1/84
Looks like no tags are added yet.
Induction of hypotheses from curiosity
Empirical Observation
drives scientific inquiry from observation to theory.
Empirical Cycle
Deduction of specific predictions
Theory Building
Experimental testing of predictions
Empirical Testing
Evaluation of theory
Theory Testing
A recording of observations, quantitative or qualitative, made by recorded operations and in defined conditions followed by appropriate statistical and mathematical rules, for the existence of significant relations.
Experimentation Definition
Systematic documentation
Recorded Operations
Eliminating outside noise
Defined Conditions
Statistical significance
Mathematical Rules
The idea that repeated exposure causes people to habituate to negative thoughts
Broad Theory (Habituation Theory)
Does expressive writing cause this habituation to happen faster?
Unobserved Component
Expressive writing about positive experiences will not reduce emotional distress as effectively as writing about traumatic experiences, because it lacks the necessary exposure to negative thoughts.
Specific Hypothesis
An existing framework or model used to interpret phenomena.
Step 1: Theory (Hypothetico-Deductive Cycle)
A strict, deductive prediction derived from the theory about a new phenomenon.
Step 2: Hypothesis (Hypothetico-Deductive Cycle)
Conducting the experiment to gather targeted data.
Step 3: Empirical Test (Hypothetico-Deductive Cycle)
Reevaluating and adjusting the original theory based on the new evidence, triggering the cycle to begin again.
Step 4: Theory Revision (Hypothetico-Deductive Cycle)
Assumes there is no real relationship in the population. Any observed effect in the sample is purely due to random sampling error.
The Null Hypothesis (H0)
The World of Chance
The Null Hypothesis
Assumes the manipulated independent variable actually caused a real change in the population.
The Alternative Hypothesis (H1)
The World of Causality
The Alternative Hypothesis
Goal of Hypothesis Testing
To determine if the data is surprising enough to reject the Null reality in favor of the Alternative.
The Burden of Proof
Achieving Statistical Significance : Statistical Significance (Reject H0)
How massive is the difference between groups? (Even small samples can detect massive effects).
Relationship Strength (Effect Size)
How many participants were tested? (Massive samples can detect incredibly weak, microscopic effects).
Sample Size (N)
Statistical significance measures probability, not importance. A statistically significant finding simply means
the result was unlikely to occur by chance alone.
You recognized there was no effect.
Correct Decision (H0 True & Retain H0)
Claiming a causal relationship exists when it is actually just random chance.
Type I Error (False Positive)
Failing to detect a real causal relationship, often due to low statistical power or a small sample size.
Type II Error (False Negative)
You successfully discovered a real effect.
Correct Decision (H0 False & Reject H0)
Successfully discovering a real effect when H0 is false and rejected.
Adequate Power
The Cause / Predictor Variable. The input that is manipulated or allowed to vary.
Independent Variable (IV)
The Effect / Criterion Variable. The output that should only vary in response to the IV.
Dependent Variable (DV)
Background noise. Uncontrollable influences that must be controlled.
Extraneous Variables
Critical threats. Extraneous variables that have accidentally co-varied with the IV.
Confounding Variables
Participants report their own thoughts, feelings, and actions (e.g., Rosenberg Self-Esteem Scale).
Self-Report Measures
Direct observation and recording of participant actions (e.g., number of physical strikes on a Bobo doll, or backward digit span task performance).
Behavioral Measures
Recording internal biological processes (e.g., heart rate, blood pressure, galvanic skin response).
Physiological Measures
Different participants in different conditions.
Unrelated Design
Unrelated Design pARTICPANT aLLOCATION
Between-subjects, Between-participants, Independent groups, Unrelated groups
Unrelated Design Rule
Use unrelated statistical tests
The same group of participants across different conditions
Related Design
Related Design Alternate Terms
Within-subjects, Within-participants, Repeated measures, Dependent groups
Related Design Rule
Use related statistical tests.
wHAT DESIGN THAT Each participant experiences only a single condition.
Basic Between-Subjects Design
The initial participant pool gathered for the study.
The Pool
assignment separating the participant pool into condition tracks.
Random Assignment
This separation physically isolates the groups, preventing cross-contamination of the treatment's effects.
Between-Subjects Separation
The intentional creation and variation of the independent variable.
