130 likes | 710 Views
Within Subjects / Repeated Measures Design. What are they? Advantages Deal with error variance Increased power Disadvantages: Carry over effects. Sources of Carry Over. Learning Fatigue Habituation Adaptation Sensitization Contrast. Dealing with Carry Over.
E N D
Within Subjects / Repeated Measures Design • What are they? • Advantages • Deal with error variance • Increased power • Disadvantages: Carry over effects
Sources of Carry Over • Learning • Fatigue • Habituation • Adaptation • Sensitization • Contrast
Dealing with Carry Over • Complete Counterbalancing (balanced across individual subject) • Block randomization • ABBA counterbalancing • Incomplete Design (different order for different subjects) • All possible orders • Partial counterbalanced • Latin square design
Block Randomization • Break up the number of times the participant is exposed to the various levels into blocks • # of blocks = # times expose to each level • Size of the block = # conditions • Averaging process • Each level has an equal likelihood of being at the beginning, middle and end of presentation
Example • IV: encoding strategies with 3 levels • Imagery, rehearsal, control • Want to expose you to these techniques 4 dif times • Trial Condition • 1 C • 2 I • 3 R • 4 R • 5 I • 6 C • 7 I • 8 R • 9 C • 10 C • 11 R • 12 I
ABBA • Balance effects by presenting one sequence and then following it with its opposite • Appropriate when practice effects are linear • ACBBCA • Even # of repetitions • ACBBCAACBBCA • Not ACBBCAACB • Problem: anticipation effects
Incomplete Designs • Each condition must appear in each ordinal position equally often (different Ss) • All possible orders • With 2 conditions 2 orders (AB or BA) • With 3 conditions 6 orders (ABC, ACB, BCA, BAC, CAB, CBA) • N! # of orders, so limited to 4 or fewer
Example • Subjects T1 T2 T3 • 1 1 2 3 • 2 1 3 2 • 3 2 1 3 • 4 2 3 1 • 5 3 1 2 • 6 3 2 1 • ETC
Partial Counterbalanced • Latin Square • Construct in 4 steps • Arrange first N letters of alphabet in alphabetical order; then rotate first letter into last position • Randomly rearrange the column order using a random number table • Randomly rearrange the rows • Randomly assign treatments to letters
Column • 1 2 3 4 • 1 A B C D • 2 B C D A • 3 C D A B • 4 D A B C
Treatment order as IV • Mixed design • 2 X 2 • Type of encoding X Order • Measure size of the carry over effect • Limitations: keep number of treatment orders small, otherwise a complex, demanding experiment
Minimize Carryover • Reduce carryover, reduce error variance, which increases power • Not all sources can be minimized • Give practice trials • Allow time to habituate, adapt, and/or rest (fatigue) • Typically a combination of methods
When to Use • When subject variables or individual difference variable correlated with DV • Economizing on number of subjects • Assessing effects of increasing exposure