CHP. 10 - Simple Experiments Flashcards
Design Confound
Part of the research design or methodology that results in a confounding variable that varies systematically w/the levels of the IV
- something else is changing other than the IV
Control Group vs. Comparison Group
All comparison groups are control groups, but not vice versa
Systematic Variability
Assignment of subjects is somehow unequal across groups
- causes a difference between groups
- affects 1 group, not the other
- threatens internal validity, Type 1 Error
Unsystematic Variability
Normal individual differences between subjects within a group/condition that isn’t a confound
- effects all groups equally
- can cause a null result, Type 2 Error
Selection Effects
Happens when participants choose which group they want to be in
- problem w/selection process
Matched Groups
Used to ensure equal representation of an important variable
- randomly assign participants to each level of the IV
Independent Groups Design
Subjects placed into different levels of IV
- Post-Test Only vs. Pre-Test & Post-Test Design
Within-Groups Design
All subjects are presented with each level of the IV
- The order of each level is randomized
- Repeated Measures vs. Concurrent Measures Design
Post-Test Only Design
DV measured only after manipulation
Pre & Post-Test Design
DV measured before & after manipulation
Training Effects
Pre-testing along can change the behavior on the post-test
Repeated Measures Design
Dependent measures are measured for all subjects for each level of the IV
- separate orders
Concurrent Measures Design
Participants exposed to all levels of the IV either at the same time or in a sequence within the same experimental session
Order Effects
“Training/practice” or “fatigue”
Carryover effects - earlier trials affect future trials
- Concern for Withing Groups Designs
- Solution: Counterbalance Design
Full Counterbalance
All possible orders represented
Partial Counterbalance
Used when there are too many conditions relative to the # of subjects
Demand Characteristic
Experiments have cues that may lead participants to guess the hypothesis & may act differently
- Least likely to occur in an independent group post-test only design
Pilot Study
Simple study in small group of subjects
- proof of concept to confirm the effectiveness of manipulations
Manipulation Check
Measures a variable to confirm if the animal model was successfully created
Cohen’s “d”
Standardized effect size
- 0.2 = small/weak
- 0.5 = medium/moderate
- 0.8+ = large/strong
