Using the appropriate regression coefficient as your measure of association, determine whether confounding exists. Explain.

Applied Regression Analysis and Other Multivariable Methods, 5^th ed, Kleinbaum et. al.

Incorporate any generated computer output.  Show all work.

Chapter 11:   Confounding and Interaction                                   Due ________________

Pages: 242–256           2, 3, 4

Problem 2:

Using the appropriate regression coefficient as your measure of association, determine whether confounding exists.  Explain.

Suppose that confounding was defined to require a comparison of crude versus adjusted (partial) correlation coefficients.  What conclusion would you draw?

Problem 3:

Using the appropriate regression coefficient as your measure of association, determine whether confounding exists.  Explain.

Suppose that confounding was defined to require a comparison of crude versus adjusted (partial) correlation coefficients.  What conclusion would you draw?

What does this example illustrate about using a test of the hypothesis H₀: b₂ = 0 in order to assess confounding.

Problem 4:

Fill in the following ANOVA table for the model .

Test if X₂ should be entered into a model already containing X₁.

Test if X₃ should be entered into a model already containing X₁ & X₂.

Test if X₂ & X₃ should be entered into a model already containing X₁.

Based upon the previous tests, what is the most appropriate regression model?

Based upon the information provided, can you assess whether X₁ is a confounder of the X₂–Y relationship?  Explain.

Assignments

Applied Regression Analysis and Other Multivariable Methods, 5^th ed, Kleinbaum et. al.

Incorporate any generated computer output.  Show all work.

Chapter 11:   Confounding and Interaction                                   Due ________________

Pages: 242–256           2, 3, 4

Problem 2:

Using the appropriate regression coefficient as your measure of association, determine whether confounding exists.  Explain.

Suppose that confounding was defined to require a comparison of crude versus adjusted (partial) correlation coefficients.  What conclusion would you draw?

Problem 3:

Using the appropriate regression coefficient as your measure of association, determine whether confounding exists.  Explain.

Suppose that confounding was defined to require a comparison of crude versus adjusted (partial) correlation coefficients.  What conclusion would you draw?

What does this example illustrate about using a test of the hypothesis H₀: b₂ = 0 in order to assess confounding.

Problem 4:

Fill in the following ANOVA table for the model .

Test if X₂ should be entered into a model already containing X₁.

Test if X₃ should be entered into a model already containing X₁ & X₂.

Test if X₂ & X₃ should be entered into a model already containing X₁.

Based upon the previous tests, what is the most appropriate regression model?

Based upon the information provided, can you assess whether X₁ is a confounder of the X₂–Y relationship?  Explain.