Previous workshops: 2022

Introduction to Multilevel Modelling Using MLwiN, R or Stata, 5-7 July 2022, Online via Zoom

Instructors

Professor George Leckie, Professor William Browne

Summary

This three-day course provides an introduction to multilevel modelling and includes software practicals in your choice of software: MLwiN, R or Stata. We focus on multilevel modelling for continuous and binary responses (dependent or outcome variables) when the data are clustered (nested or hierarchical). These models can be viewed as an extension of conventional linear and logistic regression models to account for and learn from the clustering in the data. Such models are appropriate when, for example, analysing exam scores of students nested within schools, or health outcomes of patients nested within hospitals. Special interest lies in disentangling social processes operating at different levels of analysis by decomposing the within- from the between-cluster effects of covariates (explanatory or predictor variables). Longitudinal data are also clustered, with repeated measurements on individuals or multiple panel waves per survey respondent. Throughout the course we emphasize how to interpret the models and how to decide on what kinds of research question they can be used to explore.

Testimonials

"The course was really excellent - clearly structured and in a logical order. Speakers were fantastic."

"The course was excellent - far exceeded expectations. The course has given me the confidence to use MLM, something I very much lacked before. I feel I understand the theory behind MLM, why each stage is so important, and the various interpretations. Without this course I would be lost. I cannot thank you all enough."

"This was a beautifully constructed course. It was clear throughout that careful thought had been given to providing a balance between lecture content, time for questions and discussion, and practical sessions. Both George and Bill delivered fantastic lectures - explanations were clear and thorough (including critiques of each approach) and content built up in complexity over time with plenty of worked examples of different kinds. The course was superb - can't rate it highly enough."

“I thought it was a really good double act between George and Bill - they are both hugely knowledgeable so having one person focused on the slides and the other manning the chat was a good approach as it meant the teaching didn't get derailed by people's questions.”

"Both George and Bill have excellent presentation styles. I really liked that they 'riffed' off of each other with gentle humour."

Topics

Overview of multilevel modelling
Variance-components models
Random-intercept models with covariates
Between- and within-effects of level-1 covariates
Random-coefficient models
Growth-curve models
Three-level models
Review of single-level logistic regression
Two-level logistic regression

Format

The course will consist of a 2:1 mix of lectures and hands-on practical sessions applying the taught methods to real datasets. The instructors alternate the lecturing. The lectures are software independent. Each lecture is immediately followed by a software practical giving participants the chance to replicate the presented analyses and to consolidate their knowledge. The practicals are offered in participants’ choice of MLwiN, R, or Stata and are self-directed: participants complete the practicals at their own pace. At the end of each practical session the instructors demo the different software. In both the lectures and practicals, participants have opportunities to interact with the instructors.

Zoom

The course will be delivered online via the freely accessible Zoom platform. The lectures will be delivered live. Participants can ask questions via Zoom’s text-based chat facility that will be monitored and answered by one of the instructors not presenting or relayed to the instructor presenting to answer live.

Participants are encouraged to join the lectures live, but recordings of the lectures will be made available shortly afterwards for eight weeks following the course if participants are unable to attend at the scheduled time. After eight weeks, video access will end and can’t be extended.

During the practicals, participants can also speak with the instructors. Participants can use these opportunities to ask specific questions about the course material or about multilevel modelling related to their own research. Each software package will be demonstrated in a different breakout room.

Materials

Participants will be emailed in advance with comprehensive PDF copies of the lecture slides together with point-and-click instructions and datasets for MLwiN, and annotated syntax files and datasets for R and Stata. During the practicals, participants are encouraged to view the lecture slides on a second screen (or tablet etc.), else print copies out to have in front of them. Those choosing to use MLwiN may also want to view on a second screen else print out their point-and-click instructions.

Software

For those choosing to use MLwiN, we will provide instructions as to how to download and install the free teaching version of this software. For those wishing to use Stata or R we assume you are already users of these software so already have these software installed.

MLwiN

MLwiN is dedicated multilevel modelling software developed by our research team over the last 25 years. On this course we will be using the free teaching version of MLwiN. This version works with all the datasets used on the course and a wide range of other teaching datasets which come with the software. We will email you the teaching version prior to the start of the course.

Should you wish to use MLwiN after the course with your own data, you will need to use the regular version of MLwiN. This is free to UK academics (but without user support) reflecting long periods of funding from the UK’s Economic and Social science Research Council (ESRC). For all other users there is a 30-day trial version, but after that you will have to purchase MLwiN if you wish to continue using it to analyse your own data. There are various price options available. http://www.bristol.ac.uk/cmm/software/mlwin/

MLwiN is Windows software, but can be run on Mac via the Wine software or through a virtual machine. Note that Wine will only work on versions of MacOS prior to Catalina. http://www.bristol.ac.uk/cmm/software/mlwin/features/sysreq.html#unix.

Pre-requisites

We assume no prior knowledge of multilevel modelling. However, participants should be familiar with estimating and interpreting linear regression models, including the writing and interpretation of model equations, hypothesis testing and model selection, and the use and interpretation of dummy variables and interaction terms.

We will email in advance a video lecture which provides a review of linear regression accompanied with software instructions and datasets to replicate all analyses in MLwiN, R or Stata.

