ST514 -- Statistics for Management and the Social Sciences II
Fall 2007 --- 3 credits

---

Instructor: Brian Reich (reich@stat.ncsu.edu)
Office hours: 209H Patterson Hall, MW 2-3
Course times/location: MW 3-4:15, Harrelson 210
Syllabus: http://www4.stat.ncsu.edu/~reich/st514/Syllabus

Take-home final exam

---

Half of you final exam grade will be a take-home data analysis project. The project is due in class on 12/14. Click here for a full description



Homework Assignments

---

• HW #1 (Due 9/5): 1.8, 1.24, 1.30, 1.42, 1.45, 1.46, 1.60, 1.93
  
• HW #2 (Due 9/17): 3.6, 3.10, 3.12, 3.18a, 3.22
  
• HW #3 (Due 9/26): 3.31a, 3.42abcd, 3.47, 3.52, 3.62, A.1a (page 726), A.13 (page 730), verify that for simple linear regression with the mean of x equal to zero, the least squares matrix solution on page 735 is equivalent to the least squares estimates on page 96).
  
• HW #4 (Due 10/10): 4.3 (use a two-sided test in b), 4.9 (also, turn in SAS output that matches the SPSS output on page 180), 4.15
  
• HW #5 (Due 10/22): 4.48, 4.50, 5.15, 5.23
  
• HW #6 (Due 11/5): 6.2, 6.3, 6.5. For 6.5 also report and compare the models selected by the five procedures discussed in class: forward, backward, and stepwise regression (using alpha = 0.05 for entry and exit) as well as all-possible-regressions selection using both Cp and Adj-R2. Trim the SAS output to be at most 3 pages.
  
• HW #7 (Due 11/19): 7.8, 7.18, 7.19, 8.6, 8.14.
  
• HW #8 (Due 11/28): 8.20, 8.21, 8.24, 8.28. Also for problem 8.24 include a plot of the residuals from part e's model to inspect the normality assumption. Is the normality assumption reasonable for these data?
  
• HW #9 (Due 12/5): 9.16, 9.17, and analyze the DVD data using proc GAM.
  • For 9.16c, use the SPSS output (assuming beta-hat1=0.0001) to specify an equation that gives the probability of owning a digital organizer as a function of annual income and plot this probability as a function of income from income = 0,..., 100,000.
  • For 9.17c give an estimate of the mean, not a confidence interval.
  • For the DVD data, fit a GAM model with dummy variables for Genre and a spline function with 5 degrees of freedom for Box. Plot the fitted curve for Box. Is there evidence that the relationship between DVD and Box is non-linear?
  
  

Course Ouline

---

Dates -- Description (Chapters)
8/22 -- Review of basic concepts (1)
8/27-8/29 -- Review of basic concepts (1), Introduction to regression analysis (2)
9/3-9/5 -- Labor Day; Simple linear regression (3)
9/10-9/12 -- Simple linear regression (3)
9/17-9/19 -- Regression using matrices (Appendix A)
9/24-9/26 -- Multiple regression (4); Exam 1 (9/26; will cover Chapters 1, 2, and 3 and Appendix A)
10/1-10/3 -- Multiple regression (4)
10/8-10/10 -- Multiple regression (4)
10/15-10/17 -- Model building (5)
10/22-10/24 -- Variable selection (6)
10/29-10/31 -- Regression pitfalls (7); Exam 2 (10/31)
11/5-11/7 -- Regression pitfalls (7)
11/12-11/14 -- Residual analysis (8)
11/19-11/19 -- Residual analysis (8)
11/26-11/28 -- Special topics in regression (9)
12/3-12/5 -- Special topics in regression (9); Time series Analysis (10)
12/14 -- Final exam, 1-4pm, 210 Harrelson

Computing/Data

---

• Basketball data
  
  
• DVD data
  
  
• Annual temperature data
  
  
• Loading a data set and making a histogram
  
  
• Loading a data set using SAS commands in the editor window
  
  
• Simple regression in SAS
  
  
• R code to plot regression lines
  
  
• Simple regression analysis of global surface temperature data
  
  
• Prediction intervals using the smoking data
  
  
• Multiple linear regression in SAS
  
  
• CARS data
  
  
• Confidence interval applet
  
  
• Multiple linear regression in SAS using proc reg
  
  
• Polynomial regression in SAS using the motorcycle data
  
  
• Modeling interactions
  
  
• Categorical predictors/Variable selection with the fast food data
  
  
• Residual analysis of the DVD data
  
  
• Residual analysis of the PGA golf data
  
  
• Logistic regression analysis of the basketball data
  
  
• Nonparametric regression analysis of the LA air pollution data
  
  
• Time series analysis of the global temperature data
  
  

Solutions

---

• Exam 1
  
  
• Exam 2
  
  
• Final exam
  
  

*Thanks to Lisa Denogean for help in course preparation.