Which regression approach enters predictors in a pre-specified order based on theoretical or prior research, with known predictors entered first?

Which regression approach enters predictors in a pre-specified order based on theoretical or prior research, with known predictors entered first?

• A. Stepwise regression
• B. Ordinary least squares
• C. Forced-entry regression
• D. Hierarchical regression

Answer: (D)

Hierarchical regression is the approach where predictors are entered in a pre-specified order based on theory or prior research. You start with known, theoretically important variables and then add others in subsequent steps to see whether they explain additional variance beyond what was already accounted for. The crucial idea is to test incremental validity: does each new block of predictors contribute meaningfully after controlling for those entered earlier? The change in R-squared from one step to the next (ΔR^2) shows whether the added variables provide a significant gain in explanatory power. This is different from methods that add predictors automatically based on statistics (stepwise regression) or those that put all predictors into the model at once without an ordered sequence (forced-entry regression). Ordinary least squares is the general estimation technique used to fit the model, not the entry strategy itself.

Hierarchical regression is the approach where predictors are entered in a pre-specified order based on theory or prior research. You start with known, theoretically important variables and then add others in subsequent steps to see whether they explain additional variance beyond what was already accounted for. The crucial idea is to test incremental validity: does each new block of predictors contribute meaningfully after controlling for those entered earlier? The change in R-squared from one step to the next (ΔR^2) shows whether the added variables provide a significant gain in explanatory power. This is different from methods that add predictors automatically based on statistics (stepwise regression) or those that put all predictors into the model at once without an ordered sequence (forced-entry regression). Ordinary least squares is the general estimation technique used to fit the model, not the entry strategy itself.