The correlation coefficient will always be negative for expensive houses

We are extrapolating far outside the range of the original data; the linear relationship may not hold for extreme values

The y-intercept is only valid for houses under $500,000

Multiple regression must be used for expensive properties