Description

This workshop is an introduction to regression modelling in nutrition research, estimation and visualization of marginal effects as well as multiple imputation to deal with missing data. The workshop was co-developped with Hannah Yang Han (McGill University).

The workshop outline was as follows:

• Introduction. brief refresher on regression analysis
  • Nomenclature and equation of regression models
  • Description of linear regression models and assumptions
  • Description of logistic regression models and assumptions
  • Description of methods to relax model assumptions
• Part 1. Marginal effects: getting insights from a model
  • Understand limitation of typical regression output;
  • Learn how to estimate marginal effects for categorical/continuous variables;
  • Learn how to visualize regression models;
  • Learn how to estimate custom hypotheses.
• Part 2. Multiple imputation to deal with missingness
  • Learn about types of missingness
  • Learn about imputation methods
  • Learn how to apply multiple imputation

Presentation

The introduction text and slides for the workshop are available online on GitHub: Introduction | Part 1 | Part 2

Code

Supporting R and SAS codes are available in a Github repository. The data used for example is a subset of 1’000 respondents aged 19-70 years from the Canadian Community Health Survey (CCHS) 2015 - Nutrition.