Research
Matching
Randomization
Potential problems of each?
Ways of controlling for a priori differences between groups.
Potential problems of each?
Match:- may create mismatch on other criteria (and subject attrition)
Randomization:
- throwing of die: may have to start all over(?)
What procedure is the Formula: n! used?
complete counterbalancing
Effect of a single IV
Main effect
What are carry-over effects?
If you have more than 1 IV, you can run THIS: where one IV is run between and the other is tested within
Mixed Design
The effect of one IV depends on the level of the other IV
Interaction
Randomization
Counterbalancing
Two procedures for controlling carry-over effects
In a 6 X 5 within-groups multifactor experiment, how many IV's are there?
How many levels of each IV are there?
How many groups are needed?
2 IV's
1st - 6 levels
2nd - 5 levels
6 X 5 = 30
randomize order twice
Block randomization
Each treatment occurs equally in each time period of the experiment
Complete counterbalancing
How many IV's are there in a 8 X 2 X 2 between-groups factorial design?
How many levels of eaCh IV are there?
How many groups are needed to carry out this design?
3 IV's
1st - 8 levels
2nd - 2 levels
3rd - 2 levels
8 X 2 X 2 = 32 groups
use 12 random orders of anagrams for each subject
each treatment occurs equally often in each portion of the experiment
Incomplete counterbalancing
Explain the balanced Latin Square formula for a within-subjects experiment involving 7 levels of the IV
Formula: A, B, G, C, F, E, D
S 1st 2nd 3rd 4th 5th 6th 7th
1 A B G C F E D
2 B C A D G F E
3 C D B E A G F
4 D E C F B A G
5 E F D G C B A
6 F G E A D C B
7 G A F B E D C
each treatment proceeds and follows every other treatment equally often
Advantage Latin Square has over randomization