Answer: It is one that is unnecessarily included in the model, whose actual coefficient and consistently approximated value is 0 in large sample sizes. If we add these variables is costly.

## Explanation An extraneous variable, in the context of regression diagnostics, is a variable that is unnecessarily included in the model. Its actual coefficient and consistently approximated value is 0 in large sample sizes. This means that the variable does not contribute significantly to the model's predictive power or accuracy. ### Key Points: - **Definition**: A variable whose true coefficient is zero in the population - **Impact**: Does not improve model prediction in large samples - **Cost**: Including such variables is computationally expensive and can lead to overfitting - **Identification**: Should be identified and removed to improve model efficiency ### Why Other Options Are Incorrect: - **Choice A**: Incorrect because extraneous variables are not eliminated to "increase effectiveness" - they are removed because they have no real effect on the dependent variable - **Choice C**: Incorrect because sample size appropriateness is not the criterion for extraneous variables - **Choice D**: Incorrect because variables needed to control confounding are important and necessary, not extraneous ### Practical Implications: Extraneous variables can: - Increase model complexity unnecessarily - Reduce degrees of freedom - Lead to multicollinearity issues - Make interpretation more difficult - Decrease model performance on new data

Author: Tanishq Prabhu