PALPSYU3349 Week 5

When looking at the interaction between 2 continuous variables through regression, what does the interaction coefficient (b) indicate?

• The Y-intercept

• The change in the slope of X1 on Y, per 1 unit of increase in X2 (and visa versa)

• The slope of X1 on Y, controlling for X2

• The slope of X1 on Y, when X2 is equal to 0

What is the null hypothesis for testing an interaction between two continuous variables (if B₃ = the interaction coefficient)?

• H_0: B₃ = 0 | B₁, B₂

• H₀: B₃ > B₁, B₂

• H₀: B₁ = 0

• H_0: B₃ ≠ 0 | B₁, B₂

What does the following output achieve?

generate weight100 = weight*100

• Summarises the variable "weight100"

• Renames "weight" as "weight100"

• Scales the variable "weight" (all points on weight have been multiplied by 100)

• Centres "weight" (mean weight now equals 0)

What does the following rvf (residuals vs fitted values) plot tell us?

• The assumption of linearity and homoscedasticity of residuals are met.

• Only the assumption of linearity is violated. Perhaps a curvilinear model would better capture the relationship.

• Only the assumption of homoscedasticity is violated.

• The assumption of linearity and homoscedasticity of residuals are violated. Perhaps a curvilinear model would perhaps better capture the relationship.

When there is no measurable improvement from one model to another (e.g. after adding predictors), what principle has been applied?

• The law of diminishing returns

• Polynomial regression

• The law of minimal change

• The assumption of homoscedasticity