Midterm 1 Flashcards
interquartile range: how to calculate
find median of upper quartile-median of lower quartile
best central tendency/measures of dispersion to use with outliers
median/interquartile range
how to calculate variance
square difference of each number, add squares, divide by N-1. Measured in units of standard deviation
standard deviation
square root of variance, measured in original units
type 1 error
rejected null hypothesis, but decision is wrong. Probability is alpha
type 2 error
failed to reject null hypothesis, but decision is wrong. Probability is beta.
Effect size
estimate of the magnitude of the difference among sets of scores, while taking into account the amount of variability in the scores. Useful for meta-analysis
Confidence interval
Based on the data for a sample, we can be 95% confident the calculated interval captures the population mean
Power
ability to find effect: probability that study will produce a statistically significant effect if the researcher’s hypothesis is true. 1-beta. Increasing sample size to increase power.
Reliability when evaluating measures
-consistency and -minimum of measurement errors. About reproduction
Validity when evaluating measures
-measuring what you want to measure. Must obtain evidence that supports hypothesis that dependent variable actually measures the construct it is supposed to measure
content validity
measures measure broadly what they are supposed to. For example: a job satisfaction survey questions a broad array of relevant beliefs and behaviors
-most basic level
criterion validity
ability to predict another aspect of construct. In example of job satisfaction, does your survey predict absenteeism? Also might mean your measure correlates with other measures/criteria of construct.
Helps establish construct validity
construct validity
Does the operational definition truly represent the abstract construct?
-Provide clear definition of abstract construct
-collect convergent and divergent data
Construct validity not established or destroyed in single study
nominal scale
two options, male or female, categories labeled as 0 or 1
ordinal scale
rank order. Order of finish in a poetry contest or race, intervals are meaningless, simply ranking
interval scale
order plus equal intervals. Zero on scale is not absence. Temperature is example. 100 degrees is 10 times warmer than 10 degrees
ratio
order plus equal intervals plus true zero. Weight is example.
When do you not need to get consent for research?
surveys, observation, educational research, archival, not placing subjects in unusual risk
Elements of consent for research
1.) basic description, 2.) length of time experiment will take, 3.) tell participants they can quit at any time, 4.) confidentiality and anonymity assured, 5.) contact info for researcher and IRB chair, 6.) opportunity to review results of study, 7.) signature
consent with children
must provide ASSENT
how to treat participants well in experiment
1) debriefing, -dehoaxing=explain deception so participant feels respected, -desensitizing=remove and decrease stress, bring person back to reality. -control leakage=make sure participants don’t talk to others
examples of poor consent
- Tuskeegee syphilis study
- Willowbrook hepatitis study
- mk-ULTRA-testing LSD on soldiers
mundane realism
how closely a study mirrors real-life experiences
experimental realism
the extent to which a research study has an impact on the subjects and forces them to take the matter seriously
strong theories include
- productivity: produce lots of research
- falsification-strong tests
- parsimony-simple explanation for phenomenon
convergence
when multiple experiments, all using different operational definitions and experimental procedures, converge on a common conclusion
can theories be proven?
Not conclusively, can be disproven and can be supported
Belmont report
includes three basic principles for research with human subjects: Respect for persons, Beneficence, and Justice
Critical incidents technique
used empirical data from incidents observed by 7500 psychologists to guide APA ethics code
Wundt
structuralism/introspection
William James
functionalism/stream of consciousness
Skinner/Watson
behaviorism/observation
ways of knowing
- authorities
- logic
- empiricism/experience
belief perseverance
stick to belief in face of contrary evidence
example: Covid-19 is not that serious and Trump never lies
confirmation bias
try to find facts to support opinion, ignore facts that don’t
example: most conspiracy theories, QANON
availability heuristic
use information read/heard recently
example: theory explains event because you just studied it
