Advanced Course in Multiple Regression and Causal Inference

Traditional training – (5 days, R15,000.00 excl. VAT)
Online training – (4 weeks, R11,500.00 excl. VAT)

About This Course

Every researcher wants to be sure that observed associations are valid. After using bi-variate tests (such as chi-squared test, t-test, ANOVA, etc) to demonstrate association between two variables of interest, the next critical question is whether those associations are valid i.e. whether they can in fact be explained by other variables (confounding) or study methodology (bias). In other words, in strengthening causal inference, it is vital to eliminate the role of confounding and bias. Multiple regression remains the most well-known approach for controlling for confounding and estimating independent effects.

Multiple regression models are used to determine risk factors after adjusting for confounding. They are also commonly used to build models for predicting an outcome. Therefore, researchers should be able to build, understand and interpret multiple regression models.

Course Content

Traditional Course Content

i

Day 1

  • Review of bi-variate tests and measures of effect
  • Bias, confounding and interaction
  • How to control for confounders: design and analysis
  • Application of regression models
  • Linear regression models for continuous outcomes
  • T-test, ANOVA, correlation and linear regression
i

Day 2

  • Interpretation of ANOVA and linear regression outputs
  • Simple and multiple linear regression
  • Model fitting and variable selection
  • Assumptions of linear regression and post-regression diagnostics
i

Day 3

  • When linear regression assumptions are not met
  • Quantile regression
  • Fitting simple and multiple models
  • First project report
i

Day 4

  • Logistic regression models for binary outcomes
  • Odds, log odds, and logistic regression
  • Chi-squared test simple logistic regression
  • Multiple logistic regression
i

Day 5

  • Model fitting and variable selection
  • Assumptions of linear regression and post-regression diagnostics
  • Second project report

Online Course Content

i

Week 1

  • Review of bi-variate tests and measures of effect
  • Bias, confounding and interaction
  • How to control for confounders: design and analysis
  • Application of regression models
  • Linear regression models for continuous outcomes
  • T-test, ANOVA, correlation and linear regression
i

Week 2

  • Interpretation of ANOVA and linear regression outputs
  • Simple and multiple linear regression
  • Model fitting and variable selection
  • Assumptions of linear regression and post-regression diagnostics
i

Week 3

  • When linear regression assumptions are not met
  • Quantile regression
  • Fitting simple and multiple models
  • First project report
i

Week 4

  • Logistic regression models for binary outcomes
  • Odds, log odds, and logistic regression
  • Chi-squared test simple logistic regression
  • Multiple logistic regression
  • Model fitting and variable selection
  • Assumptions of linear regression and post-regression diagnostics
  • Second project report

The course is taught using Stata. So, it is a pre-requisite that the participants have at least one of the following:

  • Have attended CESAR’s level 1 course: Data Analysis and Management Course using Stata
  • Have previous experience working with Stata
  • Have no previous experience with Stata but can conduct bi-variate tests using another statistical package.

Course outcomes

The course will teach participants the following:

This course teaches participants to fit multiple regression models and estimate measures of effect. Simple and multiple regression models for crude and adjusted effects will be covered. Three types of regression models will be covered – linear, quantile and logistic regression. The course uses a fine blend of didactic lectures, group exercises for every session and two project reports that simulate real life scenarios. 

After the course, participants will be able to understand the link between bi-variate tests and regression model, for example, the link between t-test and linear regression and chi-squared test and logistic regression. They will also be able to build causal or predictive models and interpret the outputs. Furthermore, they will be able to adjust for confounders and estimate independent effects, thereby strengthening causal inference in their research.

Who Should Attend?

  • Researchers
  • Biostatisticians
  • Research analysts
  • Data analysts
  • Economists
  • HODs
  • Clinicians
  • Epidemiologist
  • Programme managers
  • Postgraduate students
  • Market researchers
  • Clinical and medical researchers
  • Scientists
  • Government practitioners

Application

For more details about our services contact:

Dave Temane
Email: info@cesar-africa.com
Tel: +27 11 403 1411

Price Includes

  • Course attendance
  • Full refreshments: lunch
  • Welcome tea
  • Two breaks for tea including pastries
  • Course lecture notes and training manual
  • Complimentary parking
  • Certificate of attendance

Application for Open Training

Click to confirm

1 + 13 =

Application for Online Training

Click to confirm

7 + 12 =