Advanced Multiple Imputation Methods to Deal with Missing Data

An online short course

This course aims to extend the application of multiple imputation methods to deal with missing data in complex analyses, including data missing not at random and multilevel data.

Course date 2 March 2021
Course fee £220
Course Organisers Dr Rachael Hughes & Dr Rosie Cornish

Prerequisites

Please ensure you meet the following prerequisites before booking:

Knowledge Prior attendance of the Multiple Imputation for Missing Data short course (or equivalent introductory course to missing data concepts and multiple imputation) or be familiar with the concept of multiple imputation, and have used it in practice.
Software You must have Stata (version 13 or later)* and R (version 4.0.3 or later) installed in advance of the course.
*Internal University of Bristol participants will be provided with access to Stata version 16 on the first day of the course.

Course format

This 1-day course will be online and consist of pre-recorded lectures, live summary plus question & answer sessions, and live computer practical exercises. Each pre-recorded lecture will be followed by a live session which will summarise the main points of the lecture and allow participants to ask questions about the contents of the lecture.

Participants have the choice of completing the course in one day (i.e., following the timetable as scheduled) or listening to the pre-recorded lectures before the start of the course and attending only the live sessions on the day of the course.

The course is timetabled to start at 08:30 and end at 17:30, including time for coffee breaks and lunch.

Course objectives

By the end of the course participants will be able to understand and use advanced multiple imputation methods when:

  1. data are missing not at random;
  2. the substantive analysis of interest is a multilevel model;
  3. the substantive analysis of interest is a propensity score analysis with partially observed covariates.

Who the course is intended for

The course is intended for statisticians, health economists, epidemiologists and other researchers who are involved in performing statistical analyses of epidemiological datasets with missing data. It is assumed that participants will have attended the 'Multiple Imputation for Missing Data' course (or equivalent introductory course to missing data concepts and multiple imputation). Participants should be familiar with: the concept of multiple imputation, and have used it in practice; standard regression methods for dichotomous and continuous outcomes beyond the basic introductory level; and using software packages, Stata or R, for statistical analyses of the data.

Course outline

This course builds on prior knowledge of multiple imputation for dealing with missing data, and extends this to the application of multiple imputation in complex analyses.

The course will include: 

  1. brief revision of theory and practice of multiple imputation in simple scenarios;
  2. guidance on conducting analyses when data are missing not at random;
  3. introduction to multilevel multiple imputation;
  4. guidance on using R package jomo to impute multilevel data;
  5. recommendations on the optimal way to implement multiple imputation for a propensity score analysis.

Recommended reading

Hughes RA, Heron J, Sterne JAC, Tilling K. Accounting for missing data in statistical analyses: multiple imputation is not always the answer. International Journal of Epidemiology 2019; 48: 1294–304.

Online Course Bookings


Bookings are open for online courses running in 2021.

Well thought through, well-structured and well presented. Good to extend coverage from the first Multiple Imputation course to the more complex situations covered here.

Course feedback, May 2019

Related short courses

Coronavirus (COVID-19)

We may need to make responsive changes to our courses at short notice in order to follow the latest Public Health, Government and University guidance on coronavirus (COVID-19).

Edit this page