LEMMA: Multilevel Modelling online course

Our on-line multilevel modelling course is now available: 'LEMMA' (The Learning Environment for Multilevel Methodology and Applications) contains a set of graduated modules starting from an introduction to quantitative research progressing to multilevel modelling of continuous data.

"A great course. Just what I needed. Explained everything in detail. The tests were particularly good as they highlighted aspects you thought you understood but hadn't really grasped. A great way to get back into stats. Thanks" (PhD student, 2009)

Log on or register to course

Whether you are new to statistical modelling or an advanced user, we hope that you will find our materials useful. We assume that users have attended at least an introductory statistics course in the past, but Modules 1 and 2 are provided as refreshers. More experienced quantitative researchers may wish to skip to Modules 3 or 4.

Before starting to work through the materials, we strongly encourage you to test your current understanding of statistics by taking our prerequisites quiz. If you find the quiz questions difficult, we suggest that you refresh your knowledge by working through Modules 1 and 2. On the other hand, those who find the questions easy may prefer to go straight to Module 3 (Multiple regression) or Module 4 (Multilevel structures and classifications). You will also find quiz questions throughout the modules to allow you to assess your understanding of the material.

The system is under continuous development, and will be extended to include further modules on advanced multilevel modelling and applications of multilevel modelling to data from different contexts. Other planned enhancements include overview videos for each module and a more extensive glossary.

FAQs or frequently-asked questions about the LEMMA course

What topics are covered?
What is the target audience for the materials, do I need to have a strong background in statistics?
- Are you a beginner in statistics?
- Are you an experienced quantitative researcher?
I have attended one of the Centre for Multilevel Modelling workshops. Do the online materials cover the same topics?
I am planning to attend on of the Centre for Multilevel Modelling workshops. Should I work through these materials first?
How can I test my understanding of the material?
I do not have a copy of MLwiN, how can I do the practical exercises?
I do not use MLwiN, will I still find these materials useful?
Do I have to pay to access the materials?
Why do I need to register?
Can I use the materials in my own teaching?
Who designed the materials?
Contact us for technical help
Why can't we answer questions about the content of the materials?
How do I get to the course?
When I try to open the worksheets for the LEMMA course I get an error message
When I double click on one of the worksheets for the LEMMA course, it opens with notepad not MLwiN

What topics are covered?

There are currently seven modules available:

Using quantitative data in research
Introduction to quantitative data analysis
Multiple regression
Multilevel structures and classifications
Introduction to multilevel modelling
Regression Models for Binary Responses
Multilevel Models for Binary Responses

Further modules will be added as they are developed.

1. Using quantitative data in research: concepts and definitions

In Module 1 we look at quantitative research and how we collect data, in order to provide a firm foundation for the analyses covered in later modules. The aims of Module 1 are:

To give a broad overview of how research questions might be answered through quantitative analysis. Such questions as the following are explored: How does quantitative analysis relate to other methods of inquiry? Why is it required and what sorts of evidence can it supply?
To introduce the vocabulary of quantitative analysis and specify the common terminology to be used in later modules. Of particular importance is the operational definition of research concepts (how we get from real world characteristics to numbers in our data set) and how this leads to observable variables at different levels of measurement.
To introduce sources of data and concepts relating to how it may be possible to generalise results from samples of various kinds to the populations they are drawn from.
To discuss how variables are defined, what different types there are, and how this may influence how they are analysed.
To give some emphasis to certain ideas such as the nature of variability or the recognition of hierarchical units of analysis that are central to multilevel modelling.

2. Introduction to quantitative data analysis

The aim of this module is to give an overall view of some the principles of effective data analysis. The focus is on how we summarise data to uncover patterns of explanation and relationships between variables. Some key statistical vocabulary is introduced and concepts are illustrated by example. You will learn about the following:

Ways of summarising the shape and pattern of a single variable, at different levels of measurement.
Understanding the role of variability in comparative analysis and the study of relationships.
The elaboration of relationships, the importance of control variables, and the concepts of confounding, suppression and interaction.
The essential parts of a statistical model: pattern and residual variation.
Key concepts in inference from samples.

