CHAPTER 11 Small N Designs
Research in Psychology Began with Small N Most early research in psychology… Small N Data not summarized; data for each subject presented Additional subjects used for replication Dresslar’s facial vision study Also a good example of falsification thinking
Reasons for Small N Designs Misleading results from statistical summaries of grouped data A failure of individual-subject validity e.g., discrimination learning research in children Grouped data suggested gradual learning Individual data suggested hypothesis testing and quick solutions once proper hypothesis found Solution children reach criterion at different rates
Reasons for Small N Designs Practical problems with large N designs Participants with a particular attribute are rare e.g., obsessive-compulsive disorder occurs in < 2% of population Members of a specific animal species are rare, costly, or require much time for training
The Experimental Analysis of Behavior Operant conditioning Behavior conditioned, in a particular environment, by consequences (e.g., reinforcement) Primary DV rate of response Recorded cumulatively
The Experimental Analysis of Behavior Applied behavior analysis Contemplative (understanding basic causes of events) vs. technological ideal (using science for control) Skinner using science to achieve control The issue of control (Box 11.2) Comparing Watson and Skinner Attempt to improve society Use of punishment for controlling self-destructive behaviors Justified? Judge Rotenberg Center
Small N Designs in Applied Behavior Analysis Elements of single-subject designs Importance of operational definitions (again); what constitutes a child’s “problematic behavior”? Baselines established Introduce some treatment Simplest A-B design A = baseline B = treatment Withdrawal designs A-B-A Treatment (B) introduced, then withdrawn To see if behavior change correlates with treatment change
Small N Designs in Applied Behavior Analysis Withdrawal designs A-B-A-B design Treatment evaluated twice Experiment ends with treatment in place Ideal pattern:
Small N Designs in Applied Behavior Analysis A-B-A-B design Research example 34: Differential Reinforcement of Alternative Behavior Peer reinforcement of on-task behaviors for ADHD children External validity issue Laboratory study might not generalize to classroom
Small N Designs in Applied Behavior Analysis Multiple baseline designs Three varieties of multiple baseline One behavior, two or more subjects Two or more behaviors, one subject Two or more environments, one behavior, one subject
Small N Designs in Applied Behavior Analysis Changing criterion designs For behaviors that must be changed gradually Operant principle of shaping
Small N Designs in Applied Behavior Analysis Alternating treatments design Compares relative effectiveness of two or more of treatments
Small N Designs in Applied Behavior Analysis Evaluating single-subject designs External validity issues Visual inspection instead of statistical analysis, although this is changing Hard to test interactive effects; effect on one type of person but not another?
Case Study Designs Detailed analysis of single individual or event Incorporates other methods (e.g., observation, psychometric tests) Individual Both clinical and non-clinical cases Effects of head trauma in boxer: Demonstrated targeted treatment plan for speech problems, independent of cognitive functioning The mind of a mnemonist (Box 11.3), autstic savants Event Response to catastrophic event
Case Study Designs Evaluating case studies Level of detail not found elsewhere Can serve falsification (Remember Hans?) Limited control External validity issues Faulty memory problems