STAT 516 -- FINAL EXAM REVIEW SHEET

(The final exam will be ROUGHLY 70% new material and 30% from Exam 1 
and Exam 2 material.  Also study the review sheets for Exams 1 and 2.)

I.  Design Of Experiments
  A. Purpose of Designing Experiments
   1. Disadvantage of factorial design
   2. Reliability of an experiment
   3. Experimental error variation
   4. Why should we reduce experimental error?
  B. Randomized Block Design
   1. Purpose of blocks
   2. Examples of block designs
   3. Linear model for RBD
   4. Random block effects
   5. Treatment-block interaction
   6. Expected MS in RBD
   7. Testing for treatment effect
   8. Testing for variation among blocks
  C. Randomized Block Design with Sampling
   1. Replication within treatment-block combinations
   2. Experimental error vs. Sampling error
   3. Expected MS in RBD with Sampling
   4. Appropriate denominators for various F statistics
  D. Latin Square Designs
   1. Two blocking factors
   2. Properties of Latin Square design
   3. Number of levels of each factor
   4. Linear model for Latin Square design
   5. Advantage of Latin Square
   6. Disadvantage of Latin Square
II. Other Linear Models
  A. General Linear Model Concept
  B. Dummy Variables
   1. Linear Model for One-Way ANOVA using dummy variables
   2. Extra restriction on parameters
  C. Unbalanced Data
   1. Adjusted means (LSMEANS)
   2. Why must we adjust our marginal means?
   3. Type I vs. Type III SS
  D. Analysis of Covariance Models
   1. Type of data for which ANCOVA is suitable
   2. Combining one-way ANOVA and SLR
   3. Covariate's connection with blocking
   4. Testing for covariate's effect
   5. Testing for factor's effect
   6. Interpreting coefficient for covariate
   7. Using more than one covariate
   8. Unequal-slopes ANCOVA model
   9. How to test whether we need unequal slopes?
  E. Logistic Regression
   1. Binary response variables
   2. Disadvantage of standard SLR model
   3. Form of logistic regression model
   4. What does E{Y|X} represent?
   5. Shape of logistic curve and role of beta_1
   6. Odds, log odd, and logit transformation
   7. Maximum likelihood estimation
   8. SAS analysis: PROC LOGISTIC
   9. Odds ratio and its interpretation
   10. Hosmer-Lemeshow test
   11. Testing whether X is significant in logistic regression model
III. More Advanced Topics
   A. Multiple Logistic Regression
    1. Extension of Logistic Model to Multiple Logistic Regression
    2. Inferences in Logistic Regression
      a. LR test about all betas
      b. Wald test about a single beta_j
      c. CI for beta_j or for associated odds ratio
    3. Model Selection and AIC
    4. Pearson Residuals and Outlier Diagnostics
    5. CI for "Mean Response" E(Y_h|X)
      a. Point estimate
      b. Interpretation of CI
   B. Poisson Regression
    1. Count response variable
    2. Interpretion of beta_1-hat in terms of exp(beta_1-hat)
    3. Inference about regression parameters
    4. Goodness of fit: residual deviance and Pearson X^2
    5. Deviance residuals and Pearson residuals
    6. Predicted mean response values and CI for mean response