3. Multiple Regression

Multiple regression is a technique used to study the relationship between an outcome variable and a set of explanatory or predictor variables. Module 3 covers the following topics:

Regression with a single continuous explanatory variable
Comparing groups: regression with a single categorical explanatory variable (dummy variables)
Regression with more than one explanatory variable (multiple regression), including:
- A discussion of the idea of statistical control
- The multiple regression model for continuous and categorical explanatory variables
- Modelling non-linear relationships
Interaction effects (allowing the effect of one explanatory variable X₁ to depend on the value of another X₂)
- Allowing the slope of the relationship between Y and X₁ to vary across groups defined by a categorical variable X₂ by 1) fitting separate models for each value of X₂, and 2) fitting an interaction between X₁ and X₂
- Testing for interaction effects
Checking model assumptions in multiple regression
- Checking the normality assumption
- Checking the homoskedasticity (equal residual variance) assumption
- Outliers

The ideas are illustrated in analyses of hedonistic attitudes in Europe (using data from the European Social Surveys) and of trends in educational attainment (using data from the Scottish Youth Cohort Study).

4. Multilevel structures and classifications

Multilevel modelling is designed to explore and analyse data that come from populations which have a complex structure. This module aims to introduce:

a range of multilevel structures and classifications and how they correspond to real-world situations, research designs, and/or social-science research problems;
the different types of data frames associated with each structure and how subscripts are used to represent structure;
targets of inference;
the distinction between levels and variables, and fixed and random classifications;
the notion that multilevel structures are likely to generate dependent, correlated data that requires explicit modelling;
the difference between long and wide forms of data structures;
the advantages, both technical and substantive, of using a multilevel model, and the disadvantages of not doing so.

5. Introduction to Multilevel Modelling

In the social, medical and biological sciences multilevel or hierarchical structures are the norm. Examples include individuals nested within geographical areas or institutions (e.g. schools or employers), and repeated observations over time on individuals. Other examples of hierarchical and non-hierarchical structures were given in Module 4. When individuals form groups or clusters, we might expect that two randomly selected individuals from the same group will tend to be more alike than two individuals selected from different groups. For example, children learn in classes and features of their class, such as teacher characteristics and the ability of other children in the class, are likely to influence a child's educational attainment. Because of these class effects, we would expect test scores for children in the same class to be more alike than scores for children from different classes. Multilevel models – also known as hierarchical linear models, mixed models, random effects models and variance components models – can be used to analyse data with a hierarchical structure. Throughout this module we refer to the lowest level of observation in the hierarchy (e.g. student) as level 1, and the group or cluster (e.g. class) as level 2.

Module 5 will cover the following topics:

Comparing groups (level 2 units) using multilevel modelling
- Estimating group effects
- Random effects (multilevel) models vs. fixed effects models (i.e. analysis of variance or regression with dummy variables for groups)
Multilevel regression with a single level 1 explanatory variable
- Allowing the regression intercept to vary across groups: random intercept models
- Why are standard errors underestimated if clustering is ignored?
Allowing for different slopes across groups: random slope models
- Allowing both the intercept and slope of a regression line to vary across groups
- Centring explanatory variables
- Example of a random slope (coefficient) for a dichotomous explanatory variable
- Between-group variance as a function of explanatory variables
Adding level 2 explanatory variables
- Contextual effects
- Models with level 1 and level 2 versions of the same explanatory variable: within-group, between-group and contextual effects
- Cross-level interactions: allowing the effect of a level 1 variable to depend on the value of a level 2 variable
Complex level 1 variance: allowing within-group variance to depend on explanatory variables

The ideas are illustrated in analyses of hedonistic attitudes in 20 European countries (using data from the European Social Surveys) and of between-school variation in trends in students' educational attainment (using data from the Scottish Youth Cohort Study). The same datasets are analysed in Module 3 using multiple regression, ignoring country and school effects respectively. In this module, we emphasise the substantive insights that can be gained from a multilevel modelling approach.

