2.10 Conclusion

• Do not assume a linear regression model is restricted to linear effects
• Useful to distinguish between linear and nonlinear effects
  • And between additive and interactive effects
• Linear regression models can produce nonlinear and/or interactive effects
• All models for categorical DVs produce nonlinear and interactive effects
  • Tools for interpretation are largely the same as we covered in this section:
    • Predictions
    • Visualizations
    • Marginal effects