UI Lisa Special Training Workshop on Practical Guide to Logistic Regression Modeling

University Of Ibadan, UI Lisa Special Training Workshop on Practical Guide to Logistic Regression Modeling


Logistic regression is one of the most widely used statistical procedures in research. It is a major important statistical routine in fields such as health-care analysis, medical statistics, ecology, social sciences, education, econometrics and other similar areas. It is for this reason that logistic regression forms a major component of nearly all, if not all, general purpose commercial statistical packages. Logistic regression is typically used by researchers and analysts in general for three purposes: to predict the probability that the outcome or response variable equals 1; to categorize outcomes or predictions; and to access the odds or risk associated with model predictors.

This workshop is not intended to dwell on the esoteric details of the theory behind estimation and inference of logistic regression but rather focus on the fundamental logic of the model and its appropriate application. More specifically, the participants will learn the most important features of the logistic model – how to construct a logistic model, how to interpret coefficients and odds ratios, how to predict probabilities based on the model, and how to evaluate the model as to its fitness or adequacy.

R statistical software is adopted for this training/workshop for obvious reasons. It is currently the most popular and widely used statistical software in majority of articles published in highly rated journals. It is an open ware, meaning that it is possible for users to inspect the actual code used in the analysis and modelling process. It is also free, costing nothing to download into one’s computer.

The training will be hands-on and practical oriented. Participants are expected to come with their laptops pre-installed with the R software, although, installation assistance will be provided at the training venue for those who do not have the software on their computer. Other details are as follow:

Date: January 26 & 27, 2016

Venue: PG Computer Lab, Department of Statistics, UI.

Time: 9.00am – 4.00pm Daily

Fee: N10,000 only

Schedule of Programme

Tuesday 26 January, 2016
PeriodProgrammeResource Person
9.00 – 9.30Arrival and RegistrationCollaborators
9.30 – 10.00Opening and General IntroductionDr.J.F. Ojo
10.00 – 12.00Basics of Logistic Regression ModellingProf. G.N. Amahia
12.00 – 1.00Introduction to R packageDr. K.O. Obisesan
1.00 – 2.00LUNCH BREAK
2.00 – 3.00glm function in RDr. K.O. Obisesan
3.00 – 4.00Logistic Models with a Binary and Continuous PredictorsDr. O.E. Olubusoye
Wednesday 27 January, 2016
9.00 – 11.00Statistics in a Logistic Model and their InterpretationsDr. O.I. Shittu
11.00 – 1.00Checking Logistic Model FitDr. S.O. Oyamakin
1.00 – 2.00LUNCH BREAK
2.00 – 4.00Classification StatisticsDr. O.E. Olubusoye

For more information, please contact:

Room 103, Department of Statistics,

Mathematics/Statistics Complex, U.I.

Email: uistat.lisa@gmail.com

Phone: 08138840067

Expected Coverage of the Topics

1. Basics of Logistic Regression Modelling

· Methods of statistical analysis for response variables measured on continuous, binary, categorical (> 2), ordinal and counts scales.

· Uses of binary response logistic regression model

· The Bernoulli Distribution, the link function, the mean function and the variance function

· The linear predictor, the logit function and its properties

· Distinction between logit and probit models

2. Introduction to R package

· Installing R

· Work Session in windows

· Help – Online and CRAN

· Reading Data from file

· Exporting Results

· Manipulating variables

3. glm function in R

· The structure of GLMs

· The major glm arguments: formula, family, data, and subsets

· Fitting the binary logistic regression Model

· Parameter estimates for logistic Regression

4. Logistic Models with a Binary and Continuous Predictors

· Single predictor (binary) case: Calculating and interpreting coefficients, odds, odds ratio and probabilities

· Single predictor (continuous) case: Calculating and interpreting coefficients, odds, odds ratio and probabilities

· Logistic Models: Multiple Predictors

5. Statistics in a Logistic Model and their Interpretations

· Basic Statistics: standard errors, z –statistics, p-values, confidence intervals

· Information Criterion Tests: AIC, Finite Sample, BIC, etc

· Risk Factors, Confounders, Effect Modifiers, and Interactions

6. Checking Logistic Model Fit

· Pearson Chi2 Goodness-of-fit Test

· Likelihood Ratio Test

· Residual Analysis

· Conditional Effects Plot

7. Classification Statistics

· S-S Plot

· ROC Analysis

· Confusion Matrix

· Hosmer-Lemeshow Statistic

Leave a Reply