Experimental Design

Question 1

What are the two types of experiments?

Answer 1

1. Natural experiment = no treatments applied

2. Manipulation experiments = treatments applied

Question 2

Why is it necessary for replication?

Answer 2

1. to estimate experimental error = variability

2. insurance against the intrusion of chance events in experiment (noise)

Question 3

What are types of experimental designs?

Answer 3

1. completely randomized design
2. randomized block design
3. Latin-square design
4. systematic design

Question 4

What is the completely randomized design?

Answer 4

• experimental units are randomly placed

- NOT recommended when few replicates

Question 5

What is randomized block design?

Answer 5

• experimental units are grouped together in blocks
• each block has experimental units = treatments
• corrects for effects of single environmental gradient
• place blocks perpendicular to gradient

Question 6

What is the disadvantage of completely randomized design?

Answer 6

• lead to loss of disproportionate number of one particular treatment if grouped together by chance

Question 7

What is systematic design?

Answer 7

• samples placed in systematic fashion

- could have natural trends where experimental units are randomly placed

Question 8

What is the Latin-square design?

Answer 8

• used when TWO factors (gradients) are known to exist
• corrects for effects of two gradients
• first two blocks are horizontal and vertical starting at upper left

Question 9

What are the requirements of Latin-square design?

Answer 9

1. all treatments must be present in each block
2. treatments must NOT be duplicated in a block
3. when assigning treatments to adjacent blocks the same treatment cannot share a common border

Question 10

What are pseudo replicates?

Answer 10

• replicates that are NOT independent

- consequence of assuming that sub samples are independent of each other and come from independent EU

Question 11

What are examples of pseudo replicates?

Answer 11

1. A 1000 ha plot of forest is burned and another plot is not burned, Sample 15, 10 meter square quadrats on each plot
2. A 1 ha plot of grassland is fenced and an adjacent 1 ha plot is not fenced. Sample 50, 1 meter square quadrats on each plot

They have n = 1 and df = 0

Question 12

What is the solution to pseudo replication?

Answer 12

Establish more experimental units