Introduction & Research Methods Flashcards
Goal in Science (theory-data cycle)
- coming up with theory and test it
-> disproving a theory
1. observation/existing theory
2. hypothesis (expected outcome of study)
3. generating predictions; data evaluation (new theories, new hypothesis, …)
-> we cannot prove theories (just say if they are correct) -> never find full truth
hypothesis testing trade-off
internal validity: experimental control (causal statement)
vs.
external validity: generalizability (allowing for general statement -> applying to other groups/all)
Observational studies (correlation research)
-> data collection:
no one knowing -> natural behavior
obvious -> maybe less natural
- naturalistic observation; watching people in the wild (unstructured)
-> generalizable, no impact caused, no internal validity, might have biased observations (influencing) - participant observation: sociologist incorporates in group and watches behavior if members
-> in between controlled and field - structured observation: lab studies (planned happenings)
-> high internal validity, maybe not very generalizable
-> replication is key to science
Festingers study (when prophecy fails)
- attitude changes (alien cult idea)
-> solved discomfort of misbelief by getting to the belief that they saved the world
-> external validity; human decisions can be shaped towards being positive
correlational study types, beyond observational studies
- surveys; quick data collection, interview studies (take long? not always truthful)
- analyzing pre-existing data; associations, not causation! (there might be an in between variable causing that effect)
- directionality if correlation
experiments
= to get rid of all sorts of other explanations (maximize internal validity) -> not real world representative
dependent variable
measured; value dependent on independent variable
independent variable
manipulated; different combinations (all needed to make causal statement)
control variable
same experience for all participant (just independent variable varies)
confounding variables
any variable systematically varying with the independent variable that is out of what you wanted to study -> important to ensure that they are not there
random - sampling vs assignment
getting random sample around the world vs random assignment to independent variable manipulation (group; chance that kind of people in groups are same)
experimental control (2 components)
- cover story; story told to participants (deception = lying within that -> avoid confounders)
- confederate; routing the experiment, ensures same experience for all
brainstorming study - disruptions vs no disruptions
- lab experiment, study alone, with group, alone with interruptions (IV)
- DV; quality of brainstorming ideas
-> people alone showed better brainstorming skills than those with interruptions of both kinds -> no significant difference between type of interruption -> causal statement allowed (fact)
Quasi vs field experiments
quasi = no random assignment possible
-> interested in subject variable (reduces causal claim option)
field = no variable control
-> messy, uncertainty in data
-> real-world -> generalizable (no causal claims though)
peeing in public (field experiment)
-> quasi-field experiment when demographically data are collected in the debrief
- invasion of private space (higher arousal; not thinking about peeing anymore)?