For those choosing to use MLwiN, we assume no prior knowledge of using this software and so provide step-by-step instructions to allow you to replicate all presented analyses in MLwiN. For those choosing R or Stata, we assume you are already users of these software and so know the basics.

Timings

The course starts and ends each day at 09:30 and 16:00 with a 15-minute morning break and a one-hour break for lunch from 13:00 to 14:00.

Fees

For UK-registered MSc and PhD students - £180
For UK university academics, UK public sector staff, and staff at UK registered charity organisations - £360
For all other participants - £660

Please note, in order to be eligible for the reduced pricing brackets please submit your application using your UK academic/organisational email address.

Cancellation/refunds

A full refund will be given if cancellation occurs three weeks prior to the event. No refund is given after this date. By completing the application form, you are accepting these cancellation terms.

Applications

If you would like to attend either workshop, please complete and submit the online application form (see below). Please note the closing date for applications is 1st May 2022.

If application numbers are high and we cannot accept everyone without compromising the quality of the course, we will offer places on a first come, first service basis. We will email you by 9th May to tell you whether or not your application has been successful. If you are offered a place on either workshop, it will not be confirmed until you have accepted and paid the relevant fee.

If you have any queries, please email info-cmm@bristol.ac.uk.

Go to booking form >>

Terms and Conditions

Please click here to read the booking terms and conditions before completing the booking form. Note that it is the participant’s responsibility to ensure that the Zoom and their choice of MLwiN, R, or Stata software work on their computer in advance of the course, as the Centre for Multilevel Modelling is unable to provide technical support.

MLwiN

MLwiN is Windows software, but can be run on Mac via the Wine software or through a virtual machine such as Parallels, depending on the Mac model and version of MacOS on your machine. http://www.bristol.ac.uk/cmm/software/mlwin/features/sysreq.html#unix .

Introduction to Multilevel Modelling Using MLwiN, R or Stata, 18-20 January 2022, Online via Zoom

Instructors

Professor George Leckie, Professor William Browne

Summary

This three-day course provides an introduction to multilevel modelling and includes software practicals in your choice of software: MLwiN, R or Stata. We focus on multilevel modelling for continuous and binary responses (dependent or outcome variables) when the data are clustered (nested or hierarchical). These models can be viewed as an extension of conventional linear and logistic regression models to account for and learn from the clustering in the data. Such models are appropriate when, for example, analysing exam scores of students nested within schools, or health outcomes of patients nested within hospitals. Special interest lies in disentangling social processes operating at different levels of analysis by decomposing the within- from the between-cluster effects of covariates (explanatory or predictor variables). Longitudinal data are also clustered, with repeated measurements on individuals or multiple panel waves per survey respondent. Throughout the course we emphasize how to interpret the models and to decide on what kinds of research question they can be used to explore.

Testimonials

"The course was really excellent - clearly structured and in a logical order. Speakers were fantastic."

"Both George and Bill have excellent presentation styles. I really liked that they 'riffed' off of each other with gentle humour."

Topics

Overview of multilevel modelling
Variance-components models
Random-intercept models with covariates
Between- and within-effects of level-1 covariates
Random-coefficient models
Growth-curve models
Three-level models
Review of single-level logistic regression
Two-level logistic regression

Format

The course will consist of a 2:1 mix of lectures and hands-on practical sessions applying the taught methods to real datasets. The instructors alternate the lecturing. The lectures are software independent. Each lecture is immediately followed by a software practical giving participants the chance to replicate the presented analyses and to consolidate their knowledge. The practicals are offered in participants’ choice of MLwiN, R, or Stata and are self-directed: participants complete the practicals at their own pace. At the end of each practical session the instructors will demo the different software in breakout rooms. In both the lectures and practicals, participants have opportunities to interact with the instructors.

Zoom

Materials

Software

MLwiN

Pre-requisites

Some participants may wish to refresh themselves of linear regression by reading module 3 of our LEMMA online course. https://www.bristol.ac.uk/cmm/learning/online-course/course-topics.html.

For those choosing to use MLwiN, we assume no prior knowledge of using this software and so we will email in advance a video lecture which provides an introduction to MLwiN in terms of fitting linear regression models. We will also provide step-by-step instructions to allow you to replicate the presented analyses in MLwiN.

Timings

The course starts and ends each day at 09:30 and 16:00 with a 15-minute morning break and a one-hour break for lunch from 13:00 to 14:00.

Fees

For UK-registered MSc and PhD students - £180
For UK university academics, UK public sector staff, and staff at UK registered charity organisations - £360
For all other participants - £660

Please note, in order to be eligible for the reduced pricing brackets please submit your application using your UK academic/organisational email address.

Cancellation/refunds

A full refund will be given if cancellation occurs three weeks prior to the event. No refund is given after this date. By completing the application form, you are accepting these cancellation terms.

Applications

If you would like to attend either workshop, please complete and submit the online application form (see below). Please note the closing date for applications is 14th November 2021.

If application numbers are high and we cannot accept everyone without compromising the quality of the course, we will offer places on a first come, first service basis. We will email you by 22nd November to tell you whether or not your application has been successful. If you are offered a place on either workshop, it will not be confirmed until you have accepted and paid the relevant fee.

If you have any queries, please email info-cmm@bristol.ac.uk.

Go to booking form >>