II - Multivariate Correlational Research Flashcards
Longitudinal design
A study in which the same
variables are measured in the same people at
different points in time.
E.g., Eron and colleagues’ (1972): Measured
children children s aggression and interest in violent ’s aggression and interest in violent
television programs in the same group of
people across time-10 years apart
Cross-sectional correlations:
: In a longitudinal
design, a correlation between two variables
that are measured at the same time. (like those taken at the beginning at of the study, then at the end)
Autocorrelation:
The correlation of one variable with itself, measured at two different times.
Cross-lag correlations:
i.e. TvViolence2001 w/ Aggression2011
In a longitudinal design,
a correlation between an earlier measure of
one variable and a later measure of another
variable.
Addresses the directionality problem and help
establish temporal precedence
Longitudinal Design and 3 Prereqs for Causation
Covariance: Significant relationships in
longitudinal designs help establish covariance.
Temporal precedence: A longitudinal design can
help researchers make inferences about temporal
precedence.
Wh en one cross-lag correlation is stronger, this helps to
determine which variable comes first.
Internal validity: When conducted simply
longitudinal studies do not help us rule out third
variables.
Longitudinal Designs: How structure so that can find causation
Careful researchers may g be able to design their
studies or conduct the subsequent analyses in
ways that address some third variables.
E.g., Eron and his colleagues studied boys and
girls separately; they found that a preference
for violent TV predicts aggression for boys but
not for girls.
Rules out gender as a third variable; gender
moderates the effects of TV preference on violence.
Why not just do an experiment (instead of a longitudinal study)?
.Exp best way to confirm/disconfirm causal claims
.But, many cases ppl can’t be randomly assigned or it would be unethical
.So longitudinal design becomes a good option
Multiple regression (or multivariate regression):
A statistical technique used to test for the
influence of third variables.
-rule out third variables and address questions of internal validity
-ex: more recess and less behavior problems
-possible third variables: class size, private schools, income, etc
Criterion Variable
The variable in a multiple
regression that the researchers are most
interested in understanding or predicting.
May also be called the dependent variable.
Predictor Variable
Predictor variable: A variable in an analysis
using multiple regression that is used to
explain variance in the dependent or criterion
vari bl ae.
May also be called the independent variable.
Beta?
positive B indicates positive relationship btw specific predictor variable and DV, when other variables are statistically controlled for
neg B indicates neg relationship two variables when other predictors are statistically controlled for
-betas change depending on what other predictor variables are being used and controlled for in the regression
-can compare two Bs within same table to each other
-little b shows similar associations but cannot be compared with same table to each other
Mediating variable (or mediator):
A->B(med var)->C
A variable that
helps explain the relationship between two
other variables.
Can be investigated with experiments, but
mediation analyses often rely on multivariate
tools such as regression analyses.
mediating step between the IV and DV that causes the DV
-overall relationship “c” only due to relationship “a” btw IV and mediator and relationship “b” btw mediator and DV
-ex: availability of recess and behavior problems, mediator = physical activity
Statistical Significane of Beta
column labeled “sig” or “p”
- tell whether each beta is statistically significantly different from zero
- when B is zero, shows that the relationship is explained by another variable
mediator vs moderator
med: “why are these two variables linked?”
mod: “are these two variables linked the same way for everyone, or in every situation?”
Third-Variable Problem
internal validity rule, when you can come up with an alternative explanation for the association between two variables, that alternative explanation is the third variable
