Lecture 02 Methods in Cognitive Psychology Flashcards
Scientific Principles
- Rationalism
- Empiricism
- Self-regulatory
- Falsifiability
Rationalism
Logical and systematic theories
Empiricism
Predictions and hypotheses are tested against empirical observations
Self-regulatory
- Theories and data are scrutinized by scientific communities
- Peer-reviewed journal, open data, pre-registration.
Falsifiability
A theory is scientific only if it can be, in principle, proven false
Research Designs
- Descriptive Research
- Relational research
- Experimental Research
Descriptive Research
To describe the phenomenon but not to explain the cause
- Observation
1) Controlled
2) Naturalistic
3) Participant
- Case study
1) Rare events
2) Few cases being studied in details
Strength and Weakness of Descriptive Research
Strength
- A starting point for later research
Weakness
- Can’t explain or imply causation
Relational Research
Association between two or more variables
- Linear relationship
- Nonlinear relationship
Limitation of Relational Research
Only explaining linear regression part of exponential
- Constraining or controlling for other variables
Criteria for Causal Inferences
- Time precedence
- Covariation
- No alternative explanations
Covariation
The amount or the absence of the cause affect the amount or the absence of the outcome
Causal Modeling with SEM
Constraining or controlling for other variables
Experimental Research
- True Experiment
- Quasi-experiment
True Experiment
- Manipulation of independent variable
- Control/comparison groups
- Random assignment
Purpose of Random Assignment
Equalizer
- Because some variables can’t be controlled, like pre-detemined characteristics.
Limitation of True Experiment
Unable to manipulate some variables, such as race
- It will be more of naturalistic experiment
Quasi-experiment
- Experimental design without random assignment
- Practical or ethical reasons
- Usually in field research
Limitation of Quasi-experiment
- Threats to internal validity
1) Self-selection
2) Pre-existing differences - It does provide strong claim as pure experimental research
- In naturalistic setting, something can’t be manipulated due to ethics
Between-subject Design
- Prevent carry-over effect
- “Blind” to conditions
- More variation due to individual differences
Within-subject Design
- Subjects are their own control
- Higher statistical power -> less N
- Prone to carry-over effect
- No confounding of individual difference
- Preferred by cognitive experiments
Validity
- Internal validity
- External validity
- Construct validity
- Statistical conclusion validity
Internal Validity
Free of confounds
- Running multiple experiment to examine alternative explanation
External Validity
Generalizability
- Concern in the field study or applied study
- Ecological validity: Applying knowledge or findings in real world
- Cross cultural research