lecture 25 (stats)

Question 1

mediation?

Answer 1

• relationship between predictor and outcome is explained by a third variable
• mediation analysis: dividing association between direct effects and indirect effects
• direct effect: from predictor to outcome when controlled for mediator
• indirect effect occurs: through the mediator
• check notebook

Question 2

mediation paths regression formulas?

Answer 2

1. Total effect:
   Outcome=b0+bt*Predictor
2. Mediator effect (a):
   Mediator=b0+ba*Predictor
3. Combined effect (b + c):
   Outcome=b0+bcPredictor+bbMediator

• bt = (ba*bb) + bc

Question 3

multicollinearity?

Answer 3

• you use mediation to deal/assess multicollinearity
• high multicollinearity inflates SE and weakens t statistics, messing with the predictive power of the model
• ba (mediator effect) specifically models ffor multicollinearity, since it checks the association of the predictor with the moderator

Question 4

calculate indirect effect?

Answer 4

• check notebook for unstardized and standardized coefficients
• Jasp already standardizes regression weights

Question 5

propotion of mediator?

Answer 5

• The proportion of the total effect that can be attributed to the mediator
• however it has its issues, since it is not a real propotion given the fact that it can exceed 1 and even take negative values
• check notebook for formula

Question 6

mediation in jasp?

Answer 6

• process (beta) -> classical process model -> specify variables -> models tab -> model 4
• estiamtes = regression coefficients (bs)
• if confidence interval has 0 it is not a significant effect

Question 7

mediation effect VS interaction effect?

Answer 7

just because there ia a very strong association between predicive variables it doesnt mean that there is an interaction effect