Quantitative Research

advertisement
Quantitative Research
SPED 500
Dr. Sandra Beyda
• Designs that
• Modes
maximize
– Experimental mode
objectivity by using
– Non-experimental
mode
numbers, statistics,
structure, and
experimenter
control
Structure of a Data Based Study
•
•
•
•
•
•
•
Abstract
Introduction
Method
Results
Discussion
Summary & Conclusions
References
True experiment
Quasiexperimental
Single Subject
True Experiments
• Internal validity
• External validity
–
–
–
–
–
–
Randomization
Systematically vary or manipulate
Hold constant
Assume invariance or irrelevance
Matching
Statistical control
Threats to Internal Validity
When there is a statistically significant result (X >
Control)
History
Maturation
Testing
Nonequivalence
Regression
Interactions with Selection
Instrumentation
Mortality
Threats to Internal Validity
When there are non-significantly different
results:
Small sample size
Insensitive dependent variable
Treatment implementation
John Henry Effect
Treatment not long enough
Treatment diffusion
Threats to External Validity
• Population validity
• Ecological validity
• Operations validity
•
•
•
•
Posttest only
Pretest-Posttest
Solomon 4 Group Design
Factorial Design
• No Random Assignment!
• One group
• Time series approach
• Two groups
• Static group comparison
• Non-equivalent control group
• Reliable
measurements
• Repeated
measurements
• Baseline
• Introduction of
treatment
• Reversal or
withdrawal
• Maintenance
• Replication
Single Subject Research Designs
AB
ABA
ABAB
Multiple Baseline
How about a little practice exercise?
Don’t worry. I’ll wait!
• Descriptive
• Comparative
• Correlational
• Ex Post Facto
Descriptive Research
• Single Group
• Cross-sectional
• Longitudinal
Comparative Research
Correlational Research
• Purpose: What is the relationship between variable
A and B?
• Uses:
– Theory building
– Establish predictive relationships
Correlational Research
• Bivariate correlational studies
• Prediction studies
– Predictor variable
– Criterion variable
Correlational Research Limitations
• Correlation and
Causation
• Spurious
Correlations
• Size of Correlation
Coefficients
• Purpose:
– To examine how an identified independent
variable effects the dependent variable when
circumstances do not allow for the manipulation
of the independent variable
• Conditions for establishing cause-effect connections
– (1) statistical significance– (2) IV before DV.
– (3) Alternative causes
Ex-Post Facto Research
Designs
• Single Group
• Between Group
• Problems
– Self selection
– Correlation-causation fallacy
– Direction of causality
Download