2) Research Methods Flashcards
What are the two modes of thinking?
Intuitive & Analytic
Intuitive thinking
System 1
- Quick & reflexive
- Gut hunches
- Not much mental effort since brain is on autopilot
- Not always right
- Use of heuristics
ex/ form first impression on smb
Heuristics
Mental shortcuts or rules of thumb that help us make sense of world
- Can lead to errors
Analytical thinking
System 2
- Slow & reflective
- Takes mental effort, trying to reason through a problem
- Can override intuitive thinking and reject gut hunches
ex/ first learning to drive car
What is an operation definition and what kinds of variables are included?
A concrete way to measure an abstract concept
- Includes situational, response, & participant variables
Describe whether participant, response, and situational variables can be manipulated and/or measured
Situational: can measure & manipulate
Response: measure only
Participant: measure only, cannot be manipulated
What is the key to generalizability?
Random selection
What are the three key characteristics of a scientific method?
- Random Selection
- Evaluating measures (reliability & validity)
- Openness & transparency
What 2 things make a good measure?
validity & reliability
Describe the difference between validity and reliability
Validity = accuracy, whether it’s measuring what it’s supposed to
Reliability = consistency, stability of the measure
What is external validity?
Generalizability
- degree to which a study’s findings can be generalized to population from which sample was drawn
Ecological Validity
degree to which the setting of a study mirrors real life settings
- subtype of external validity
Internal Validity
the extent to which a causal conclusion can be made
- usually high in well designed experiments
What is our first concern with any measure, reliability or validity?
reliability
Population vs Sample
Sample is drawn from population, then generalize back to population
- bigger sample = more likely to be representative
What is the differences between random selection, sampling, and assignment?
Selection
- how sample is obtained from population
- impacts generalizability
- random = has equal chance of being selected to be in experiment
(Sampling is same as selection^)
Assignment
- in experiments ONLY
- how sample is split into conditions
- impacts internal validity
- random = has equal chance to be in conditions
What is sampling bias and what does it impact?
- how a sample is obtained might be biased
- impacts generalizability
What are WEIRD samples?
most research studies are conducted on:
Western
Educated
Industrialized
Rich
Democratic
example: arrow longer vs shorter
What is a non experimental design?
design where experimenter does not manipulate or control any of the variables
What are the types of non experimental designs?
case study
observation
correlation
Naturalistic Observation
- watch behaviour in real world WITHOUT manipulation
- Some behaviours are better understood irl cuz can’t replicate envo in lab
- does not involve sampling!
- non experimental
Naturalistic observation pros & cons
pros
- high in external validity
cons
- low in internal validity
- cannot make causal claims
Case Study
- examine one person or small number of ppl in depth (sometimes over long time)
- can study rare conditions that cannot be recreated in lab
- cannot draw sample!!
- non experimental
Case study pros & cons
pros
- can provide existence proofs
- allow us to study rare or unusual phenomena
- can offer insights for a way to generate hypotheses / later testings
cons
- can be misleading cuz anecdotal
- cannot generalize
- cannot make causal claims
What are existence proofs?
demonstrations that a given psychological phenomenon can occur
- usually from case studies
Self Report Measures
Assess characteristics by asking directly, aka questionnaires
- Closely related is surveys (used to measure opinions and attitudes)
Self report measures pros & cons
pros
- Easy to administer
- measures of personality traits and behaviours work well
cons
- People may not have enough insight to report accurately
- Can obtain different answers depending on how questions are phrased
- People may not understand the answers they give
- Dishonesty: response sets or malingering
What is a response set?
tendency of research participants to distort responses to make them look better
What can counter some problems with self report data and why?
rating data
- ppl don’t have the same blind spots when rating someone else rather than themselves
(more objective)
What is the drawback of rating data?
Halo effect
Horns Effect
- Opposite of halo
- Tendency of ratings of one negative trait to spill over and influence to ratings of other negative traits
Halo Effect
- tendency of ratings of one positive characteristic to influence ratings of other ones
Correlational Designs
- Examines extent to which two or more variables are associated
- Measure variables without manipulating them
- non experimental
Correlational design pros & cons
pros
- can help us predict behaviour
cons
- cannot imply causation!
describe the correlation constant (r)
- statistic that index the strength of relationship
- strength is absolute value
- only useful for LINEAR relationships
-1 (perfect negative)
0 (no relationship OR it’s not linear)
1 (perfect positive)
How is correlational data collected & analyzed?
on a scatterplot
what does it mean when r is 0 in correlational data?
no relationship
OR
there is a relationship but it is not linear
Illusory Correlation
- perception of statistical association between variables when none exists
- form basis of many superstitions
Why are we prone to illusory correlation?
Our minds are
- not good at remembering nonevents
- pay more attention to times when things happen
To minimize tendency, we should pay more attention to disconfirming instances
Experimental Design
- characterized by random assignment and manipulation of an independent variable
- allows cause and effect inferences
Experimental vs correlational design
- differences are created (exp) vs measured (corr)
- can draw causal inferences (exp) vs cannot
(i.e. high internal validity vs low)
Describe internal validity in an experimental design
Ability to infer the IV causes changes in the DV
Must have:
1. Covariation between the 2 variables
2. Temporal precedence (A must come before B)
3. Eliminate plausible alternative explanations
Can the difference in the DV be attributed ONLY to the manipulation of the IV?
How can we achieve internal validity in experiments?
- Experimental Control
Manipulate IV with no confounds - Random assignment
Balances influence of variables across conditions
Confounds
- intertwined with IV and impacts DV
- could explain part of the result
- source of false conclusions
Random assignment is more effective (for achieving internal validity) as the sample size ____
increase
Between subjects design vs within subject design
btwn
- assign diff groups to control & exp conditions
within
- each participant acts as their own control
What are some pitfalls in experimental designs?
Placebo effect
Nocebo effect
Experimenter Expectancy Effect
Demand Characteristics
Experimental design pros & cons
pros
- high internal validity, can infer causation
cons
- can sometimes be low in external validity
Placebo Effect
Improvement resulting from the expectation of improvement
ex/ we think more expensive meds work better
How can we avoid the placebo effect in experimental designs?
Use a placebo: participants shouldn’t know whether they have the real or placebo treatment
True or false: Placebo effect is all in people’s heads
false, placebo effects are just as real as drugs
Nocebo effect
evil twin of placebo effect
- harm resulting from expectation of harm
ex/ ppl allergic to roses sneeze when presented with fake ones
Experimenter Expectancy Effect
Form of involuntary and unconscious cueing by the researcher that impacts participant’s behaviour
ex/ horse that can do math
How can we avoid the experimenter expectancy effect in experimental designs?
double blind
- neither researchers nor participants are aware of who’s in which conditions
What are demand characteristics? What aspect of experimental design is it a threat to?
any feature of an experiment that might inform participants of the purpose of the study
- a threat to internal validity
How can we minimize the effects of demand characteristics in experimental designs?
use deception
cuz participants might try to help OR hurt,
can be due to social desirability
Describe the difference between quasi-experimental vs experimental designs
- Also use a control group but NO random assignment
- selected based on pre-existing values of IV rather than assigned values
- should still be some kind of intervention,
ex/ can’t ask eye colour cuz ppl born with it - Less internal validity, cannot make causal statements
- allows its wide use in ethical issues
common example: study brain before and after getting a concussion