week 11 Flashcards
(8 cards)
Moderation
the relationship between a predictor variable and an outcome variable is dependent on the level of moderating variable
Mediation
Mediation refers to a situation when the relationship between a predictor variable and an outcome variable can be explained by their relationship to a third variable
Baron & Kenny (1986) Method
A linear model predicting the outcome from the predictor (total: c)
A linear model predicting the mediator from the predictor (1st indirect: a)
A linear model predicting the outcome from both the predictor (direct: c’) and the mediator (2nd indirect: b)
the predictor must predict the outcome less strongly in model 3 than in model 1.
Sign of the paths in mediation
In a mediation pattern, the indirect path (a*b) is the same sign (+/-) as the direct path, and they add together for a larger total path.
types of mediation
Full Mediation
Direct effect of the predictor on the outcome becomes nonsignificant when the mediator is entered into the analysis.
Partial Mediation
Direct effect of the predictor on the outcome is reduced but still significant when the mediator is entered into the analysis.
Mediation cautions
Mediation is a causal model; if the causal order of concepts being measured is not established, it proves nothing.
Regression or correlations will be a better way to look at relationships among variables.
reasons to clam causal precedence
Logical (gender, nationality, etc.)
Manipulation (with random assignment)
General can take precedence over specific (general sexism vs. dating beliefs)
Time course of occurrences being studied – not necessarily of measurements in the same session (so, beliefs at time 1 can cause beliefs 2 weeks later, not the other way around
Bootstrapping
Bootstrapping is a statistical procedure that resamples a single dataset to create many simulated samples.
It is useful when your data are highly skewed with low N
The higher your N, the more it’s similar to normal statistics (remember the central limit theorem)