6. Regression Models for Binary Responses

In Module 3 we considered multiple linear regression models for the relationship between a continuous response variable and a set of explanatory variables which may be continuous or categorical. Regression models need to be adapted to handle categorical response variables and, in this module, we consider methods for a particular type of categorical variable: binary or dichotomous responses, that is variables with only two categories. Module 6 covers the following topics:

Preliminaries: mean and variance of binary data
Moving towards a regression model for y: the linear probability model
- Revision of linear regression
- Applying linear regression to binary y: the linear probability model
- Example: application of the linear probability model to US voting intentions
Generalised linear models
- A general model for the response probability
- The logit/logistic model
- The probit model
- The complementary log-log model
- Choice of link function
- Estimation of the generalised linear model
Latent variable representation of a generalised linear model
Application of logit and probit models to state differences in US voting intentions
- Probabilities, odds and odds ratios
- Interpretation of a logit model
- Comparison of probit and logit coefficients
- Interpretation of a probit model
- Significance testing and confidence intervals
Adding further predictors in the analysis of US voting intentions
- Interpretation of a logit model using odds ratios
- Interpretation of logit and probit models using predicted response probabilities
Interaction effects
Modelling proportions
- The binomial distribution
- Two approaches to analysing proportions
- Example: analysis of state-level voting intentions in the 2004 US general election
- Extra-binomial variation (over and under dispersion)
- Multilevel modelling

The ideas are illustrated in analyses of voting intentions in the 2004 US general election (using data from the National Annenberg Election Study) and uptake of antenatal care in Bangladesh (using data from the 2004 Bangladesh Demographic and Health Survey).

7. Multilevel Models for Binary Responses

In Module 6 we saw how multiple linear regression models for continuous responses can be generalised to handle binary responses. In this module, we consider how these methods can be extended for the analysis of clustered binary data. We show that many of the extensions to the basic multilevel model introduced in Module 5 – for example random slopes and contextual effects – apply also to binary responses. However, there are some important new issues to consider in the interpretation and estimation of multilevel binary response models. Module 7 covers the following topics:

Two-level random intercept model for binary responses
- Generalised linear random intercept model
- Random intercept logit model
- Example: between-state variation in voting intentions in the US
Latent variable representation of a random intercept model for binary responses
- Two-level random intercept threshold model
- Comparison of a single-level and multilevel threshold model
- Impact of adding a level 1 explanatory variable to a two-level model
- Variance partition coefficient in terms of y*
Population-averaged and cluster-specific effects
- Marginal model for clustered binary data
- Interpretation of coefficients from random effects and marginal models
- Example: comparison of marginal and random intercept models fitted to the US election data
Predicted probabilities from a multilevel model
A two-level random slope model
- A random slope logit model
- Example: allowing the relationship between income and voting intentions in the US to vary across states
- Two random coefficients: allowing income and urban-rural differentials in voting intentions to vary across states
Adding level 2 explanatory variables: contextual effects
- A random intercept model with a level 2 explanatory variable
- Cross-level interactions
Estimation of binary response models
- Comparison of estimation procedures
- Some practical guidelines on the choice between estimation procedures

What is the target audience for the materials? Do I need to have a strong background in statistics?

Are you a beginner in statistics?

This set of modules is not aimed at the complete beginner in quantitative analysis. We rather expect most users will have some familiarity with many basic ideas and have some prior experience of traditional elementary statistical methods, perhaps up to two variable regression and correlation methods. They will usually also have had some exposure to introductory inference such as the ideas of confidence intervals and hypothesis testing.

However, it may be that someone who has received such training may need their ideas refreshing. It may also be the case that the ideas of statistical uncertainty and inference were not fully understood on the first exposure. Module 1 and Module 2 therefore provide selective introductions to quantitative research and data analysis, with a focus on key topics that will help to contextualise the ultimate task of learning about multilevel models.

If you have a good understanding of multiple regression – including the treatment of categorical explanatory variables and interaction effects – you may wish to skip to Module 4. Confident users of multiple regression, but who have not used MLwiN before, should start with the practical for Module 3.

More advanced modules will be published as they are written.

Preliminary reading

Modules 1 and 2 are designed to refresh the concepts, definitions and techniques of introductory quantitative data analysis. We expect that users will already have some familiarity with many basic ideas and have some prior experience of traditional elementary statistical analysis. More in-depth treatments of the material covered in Modules 1 and 2 can be found in the following on-line resources and texts.

