Experimental Designs

Question 1

Factorial designs

Answer 1

• when more than one independent variable is used
• 3 main types:
  
  • between subject
  • within subject
  • combination of within and between:
    1. Mixed
    2. Nested
• each variable or factor has 2 or more levels
• use ANOVAs to analyze

Question 2

Between subject designs

Answer 2

2 types:

1. completely randomized
2. Matched group designs (must ensure groups are equal so match one participant from one group with one from other group to ensure equal STD)

Question 3

Within subject designs

Answer 3

• also called repeated measures designs or randomized-blocks designs
• 3 main types:
  
  1. same subject observed under all treatment conditions
  2. same subject observed before and after a treatment (pre+post test design)
  3. subjects are matched on a subject variable (organismic variable or individual difference variable) and then randomly assigned to the treatments
     
     • is actually a between subject design requiring a within subject ANOVA

Question 4

Within subject design advantages

Answer 4

• need less subjects
• each level of the independent variable is applies to all subjects, so we can evaluate how each level of the independent variable affects each subject
• each subject is its own control
• excellent for assessing experiments on learning, transfer of training, practice effects
• may help increase statistical sensitivity or power (subjects are not divided into groups, all subjects involved in all conditions)
• small n research benefits from within designs

Question 5

Within subject design disadvantages

Answer 5

• practice effects: participants get better at it over time
  
  • (if not focus of study is a problem)
  • solution: appropriate counterbalancing procedures can counteract effects + make treatment order an independent variable (change it to see if it has practice effect)
• differential carry over effects: lingering effect of one or more treatment condition (often an issue in drug studies)
  
  • solution: recovery periods
  • anti-depressants good example because some have long lasting effects
• violation of statistical assumptions: not enough people etc
  
  • solution: use more strict significance level

Question 6

Carryover effects

Answer 6

• fatigue: decreased performance with time
• contrast: treatments are compared by subjects
• habituation or sensitization: more exposure to stimulus causes increased or decreased sensitivity
• adaption: tolerance (eg drug studies)

Question 7

Nested designs

Answer 7

• similar to mixed design but levels of factor A found under difference levels of a factor B ARE NOT THE SAME
• usually because of a constraint
• eg. Difference location

-groups would be: a1b1, a1b2, a1b3… a2b4, a2b5, a2b6

• spatial: multiple samples of a single tissue type within a rat
  
  • estuaries are unique to each river
• temporal: sub-samples in time can only be sampled at one time and not another

-more economical but some interactions cannot be evaluated

Question 8

Interactions between variables

Answer 8

• interconnectivity
• variables: x, y, z
• main effects: x, y, z
• interactions: xy, xz, yz, xyz
• you have an interaction when the effect of two or more variables is not additive
• interactions make interpretation of experimental data more challenging
• a significant interaction will often mask the significance of main effects

Question 9

Additive interactions

Answer 9

• effect of each independent variable/factor doesnt interact with the other
• eg. Effect of phototherapy and melatonin on sleep quality. then the effects of PT doesnt impact melatonin, and melatonin doesnt impact effects of PT

Question 10

Interactive interactions

Answer 10

• when the effect of on factor plays a significant role on the effect of another factor
• may look like the lines of two factors crossing/meeting
• as one factor increases it may change how the other factor works or affects the dependent variable
• if the results of one factor depend on another factor there IS an interaction

Question 11

Ordinal interaction

Answer 11

• on a graph there will be no overlap of the of the data

- eg. Lines of PT vs no PT with melatonin dont cross/overlap but there is still an interaction

Question 12

Disordinal interaction

Answer 12

• looking at a graph of both variables there will be an overlap in data
• eg. Lines of PT vs no PT with melatonin dosage will cross over or overlap at at least one point

Question 13

Antagonistic interactions

Answer 13

• certain kinds of interactions can MASK the main effects of one or more variables
• eg. (Graph with lines creating X shape) the independent variable melatonin is effective, but the statistical analysis fails to reveal statistically significant main effects for melatonin
• analysis shows that there is no numerical difference in the averages, but the graph shows that the variables clearly have different effects

Question 14

Having too many factors/levels

Answer 14

• interactions will be harder to explain
• more factors equals more possible interactions
• need more power for more interactions