research methods
DESCRIPTION
Research Methods. Descriptive Methods Observation Survey Research Experimental Methods Independent Groups Designs Repeated Measures Designs Complex Designs Applied Research Single-Case Designs and Small-n Research Quasi-Experimental Designs and Program Evaluation. Experimental Methods. - PowerPoint PPT PresentationTRANSCRIPT
Descriptive Methods◦ Observation◦ Survey Research
Experimental Methods◦ Independent Groups Designs◦ Repeated Measures Designs◦ Complex Designs
Applied Research◦ Single-Case Designs and Small-n Research
◦ Quasi-Experimental Designs and Program Evaluation
OVERVIEW
WHY RESEARCHERS USE REPEATED MEASURES DESIGNS
THE ROLE OF PRACTICE EFFECTS IN REPEATED MEASURES
DESIGNS
◦ Defining Practice Effects
◦ Balancing Practice Effects in the Complete Design
◦ Balancing Practice Effects in the Incomplete Design
DATA ANALYSIS OF REPEATED MEASURES DESIGNS
◦ Describing the Results
◦ Confirming What the Results Reveal
THE PROBLEM OF DIFFERENTIAL TRANSFER
Repeated Measure Designs
Repeated Measure Designs◦ Within-subjects designs
Subjects are repeatedly tested Each subject in all conditions Subjects as their own controls practice and fatigue effects
◦ Improvement with practice◦ Worse with fatigue/reduced motivation
No elimination but balancing◦ Averaged across the conditions
Conduct experiment with few participants
◦ Special populations (individuals with brain injuries)
Conduct experiment more efficiently
Increase sensitivity
◦ Ability to detect the effect
◦ Minimize error variation
Study changes in behavior over time
Same individuals in each condition:◦ No confounding on individual differences variables
Practice effects: Change because of repeated testing (not because of the independent variable)
Practice effects = threat to internal validity ◦ If different conditions are presented in the same
order to all participants Two types of RMD (complete and incomplete)
◦ Differ in the ways to control for practice effects.
Block randomization◦ Random order of all condition on each presentation◦ Blocks = No of administrations of each condition (Tr/Cd)◦ Blocks Size = Number of conditions◦ Balancing = Avg presentation of each condition shall be
equal ◦ Avg Pc = Σ (cd number) / Blocks
ABBA counterbalancing◦ Random sequence followed by opposite sequence◦ Suitable for small number of conditions and trials◦ Balancing = 2 trials
Non-linear practice effects or Anticipation effects◦ Block randomization >> ABBA
Balanced across subjects Each condition in each ordinal position
◦ Orders = N! ◦ N=Conditions◦ Participants = Any Multiple of all Possible Orders
<=4 conditions : Use all possible orders Random Assignment Methods for orders selection
◦ Latin Square and Random starting order with rotation
◦ Orders = Any Multiple of Conditions
Latin Square
◦ each condition at each ordinal positions once
◦ each condition precedes and follows each other condition
exactly once
Random starting order with rotation
◦ Begin with a random order
◦ Rotate sequence systematically
Errors and outliers (data scanning)
A summary score (e.g., mean, median)
◦ Each participant (incomplete design)
◦ Each participant in each conditions (complete design)
Descriptive statistics
◦ Performance across all participants
◦ For each condition of IV
Probability testing Same as in random group design
◦ Null hypothesis testing◦ Confidence interval
Error variation◦ More sensitive◦ Not cause of individual differences◦ Difference in ways the conditions effect the
results
Persistence of effects of one condition Influence performance in subsequent
conditions Common with instructional variables
◦ Threat to internal validity of RMD Identification of DT
◦ Same variables in RMD and RGD Use RGD