Weeks 2 & 3 Flashcards
Independent Variable
“cause,” controlled, x-axis.
Depedent Variable
“effect,” measured, y-axis.
Random Sample
When everyone in a target population has equal chance of being in the study.
Convenience Sample
Selecting people for the study who are easy and ready to recruit.
Theoretical Definition
What we think the concept is, broad, but straight forward. Ex: defining happiness.
Operational Definition
Defining the variable in terms of operations we will perform to measure it or manipulate it. More specific. (Ex: Number of times somebody smiles in a day).
If two variables correlate, what are three possible cause and effect relationships between them?
IV causes DV
DV causes IV (reverse causation)
Another variable affecting IV and/or DV (third variable)
Three conditions to prove cause and effect
association, temporal precedence, rule out alternative explanations.
Association
The cause and effect must correlate.
Temporal Precedence
Cause must precede the effect (directionality).
Ruling out alternative explanations
There should be no other plausible causes of the effect. (Third variable problem).
Manipulation
Participants forced to undergo one procedure or another. Controls extraneous variables.
Random Assignment
Everyone participating has an equal chance of being in every condition - removes potential third variable as all groups are on average the same before IV.
Internal Validity
The extent to which the relation between the operationally defined IV and operational defined DV is casual. Only high in experiments.
External Validity
The extent to which the relationship discovered between the conceptual IV and conceptual DV truly exists in desired “real-world” group. Higher for survey/correlational research.
Construct Validity
Quality of the operational definition.
Statistical Validity
Extent to which the conclusions drawn from the data are supported by the quality of the data collected. Enhanced by larger sample sizes, larger effect sizes, and statistical procedure/control
Staged Manipulations
Use elaborate ruse to create psychological state (natural response).
Mundane Realism
The extent to which the study replicates conditions that occur in the real world.
Experimental/Psychological Realism
The study creates real psychological states in the participants. How much it engages the participants. Same mental process used in both lab and real-world.
Manipulation Checks
Try to directly measure the psychological state being created.
Pilot Testing
Run small portion of your study first; look at data, debrief participants.
Noise
Variability in resuslts.
Confounds
Something other than IV that differs between conditions and provides a possible alternative explanation.
Between-Subjects Designs/ Independent Groups Designs
When participants are only in one condition.
Within-Subjects Design/Repeated Measures Designs
When each participant is in each condition. Quicker, more statistically powerful.
Order Effect
The order of presenting treatments could affect your results.
Practice Effect
Performance improves.
Fatigue Effect
Performance worsens.
Counterbalancing
Randomizing the order of events for different participants.
Demand Characteristics
Expectations about the study changing the participants’ behaviour.
Experimenter Expectancy Effects
Person running study affects behaviour.
Double-Blind Technique
Both experimenter and participant are unaware of the hypothesis.
Block Randomization
All conditions randomly occur once before they are repeated for each participant.
Context Effects
Being tested in one condition affecting how participants perceive stimuli or interpret their task in later conditions.
Experiment
Study designed to specifically answer whether there is a casual relationship between two variables.
Includes manipulation, controlling extraneous variables.
Waitlist Control Condition
Participants are told they’ll receive treatment but have to wait until participants in treatment condition have received it.
Simultaneous Within-Subject Designs
Making multiple responses, presenting multiple conditions at once.
If ___ is possible always do that
Within-Subjects.
Control extraneous participant variables, limit noise, easier to detect relationship.
