QCM : Mastering Multiple Regression and ANOVA — 11 questions

Questions et réponses du QCM

1. What does a multiple regression model express about a numerical outcome?

The median as a nonlinear function of one predictor
The conditional mean as a linear function of several predictors
The variance as a step function of categorical groups
The raw outcome as an average of residuals

The conditional mean as a linear function of several predictors

Explication

A multiple regression model represents the conditional mean of the outcome as a linear combination of several predictors. The other options describe different statistical ideas that are not what multiple regression models estimate.

2. What is a multiple regression model?

A model that expresses the expected value of a numerical outcome as a linear function of several predictors.
A model that explains the variation in a response variable solely through a single predictor.
A model that compares the means of different groups using dummy variables.
A model that predicts a categorical outcome based on multiple predictors.

A model that expresses the expected value of a numerical outcome as a linear function of several predictors.

Explication

A multiple regression model expresses the conditional mean of a numerical outcome as a linear function of several predictors, allowing for the analysis of their combined effects.

3. In a regression with two predictors, what geometric form represents the fitted model?

A cluster of points around the mean
A plane with a separate slope for each predictor
A curved surface with one turning point
A line with one slope and one intercept

A plane with a separate slope for each predictor

Explication

With two predictors, the regression surface is a plane, and each predictor has its own slope. A line would apply to one predictor, not two.

4. What is the primary purpose of a multiple regression model?

To compare the means of different groups based on a categorical variable.
To analyze the relationship between two categorical variables.
To express the conditional mean of a numerical outcome as a linear function of several predictors.
To determine the causality between a predictor and an outcome.

To express the conditional mean of a numerical outcome as a linear function of several predictors.

Explication

A multiple regression model aims to express the expected value of a numerical outcome as a linear combination of multiple predictors, allowing for the analysis of their combined effects.

5. How is a residual for one observation defined in regression?

The average of all observed values minus the predicted value
The squared difference between observed and predicted values
The predicted value minus the observed value
The observed value minus the predicted value

The observed value minus the predicted value

Explication

A residual is the prediction error for one case, computed as y - ŷ. Squaring the difference is used in SSE, not in the residual itself.

6. What is the primary purpose of regression inference and the coefficient of multiple determination in a regression model?

To identify causal relationships and eliminate confounding variables from the analysis.
To assess the significance of predictors and quantify the proportion of outcome variance explained by the model.
To compare group means and test for differences across categories in the outcome variable.
To determine the best predictors by minimizing residuals and to measure the total variation in the outcome.

To assess the significance of predictors and quantify the proportion of outcome variance explained by the model.

Explication

Regression inference involves testing the significance of predictors, while the coefficient of multiple determination ($R^2$) quantifies how much of the total outcome variance is explained by the model, helping assess its usefulness.

7. What does the least squares criterion choose in a regression model?

The coefficients that minimize the sum of squared errors
The coefficients that maximize the correlation with the outcome
The model with the smallest number of predictors
The predictor with the largest individual slope

The coefficients that minimize the sum of squared errors

Explication

Least squares selects regression coefficients so that the sum of squared errors is as small as possible. It is not about maximizing correlation or minimizing the number of predictors.

8. When was the concept of causality in multivariate relationships formally established as a key criterion for interpreting statistical associations?

In the early 20th century during the development of classical statistics.
In the 19th century with the advent of experimental design.
In the 2000s with the advent of machine learning algorithms.
In the 1960s with the rise of causal inference frameworks.

In the 1960s with the rise of causal inference frameworks.

Explication

The formal emphasis on causality as a key criterion in statistical analysis gained prominence in the 1960s with the development of causal inference frameworks, such as those by Judea Pearl, which clarified the conditions needed to interpret associations as causal.

9. How does dummy coding of categorical predictors differ from using numeric codes for the same categories in regression analysis?

Dummy coding uses multiple dummy variables for each category, while numeric coding uses a single variable with multiple levels.
Dummy coding uses indicator variables with 0/1 values and omits one level as reference, while numeric codes assign arbitrary numbers to categories without implying order.
Dummy coding and numeric coding are identical in regression analysis; both assign numbers to categories without any difference.
Dummy coding assigns arbitrary numbers to categories, whereas numeric codes use 0/1 indicators and omit one level as reference.

Dummy coding uses indicator variables with 0/1 values and omits one level as reference, while numeric codes assign arbitrary numbers to categories without implying order.

Explication

Dummy coding creates indicator variables with 0/1 values and omits one level to serve as a reference, avoiding collinearity; numeric codes simply assign numbers without necessarily implying order or comparison.

10. Who is credited with proposing the use of dummy coding to represent categorical predictors in regression analysis?

John Tukey
George Box
Jerzy Neyman
Ronald Fisher

Ronald Fisher

Explication

Ronald Fisher is credited with developing the concept of dummy coding, which allows categorical variables to be included in regression models by representing categories with indicator variables.

11. What is the effect of including dummy-coded categorical predictors in a regression model on the interpretation of group differences?

It allows for direct comparison of each group to a reference category, revealing the effect of group membership on the outcome.
It converts categorical variables into continuous predictors, which can distort the analysis.
It causes collinearity issues that prevent interpretation of individual group effects.
It eliminates the need for a reference category, enabling comparison among all groups simultaneously.

It allows for direct comparison of each group to a reference category, revealing the effect of group membership on the outcome.

Explication

Including dummy-coded predictors in regression models allows for direct comparison of each group to a designated reference group, thereby revealing the effect of group membership on the outcome. This approach does not eliminate the need for a reference category nor necessarily cause collinearity if properly implemented.

Révisez avec les flashcards

Mémorisez les réponses avec 9 flashcards sur Mastering Multiple Regression and ANOVA.

Multiple regression — definition?

Predicts an outcome using multiple predictors.

Multiple Regression Model

Predicts outcome as linear function of predictors.

Residuals — role?

Measure prediction errors for individual observations.

Voir les flashcards →

Approfondir avec la fiche

Consultez la fiche de révision complète sur Mastering Multiple Regression and ANOVA.

Voir la fiche →

Cours similaires

Actions et mouvement

Physique

Gel pour les veines

Grandeurs composées et conversions

Mathématiques

Réciproque de Pythagore

Mathématiques

Réussir ses études efficacement

Les émotions humaines

Crée tes propres QCM

Importe ton cours et l'IA génère des QCM avec corrections en 30 secondes.

Générateur de QCM