. Claimed color percentages for M&Ms (Does not load on some systems!)
7. M&Ms JMP data file
8. Example of chi-square data file Illustrates the use of JMP-IN to compute chi-square values from tables.
9. Mon (March 6): One-way analysis of variance
10. Wed (March 8): Using one-way anova (Mileage comparison JMP data )
11. Mon (March 20): Two-way anova with some review of one-way (Web page experiment JMP data )
12. Wed (March 22): Using two-way anova with a sample of test flight data

    ---

13. Mon (March 27): Introduction to regression
    • Diamond sales data
    • Display space data
    • Cellular telephone subscription data
14. Wed (March 29): Introduction to inference in regression
15. Mon (April 3): Review for second midterm
16. Wed (April 5): Beta and regression
    This example introduces the notion of 'beta' as used in finance. Interpret the results of the calculations with a bit of common sense, as we have not taken into account the cost of borrowing or transactions.
    • Collection of monthly returns
    • Amazon data
17. Mon (April 10): Prediction and outliers in regression
    • Cottages data
    • Display data
    • Philadelphia housing data
18. Wed (April 12): Introduction to multiple regression
    A reading for this class covers an automobile design case .
    • Car data set
19. Mon (April 17): Multiple regression and collinearity.
    • Stock data illustrating collinearity
20. Wed (April 19): Categorical variables in regression.
    • Manager case (from text)
    • Manager data
21. Mon (April 24): Categorical variables, continued
    • Production time case (from text)
    • Production time data
22. Wed (April 26): Summary regression modeling
    • Example project data set
23. Mon (May 8): Review
    • Review of methods
      • Confidence intervals
      • Comparisons of two groups (t-test)
      • Comparisons of several groups (anova)
        • Anova summary and F-test
        • Multiple comparisons (Tukey, Hsu)
      • Regression
        • R2, anova table, and F-ratio
        • Regression coefficients, SE, and t-ratio
          • Marginal vs partial coefficients
          • Transformations
          • Collinearity (VIF)
        • Categorical predictors, interaction
        • Prediction, RMSE, and prediction intervals
        • Assumptions and residual diagnostics
        • Leverage points and outliers

    • Sample data analyses
      • Example data set (parcel example from the casebook, page 148)
      • Example data set (wage discrimination, page 180)

    • Questions on prior exams