1/70
Looks like no tags are added yet.
association claims
Cannot assign people to groups
outliers
Can have a lot of influence on association claims
extreme scores on both variables
outliers are the most problematic for bivariate correlation when there are
covariance, temporal precedence, internal validity
causal claims must have
only a correlation because of a third variable
supurious association
how well was each variable measured
construct validity question
effect size
strength of the relationship between two or more variables
strength, precision, replication, outliers, ROR, curvilinear
6 statistical validity questions
if correlation is still present in subgroups
how to determine if a correlation is not due to a third variable
We are 95% confident that the true correlation falls between
what does a 95% confidence interval mean
temporal precedence
longitudinal studies help establish
internal validity
multiple regression analyses help address
cross-sectional correlation
Ex.: Is there a correlation between mothers’ overvaluation at Time 4 and children’s narcissism at Time 4?
autocorrelation
Ex.: Is there a correlation of mothers’ overvaluation at Time 1 with mothers’ overvaluation at Times 2, 3, and 4?
cross-lag correlation
Ex.: Is there a correlation between mothers’ overvaluation at Time 1 and a kid’s narcissism at Time 2?
internal validity
longitudinal designs do not establish
multivariate designs
allow researchers to control for another variable
pattern and parsimony
Researchers can investigate causality by using a variety of correlational studies that all point in a single, causal direction
parsimony
the degree that a scientific theory provides the simplest explanation of a phenomenon
causal
mediation hypotheses are
case study, stable baseline, multiple baseline, reversal
4 types of small-n designs
internal
quasi-experiments have more potential threats to ___ validity
quasi-experiments
often examines real-world situations
quality of design and pattern of results
Whether a quasi-experiment supports a causal claim depends on two things:
nonequivalent control group interrupted-time series
Comparing opioid overdose death trends in Florida and North Carolina around Florida's new laws targeting pill mills.
selection effects
People choosing cosmetic surgery may differ from those declining it
design confound
Another variable differs systematically between conditions
design confound
Opt-out countries might also have organ donation advertising campaigns
attrition
People dissatisfied with surgery are more likely to stop participating
instrumentation
Observers become more lenient when rating behavior
placebo effects
Someone reports feeling better because they believe a treatment works
real-world, external validity, ethics
why quasi-experiments
quasi-experiments
more deliberately select groups, comparison conditions, or time periods.
quasi-experiments
often organized around a treatment, intervention, or event.
small-n
researchers repeatedly measure individual participants' behavior.
small-n
Examines individual responses to conditions
large-n
Estimates group-level effects
stable-baseline
A researcher observes a child's behavior for three weeks before introducing reinforcement, then continues measuring behavior. What design is this?
selection effects
A wait-list design is helpful in dealing with which of the following threats to internal validity?
baseline
in ABAB design, A represents the
treatment
in ABAB design, B represents the
controls for shared external event
Benefit of multiple baseline design
many participants, group averages
large-n designs differences
few participants, compare individual patterns
small-n designs differences
carefully control study
how to control for internal validity in small-n designs
convenience and volunteer
2 ways a sample can be biased
probability sampling
every member of the population has the same chance of being selected
simple, systematic, cluster, stratified, oversampling
5 techniques for random sampling
convenience, purposive, snowball, quota
non probability sampling
quota sampling
Ex. surveying students until you have 50 freshmen, 50 sophomores, etc
frequency claims
external validity is crucial for _____
sampling method
external validity comes from the _______
reactivity
When participants change their behavior because they know they are being observed.
acquiescence
A participant selects "agree" for every statement without reading carefully.
response set
A student selects 3 on every question of a 1–5 scale without considering the questions.
fence sitting
A participant selects "neither agree nor disagree" on controversial questions to avoid expressing an opinion.
socially desirable responding
A student reports exercising five times a week because they want to appear healthy, even though they rarely exercise.
faking bad
A participant exaggerates their depressive symptoms to appear more distressed than they are.
reactivity
Students behave more politely when they notice a researcher observing their classroom.
codebook
Researchers create specific written definitions of aggression and cooperation to help observers consistently record behavior.
open ended, forced choice, likert scale
3 ways a question can be worded
habituation
Allow participants time to get used to being observed so they eventually behave more naturally
unobtrusive observation
Observe behavior without drawing attention to the fact that it is being measured.
habituation or unobtrusive observation
how to control for reactivity
blind
how to control for demand characteristics
percentage of variance explained by predictors
what does r squared mean
statistically significant
p<.05
not statistically significant
p>.05
r squared
Represents the proportion of variance in the dependent variable explained by the regression model.
establishing stable baseline
how can small-n designs support internal validity
treatment, rather than outside variable, causes the change
how does ABAB pattern strengthen an argument of a treatment?