Psychology: Themes and Variations Ch 2
33 flashcards · Other · by Sophia Vona
3 Goals of research
Measurement + description understanding + prediction application + control
rationalism (R)
reason
empiricism (E)
observation
steps of an experiment
pre 1: get to know topic 1: form hypothesis 2: design study 3: collect data 4: analyze data 5: report findings all R, E except collect data (just E).
hypothesis
tentative statement about relationship between variables
operational def
definition in terms of actions used to measure smth
theory
system of interrelated ideas, explains an observation set
3 ways to test a hypothesis (empirically):
experiment quasi-experiment Non-experiment
quasi experiment
looks likean experiment, but IV not manipulated by researcher correlational
non-experiment
case study, naturalistic observation, survey correlational
naturalistic observation (Pros and cons)
no intervention +) observe phenomena as it occurs naturally +) good for brand new topics +) good for animal testing -) how to analyze qualitative data -) chaotic experiment = hard to observe -) observation can effect outcome (non-experient)
case study: case study research
singe subject, in depth analysis: amalgamation and analysis of multiple case studies +) in depth perspective +) compelling IRL evidence -) lack of objectivity increased subjectivity -) unusual case/ outlier non-experiment
survey/ questionnaires
interviews about specific aspects of behavior +) get data on hard to observe behavior +) lots of subjects -) sampling bias, low response rates -) self reported issues, wishful thinking, wording of q's -) can participants read/understand? non-experiment
experiment
+) precise control +) casual conclusions!! -) artificial: lack of ecological validity -) ethical/ practical concerns
Descriptive stats are about...
about study central tendency (mean median mode) dispersion of scores (standard deviation, variance) correlation ( r value 0 -> 1, + or = correlation)
inferential stats are about...
about greater population and the ability to make inferences (T test, Chi test, F-test, multiple regression) "universal truths" is it due to chance? p < .05, no, "statistically significant", all or nothing
effect size
magnitude of importance. does this research matter, is it enough of a change in DV to warrant changes
how to report findings
peer review journal publish or perish.
experiment
manipulate variable under controlled conditions determine causality
IV:
special treatment
"levels"
+ of ways IV is experienced
DV:
may be caused by IV
experimental group:
subjects receiving special treatment
control group
comparison groups
random assignment
no random assignment = confounding variables = inconclusive results. decreases odds of extra IV's
confoduing variable
influences results, inconclusive
extraneous
even across groups, affects DV but in equitable ways across groups.
sampling bias:
size must be large, don't rely on anecdotes
placebo
participant expectations change results
experimenter bias
experimentors expectations change outcome
replication
science is replicable, metaanalysis is important
self reported data potential errors
distortion, memory, social desirability bias...