case studies
one individual or group is studied to depth to reveal universal principles. time consuming, detailed info, individual cases can mislead.
meta-analysis
statistical procedure that combines conclusions of large # of different studies
naturalistic observation
observe and record behavior in naturally occuring setting. doesn’t explain behavior, but describes it. ex: Jane Goodall
surveys
asks people to self-report their behaviors or opinions, questioning a representative, random sample of group, can have low response rates. ex: political groups using election surveys
sampling bias
produces an unrepresentative sample when results are generalized
random sample
every person in the group has equal chance at participating
correlational studies
detect naturally occurring relationships.
how does one variable predict another?
cannot determine cause/effect
can help predict
ex: mental illness correlates with smoking
correlation
a measure of the relationship between 2 or more variables
correlation coefficient
1 # that represents 2 things: direction of the relationship & strength
positive correlation
increases in both variables
negative correlation
increase in one, decrease in other variable
correlation does not demonstrate causation
just b/c 2 variables are related doesn’t mean they cause other to occur
third variable
outside factors that can affect the study
confounding variable/third variable problem
when the third variable distorts the results
illusory correlation
perceiving a relationship that is stronger than it actually is
regression towards the mean
if one sample has extreme results, its likely the next will have avg results
double blind
subjects and experimenters don’t know if the subjects are in control
R
results have to be reliable and valid
theory
an explanation on how things work
experiment
specific, testable prediction based on theory
operational definition
how ind/dep variables will be measured in experiment