Which statistic is defined as the difference between the adjusted predicted value and the original predicted value for a case?

• A. a measure of the influence of a case on the regression parameter bi when deleting the case
• B. a covariance structure used in multilevel models with heterogeneous variances and zero covariances
• C. DFFit
• D. a variable that can only take on certain values

Answer: (C)

In regression diagnostics, DFFIT captures how much an observation changes the model’s predicted value for that same observation when that observation is removed from the data. It is the difference between the leave-one-out predicted value (the prediction you’d get if you delete that case) and the original predicted value from the full dataset. This directly shows how influential a single case is on its own predicted outcome.

This is why it’s the best choice here: it specifically measures the change in predicted value due to deleting the case. The other statements refer to different ideas—one describes influence on a specific coefficient, another a covariance structure used in multilevel models, and the last describes a type of variable—not the diagnostic that quantifies how predicted values shift when a case is removed.