In regression modeling, what term denotes including higher-order terms to capture curvature in the relationship?

In regression modeling, what term denotes including higher-order terms to capture curvature in the relationship?

• A. Slope
• B. Intercept
• C. Functional Forms
• D. Non-Linear Terms

Answer: (D)

When the relationship between the predictor and the outcome isn’t a straight line, you can capture curvature by adding terms that are higher powers of the predictor, such as x^2 or x^3. These are called non-linear terms because they introduce nonlinearity into the model’s relationship. The basic linear model uses an intercept and a slope to describe a straight-line trend; the intercept is the expected value when x is zero, and the slope is the constant change in the outcome per unit change in x. Those two terms alone can’t model bending or turning points, which is why non-linear terms are added. While the broader idea of functional forms encompasses different shapes the model can take, the action of incorporating higher-order terms to allow curvature is specifically described as including non-linear terms. For example, y = β0 + β1x + β2x^2 uses a quadratic term to let the curve bend, capturing increasing or decreasing effects of x on y depending on the sign and magnitude of β2.

When the relationship between the predictor and the outcome isn’t a straight line, you can capture curvature by adding terms that are higher powers of the predictor, such as x^2 or x^3. These are called non-linear terms because they introduce nonlinearity into the model’s relationship. The basic linear model uses an intercept and a slope to describe a straight-line trend; the intercept is the expected value when x is zero, and the slope is the constant change in the outcome per unit change in x. Those two terms alone can’t model bending or turning points, which is why non-linear terms are added. While the broader idea of functional forms encompasses different shapes the model can take, the action of incorporating higher-order terms to allow curvature is specifically described as including non-linear terms. For example, y = β0 + β1x + β2x^2 uses a quadratic term to let the curve bend, capturing increasing or decreasing effects of x on y depending on the sign and magnitude of β2.