Experimental Manipulation
The strict control of all environmental and procedural variables other than the independent variable.
Standardization of Procedures
The unbiased allocation of participants to conditions or to the order of conditions.
Random Assignment
The manipulation is the specific alteration believed to cause an effect on the dependent variable.
Rule 1: Active Creation
A manipulation must have at least two different levels (e.g., Treatment vs. No Treatment).
Rule 2: Minimum Threshold
Care must be taken to manipulate only one thing at a time to avoid confounded results.
Rule 3: Isolation
Always check the effectiveness of the manipulation during participant debriefing.
Rule 4: Verification
Rules of Experimental Manipulation
Rule 1: Active Creation, Rule 2: Minimum Threshold, Rule 3: Isolation, Rule 4: Verification
Random Assignment Problem
Participants bring their own extraneous variables (traits, backgrounds) into the lab.
Random Assignment Solution for Extraneous Variable
By randomly assigning participants to conditions, individual differences are distributed evenly. On average, the two groups become fundamentally equivalent before the manipulation even begins, preventing demographic or personal traits from becoming confounding variables.
Random assignment establishes ______ while random sampling establishes __________
causality; generalizability
Causal inference, generalized to the whole population. (The Ideal). Establishes Causation.
Random Assignment + Random Sampling
Causal inference, but only valid for this specific sample. Establishes Causation.
Random Assignment + NO Random Sampling
No causal inference (correlation only), but the correlation generalizes to the whole population. Establishes Correlation.
NO Random Assignment + Random Sampling
Correlation only, restricted only to this sample. (The Weakest). Establishes Correlation.
NO Random Assignment + NO Random Sampling
Standardization minimizes?
unintended variation in how participants are treated
Specify every experimenter action from greeting to dismissal.
Written Protocol (Standardization)
Use word-for-word scripts or participant-read documents.
Standard Instructions (Standardization)
Ensure each experimenter tests participants across all conditions.
Experimenter Training (Standardization)
Utilize software packages or slide shows to deliver the procedure consistently.
Automation (Standardization)
Develop standard answers to prevent improvised responses.
Anticipate Questions (Standardization)
Standardization Risk
Systematic but unintended differences in treatment lead directly to erroneous conclusions.
Lever of controlling extraneous variable by Completely removing extraneous physical variables from the situation. (Example-Soundproofing a room to eliminate outside noise).
Lever 1: Elimination
Lever of controlling extraneous variable by Keeping all aspects of the treatment conditions identical, except the IV. (Example-Testing all subjects in the exact same room, or at the exact same time of day).
Lever 2: Constancy of Conditions
Lever of controlling extraneous variable by Equally distributing the effects of a variable across all conditions. (Example-Running half the subjects for condition A in the morning, and the other half in the evening).
Lever 3: Balancing
Levers of controlling extraneous variable by
Lever 1: Elimination, Lever 2: Constancy of Conditions, Lever 3: Balancing
Anatomy of a True Experiment
Hypothesis (Input), conditions (Engine), Blocking noise (Shield), internal validity (Equalizer), data ready for statistical analysis (Output)
Gather the initial participant pool.
Step 1-Sample Selection
Split the pool into Group A and Group B to ensure baseline equivalence.
Step 2-Random Assignment
Group A receives the active treatment condition; Group B receives the neutral/control condition.
Step 3-Manipulation (IV)
Uniformly operationalize and measure the output for both groups.
Step 4-Measurement (DV)
Utilize inferential statistics to test the null hypothesis and check for a significant difference between the groups' averages.
Step 5-Comparison
Anatomy of Basic Design
Sample Selection, Random Assignment, Manipulation, Measurement, Comparison
Between-Subjects Advantage 1
Conceptually simple to design and analyze
Between-Subjects Advantage 2
Requires less total testing time per participant.
Between-Subjects Advantage 3
Total elimination of carryover, practice, or fatigue effects (participants aren't exhausted or biased by taking multiple treatments).
Between-Subjects Disadvantage 1
Inefficient; requires significantly larger overall sample sizes to achieve statistical power.
Between-Subjects Disadvantage 2
Vulnerable to "noise": Because different people are in each group, participant variability makes it harder to detect the true effect of the independent variable.