9. Experimental Design

Question 1

How can we reduce bias

Answer 1

• inclusion of control groups
• Randomization of treatment across units
• Block randomization
• Stratified randomization
• Blinding so neither experimenter nor subject know
  treatment assignment

Question 2

How can we reduce sampling error?

Answer 2

• Replication
• Balance in number between treatment and blocking

Question 3

What’s the advantage of multiple treatments?

Answer 3

Allows interactions and biological detail to be revealed

Leads to factorial designs

Question 4

What can you do if u cannot use designed experiments?

Answer 4

• Stratified sampling
• Matching
• Adjustment

Question 5

How can we choose an appropriates sample size?

Answer 5

check for precision using the magnitude of error
= 2SE
n=8(Var/MOE)^2

check for power using set equation
(power 0.8, sig 0.05)
n=16(Var/minDiff)^2