Chapter coverage

1
Chapter coverage

• Part A
• 1 Practical tools
• 2 Consulting
• 3 Design Principles
• Part B (4-6) One-way ANOVA
• Part C (7-9) Factorial CRD
• Part D (10-12) Unbalanced CRD

2
Chapter coverage

• Part E (13-15) Questioning Assumptions
• Part F (16-18) ANCOVA
• Part G (19-21) Random and Mixed Effects
• Part H Nested and Split Plot Designs
• Part I Repeated Measures and Cross-over Designs

3
Important Design Principles

• Replication
• Randomization
• Blocking
• Multifactor Studies
• Confounding

4
Replication

• A repetition of the basic experiment
• Obtain estimate of experimental error
• Increase power
• False replication
• Experimental Unit
• Observational Unit

5
Randomization

• Random assignment of treatments to experimental
  units
• Random run order
• Random selection of experimental units from a
  population

6
Randomization

• Randomization restrictions
• Stratification
• Clustering

7
Restricted Randomization
8
Systematic Sampling

• Easy to implement
• Possible loss of efficiency
• Beware systematic sources of variation

9
Restricted Randomization
10
Blocking or Stratification

• Inference on subgroups or strata
• Convenience

11
Restricted Randomization
12
Clustering

• Convenient, Efficient (in terms of experimental
  effort)
• Possible loss of statistical efficiency
• Important in adaptive sampling

13
Blocking

• Grouping to remove sources of heterogeneity
• Paired design is a form of blocking

14
Multifactor Studies

• One at a time (OAT) designs
• Multifactor designs
• Orthogonality allows us to use multifactor
  designs to estimate factor effects and
  interactions

15
Confounding

• Confusing factors in a design
• Good confounding
• Bad confounding

16
Complete Confounding in a 2x2 table
17
Confounding in a Randomized Complete Block Design
(BxT is confounded with Error)
18
Chapter 3--Completely Randomized Designs

• CRD Exercise
• Review 5.2,5.3,6.1-6.4
• Class
• 5.1,5.4,5.5