14 - Assumptions

Question 1

What are the 5 main assumptions? (Assumptions Statistics)

Answer 1

• Linearity/additivity
• Independence of errors
• Homogeneity of variance
• Multicollinearity/auto-correlation
• Normality of residuals

Question 2

What is the extra assumption for ANCOVA? (Assumptions Statistics)

Answer 2

Homogeneity of regression slopes

Question 3

What is the extra assumption for repeated measures or mixed design? (Assumptions Statistics)

Answer 3

Sphericity

Question 4

What is the most important assumption? (Assumptions Statistics)

Answer 4

Additivity and linearity

Question 5

What is the least important assumption? (Assumptions Statistics)

Answer 5

Normality of residuals

Question 6

How are normality of residuals measured? (Assumptions Statistics)

Answer 6

PP/QQ plots and histograms

Question 7

What is homoscedasicity? (Assumptions Statistics)

Answer 7

Ensuring variance in residuals is constant over all levels of the predictor variable

Question 8

How to detect non-linearity/addativity and how do you fix it? (Assumptions Statistics)

Answer 8

• Visual plots

- Transformations

Question 9

How to detect independence of errors and how do you fix it? (Assumptions Statistics)

Answer 9

• Residual time series plot

- Use a different model or remove one predictor

Question 10

How to correct for violation of homogeneity of variance? (Assumptions Statistics)

Answer 10

Parametric equivalent test