This course focuses on analyzing categorical response data in scientific fields. The SAS/STAT procedures addressed are PROC FREQ, PROC LOGISTIC, and PROC GENMOD. The ODS Statistical Graphics procedures used are PROC SGPLOT and PROC SGPANEL. The course is not designed for predictive modelers in business fields, although predictive modelers can benefit from the content of this course.

Изучается

• use the FREQ procedure for preliminary analyses
• recognize when logistic regression is appropriate
• write code in the LOGISTIC procedure for binary, ordinal, and nominal logistic regression
• create effect plots and odds ratio plots using ODS Statistical Graphics
• create logit plots
• use automatic model building options in PROC LOGISTIC
• assess models for fit and influential observations using PROC LOGISTIC
• assess functional form of the model effects using PROC GENMOD
• create ROC curves for measuring sensitivity and specificity
• perform exact and conditional logistic regression with PROC LOGISTIC
• analyze repeated and clustered data using Generalized Estimating Equations (GEE's) in the GENMOD procedure.

Аудитория

Biostatisticians, epidemiologists, social scientists, and physical scientists who analyze categorical response data and predictive modelers who would like to learn more about the statistical background of logistic regression

Доступные форматы обучения	Продолжительность
e-Learning:	24 часа(ов)/180 день доступ

Before attending this course, you should

• have a working knowledge of statistical modeling, including concepts of regression, analysis of variance, and contingency table analysis, which you can obtain in the Статистика 1: Введение в дисперсионный анализ, линейную и логистическую регрессию course
• have an understanding of basic syntax in SAS procedures and DATA steps
• have experience in executing SAS programs and creating SAS data sets, which you can gain by completing the Программирование на языке SAS, Часть 1. Основы course
• have experience analyzing frequency tables using SAS software
• have completed a course in statistics that covers linear regression and logistic regression, which you can achieve by completing the Статистика 1: Введение в дисперсионный анализ, линейную и логистическую регрессию course.

Используется SAS/STAT ПО

Categorical Data and Contingency Table Analysis

• introduction to categorical data
• associations among categorical variables
• stratified contingency table analysis

Binary Logistic Regression

• introduction to logistic regression
• adding categorical predictors and the CLASS statement

Model Building

• empirical logit plots
• confounding and interactions
• automatic model selection
• customized tests

Model Illustration and Assessment

• interaction illustration
• model assessment
• ROC curves
• outlier detection

Multinomial Logistic Regression

• ordinal logistic regression
• nominal logistic regression

Special Topics

• correlated observations
• GEE regression models
• conditional logistic regression
• failure to converge and small samples