Summary

Question 1

The correlation is the … covariance

Answer 1

Standardized

Question 2

T/F (explain): the correlation of z-scores = covariance of z-scores

Answer 2

True, because the z-scores are already standardized

Question 3

What are the values of the sd and mean if we standardize by using z-scores

Answer 3

Sd = 1, mean = 0

Question 4

What are the values b0 and b1

Answer 4

B1= slope (regression line)
B0=intercept

Question 5

What are y and ÿ (hat)

Answer 5

Y = observed
Ÿ = predicted

Question 6

How is the fit of the model shown

Answer 6

By the correlation between observed and predicted values

Question 7

F = …

Answer 7

Signal/noise or explained variance/unexplained variance

Question 8

Multicollinearity can be a threat to estimation of regression coefficients in a regression analysis because (3):

Answer 8

1. It causes the value of the explained variance of the model to decrease
2. It makes it difficult to determine the individual importance of the predictors
3. It causes the standard error of the b coefficients to increase, making the estimates of the b coefficients less trustworthy

Question 9

How can we model the individual effect of the predictor variables when multicollinearity is violated

Answer 9

With mediation models

Question 10

What do a, b, c, and c’ stand for

Answer 10

A = effect of predictor on mediator b = effect of mediator on outcome —> together these are called the indirect effect c = total effect c’ = direct effect

Question 11

How do we calculate the indirect effect

Answer 11

Total effect - direct effect (c-c’)
or
predictor on mediator * mediator on outcome (a*b)

Question 12

Can we infer causation from the outcome from mediation or moderation analyses

Answer 12

No, all these measures are correlational

Question 13

What are the 6 most important assumptions and how do we check for them

Answer 13

1. Linearity (linear relation between predictor and outcome variable) - scatterplots
2. Homoscadesticity (variance of residuals is equal across all expected values) - predicted values X residuals plot
3. Sensitivity (outliers) - cook’s distance
4. Multicollinearity (predictor variables should not be too highly correlated) - collinearity diagnostics: VIF < 10, tolerance > 0.2
5. Normality - q-q plots
6. Normally distributed residuals - plots, residuals histogram

Question 14

What is the difference between a mediation and a moderation analysis

Answer 14

Mediation involves an indirect effect of the independent variable on the dependent variable, whereas moderation involves an interaction effect between the moderator and the independent variable