ONLINE Introduction to Multilevel Modelling

This Advance Quantitative Methods workshop is open to all ESRC and non-ESRC funded Social Sciences PhD students within the WRDTP partner universities.

Multilevel models (MLMs) allow researchers to use data that has a multilevel structure, such as pupils within schools, individuals within neighbourhoods, or longitudinal panel data where observations are nested within individuals. A failure to take account of such structures can lead to results that are not only biased or anticonservative, but also missing important aspects of the societal processes they are trying to uncover. This course will build on standard single level regression models to provide an introduction to 2-level multilevel models, that can be used in a range of disciplines across the social sciences and beyond.

Given the short length of the course, we will not cover all aspects of multilevel models. However, participants will gain experience using relatively simple multilevel models, and be pointed to resources for more complex models.

Learning Outcomes

Learn how to implement multilevel models for data with two levels and a continuous Y variable. This will include random intercept models, and random slopes models, that allow the effect of a variable to vary across higher- level units
Learn how to use the MLwiN software to implement multilevel models
Learn how to produce a range of plots to visualise multilevel effects

Prerequisites

Participants must have a good understanding of multiple regression modelling and inferential statistics, especially regression intercepts and slopes, standard errors, t-ratios, residuals, and the concepts of variance and co-variance. No experience of multilevel modelling, or MLwiN, is expected. As a check, participants should be comfortable with completing module 3 of the LEMMA online course that can be found here: https://www.cmm.bris.ac.uk/lemma

Workshop organiser/ leader

Dr Andrew Bell

Director of Research and Lecturer in Quantitative Social Sciences, Sheffield Methods Institute

Andy is a lecturer in Quantitative Social Sciences at the University of Sheffield, having completed his undergraduate degree at the University of Bristol. His current substantive research focuses on mental health from a life course perspective, but also spans a diverse range of other subject areas, including geography, political science, social epidemiology and economics. Methodologically, Andy’s interests are in the development and application of multilevel models, with work focusing on age-period-cohort analysis and fixed and random effects models.

There are 30 places available on this training workshop

Please direct any queries related to this workshop to smi@sheffield.ac.uk

PLEASE NOTE: Students are responsible for arranging travel to and from this Advance Quantiative Methods Training session. The WRDTP cannot reimburse travel costs to this session.

BOOK YOUR PLACE

Hourly Schedule

Programme

9.00am - 9.30am
Registration
Coffee/ tea

9.30am - 10.15am
Lecture: Multilevel modelling
The concepts and the alternatives

10.15am - 11.00am
Lecture: Running simple multilevel models
Null model; random intercepts; higher level predictors; concepts: VPC, variance, residuals etc.

11.00am - 12.30pm
Practical: MLwiN, Random Intercepts, Higher level predictors
Coefficients and Standard errors; Caterpillar Plot; Varying line plot; Calculate the VPC

12.30pm - 1.30pm
Lunch

1.30pm - 2.30pm
Lecture: Random slopes

2.30pm - 3.30pm
Practical: Random slopes
Caterpillar plots; Varying line plot; Model comparison

3.30pm - 4.00pm
Comofrt break

4.00pm - 4.30pm
Lecture: Variance functions

4.30pm - 5.00pm
Practical: Variance functions
Variance function plots

5.00pm - 5.30pm
Wrap up and 'homework'
Level 1 complex variance; cross level interactions; categorical X; non-continuous Y; MCMC; non-nested structures; other software; other resources (LEMMA practicals)