lesson 8 - multivariate correlational research

longitudinal studies

in longitudinal studies reasearchers measure the same variable in the sam epeople at two or more time points

• even if there are only 2 distinct variables its still multivariate because you end up with more than one measure variable—cause of the multuple time points

why use longitudinal designs?

• helps support causal conclusions

→ helps establish temporal precedence

cross-sectional correlation

correlation of the two different variables measured at the same time point

→ ex: time 1 TV violence & time 1 aggression

*establishes covariance

autocorrelations

correlation of each variable with itself across time points

→ ex: time 1 TV violence & time 2 TV violence

cross-lag correlations

correlation of the earlier measure of one variable with the later measure of the other

→ ex: time 1 TV violence & Time 2 aggression

*helps establish temporal precedence

p-value significance and CI

p < or = 0.05 → statistically significant

if 95% CI does not include 0 → statistically significant

beta (β)

the relationship between one predictor and the outcome while controlling for the other predictors

• positive = positive relation

• negative = negative relation

• larger value = strength

mediation

mediation answers the question: “how or why does x affect y”

→ explains WHY the two variables are related

3 pathwas in mediation

a path = IV → mediator

b path = mediator → DV

c path = total effect of IV → DV

c’ (c-prime) = direct effect of IV → DV after controlling for mediator

full mediation

after adding the mediator the direct path (c’) becomes non-significant

→ the entire effect of x on y works through M

partial mediation

after adding the mediator the direst path (c’) is smaller but still significant

→ part of the effect goes through M but part is direct

moderation

identifies certain situations for which the two variables are more strongly related

→ changes (modifies) the relationship