Task 6. Setting up research & sampling Flashcards
Population
- all possible participants that fit the criteria
Sample
- smaller subgroup chosen from the larger population
Generalization
- ability to apply findings of a sample to a larger population
- requires a representative sample
- one of the important factors for external validity
Random sample
- every person in the population has an equal chance of being chosen, necessary for generalization, but difficult to obtain
Nonrandom sample
- individuals chosen from a highly specialized subpopulation
Internet research sampling
- nonrandom, highly specialized, large but not representative (only ppl w )
Animal studies sampling
- strains need to be exactly the same, usually the exact same litter because any kind of genetic mutation can affect the results
Factors to consider when acquiring a research sample
1) research setting: lab or field
2) special needs of research: screening
3) ethical guidelines: APA
Volunteer bias
- individuals that volunteer fundamentally different from the ones that don’t, lowered external validity
Situational factors of volunteers
- interesting / important topis
- rewards
- personal acquaintance with the researcher
Internal validity and volunteerism
- volunteers are trying to please the experimenter, altering the results
- volunteerism can serve as “motivation indicator”
External validity and volunteerism
- WEIRD participant characteristics -> low generalizability
Remedies for volunteerism
- making the appeals as interesting as possible
- make appeal as nonthreatening as possible
- state in what way the participants are important-> benefitting others
- if possible avoid stressful tasks
- personalized appeals
Probability sampling
- each participants has a known probability of being in the sample -> ideally, all have an equal chance
Representative sample
- closely matches the characteristics of a population
Biased sample
- non-representative sample
Sampling techniques (5)
- simple random: random selection of a certain # of individuals
- stratified: population divided into segments, each segment represented
- proportionate: conserve the percentage of the population segments
- systematic: every nth individual is included, random start
- cluster: picking a naturally occurring cluster of individuals, including the entire group (used together with simple random)
What to keep in mind when determining Sample size
- smaller samples- controlled environment
- larger samples - many variables/uncontrolled variables
-amount of acceptable error
-expected magnitude of population proportions
Economic sample
- enough participants to be valid, but no more
Internal validity
- extraneous variables?
- confounding?
- ability of research design to adequately test your hypothesis / ability to adequately measure what its supposed to
-shown in variation in independent variable - threatened by extraneous variables that can provide alternative explanations and rival hypotheses
- confounding: merging of two or more variables to the point that it is impossible to separate their effects, affects overall correlation
Threats to internal validity
general sources of confounding (7)
- history: special events occur between obs.
- maturation: performance affected due to aging/fatigue
- repeated testing
- instrumentation: missed errors in criterion/instruments
- statistical regression: extreme-result subjects selected for treatment tend to regress toward the mean upon retesting
- biased selection: groups not equivalent prior to intervention
- experimental mortality: loss of subjects during the research
External validity
- results can be generalized for the population
Threats to external validity
- reactive testing: subjects react to pretest and adjust
- interaction between participant selection bias and the IV: Jožek iz Špičkovine vs Maartje van Maasi
- reactive effects of experimental arrangements: being aware that one is a participant and that they are being studied
- multiple treatment interference