 |
| |||||
 |
| |||||||||||||||||||||
Logistic Regression: A Primer
Comment from the Stata technical groupThis monograph provides an excellent introduction to logistic regression from first principles. It is an ideal tutorial for those who are familiar with standard linear regression and wish to branch out for the first time into more complex generalized linear models, for which logistic regression (regression with a binary response) is a good starting point. Making the jump from linear regression to logistic regression introduces many issues, including model fitting (least squares versus maximum likelihood), interpretability (odds ratios), interpretation of marginal effects (which are not constant for logistic regression), and regression diagnostics. These topics are covered fully in the context of logistic regression, and where appropriate, the comparison between linear and logistic regression is made. A chapter on probit analysis is also provided. Table of contentsSeries Editor’s Introduction
Preface
1 The Logic of Logistic Regression
Regression with a Dummy Dependent Variable
Transforming Probabilities into Logits Linearizing the Nonlinear Summary 2 Interpreting Logistic Regression Coefficients
Logged Odds
Odds Probabilities Tests of Significance Standardized Coefficients An Example Summary 3 Estimation and Model Fit
Maximum Likelihood Estimation
Log Likelihood Function Estimation Tests of Significance Using Log Likelihood Values Model Evaluation An Example Summary 4 Probit Analysis
Another Way to Linearize the Nonlinear
Probit Analysis Interpreting the Coefficients Maximum Likelihood Estimation An Example Summary 5 Conclusion
Notes
Appendix: Logarithms
The Logic of Logarithms
Properties of Logarithms Natural Logarithms Summary References
About the Author
Acknowledgments
|
|||||||||||||||||||||
){kind=link}
){kind=link}