Online resources

We have come across the following sites which we think are useful. Please note, however, that they are not connected with the Centre for Multilevel Modelling.

http://www.socialresearchmethods.net/kb/contents.php
This is the Research Methods Knowledge Base and is a comprehensive web-based text book that introduces all of the topics in a typical introductory course in social research methods, including quantitative analysis and statistical inference.

http://cast.massey.ac.nz/login.html
CAST includes several computer-based statistics textbooks that uses interactive diagrams to help teach all the statistical concepts. There are three introductory textbooks aimed at different application areas.

http://sportsci.org/resource/stats/
New original approaches to statistics for researchers: the examples are taken from exercise and sports science, but the principles apply to all empirical sciences: The site says "If you're new to stats, most of what you read here will be a new view. But even if you have done some stats, there's plenty here that's new. For example, I've discarded most details of computation, in the hope that you will get a better understanding of the concepts. Let's leave the computations to the computers."

http://davidmlane.com/hyperstat/index.html
Hyper Stat online; an online statistics book with links to other statistics resources on the web.

http://faculty.vassar.edu/lowry/webtext.html
Concepts and Applications of Inferential Statistics is a free, full-length, and occasionally interactive statistics textbook. It is a companion site of VassarStats, Web Site for Statistical Computation. The materials on this site may be freely used for any non-commercial educational purpose.

http://www.stats.gla.ac.uk/steps/home.html
The STEPS consortium has developed problem-based modules to support the teaching of Statistics in Biology, Business, Geography and Psychology. The software is freely available to educational institutions, and can be downloaded from the Web site.

http://www.anu.edu.au/nceph/surfstat/surfstat-home/surfstat.html
Surfstat.australia: an online text in introductory Statistics

Books

Diamond, I. and Jefferies, J. (2001) Beginning Statistics: An Introduction for Social Scientists. London: Sage.

Fielding, J. and Gilbert, N. (2006) Understanding Social Statistics. 2nd Edition. London: Sage.

Wright, D. (1996) Understanding Statistics: An Introduction for the Social Sciences . London: Sage.

Abselson, R.P. (1995) Statistics as Principled Argument. New York: Lawrence Erlbaum.

de Vaus, D. (2002) Analysing Social Data: 50 Key Problems in Data Analysis. London: Sage.

Erickson, B.H. and Nosanchuk, T.A. (1992) Understanding Data. Buckingham: Open University Press.

Robson, C. (2002) Real World Research: A Resource for Social Scientists and Practitioner-researchers. 2nd Edition. London: Blackwell.

Other useful multilevel modelling books

Are you an experienced quantitative researcher?

If you have a good understanding of multiple regression – including the treatment of categorical explanatory variables and interaction effects – you may wish to skip to Module 4. If you do not feel fully confident in the application and interpretation of multiple regression, we recommend that you at least answer the quiz questions for Module 3 to test your understanding.

If you are a confident user of multiple regression, but have not used MLwiN before, you should start with the practical for Module 3.

I have attended one of the Centre for Multilevel Modelling workshops. Do the online materials cover the same topics?

Modules 4 and 5 cover the same material as the first part of our workshops, although in more detail and using different examples. We therefore strongly recommend working through these modules after attending the workshop to consolidate what you learnt before attempting to analyse your own data.

Further modules will be added as they are developed. These will include additional methods-based materials (e.g. single-level and multilevel logistic regression), as well as applications (e.g. the use of multilevel models in school effectiveness research).

I am planning to attend one of the Centre for Multilevel Modelling workshops. Should I work through these materials first?

Our workshops assume an understanding and familiarity with the application of multiple regression. If you have not used multiple regression, or feel in need of a refresher, we strongly recommend that you work through Module 3 before the workshop. For those with less experience in statistical analysis, you may additionally find it helpful to study Modules 1 and 2.

The workshops do not assume any prior knowledge of multilevel modelling or MLwiN. The first part of the workshop will cover much of the same material as Modules 4 and 5, although in less detail and using different examples. To get the maximum benefit from the workshop, it would be worthwhile reading Module 4 and Module 5 (Concepts) beforehand. If time is short, Lessons 5.1 and 5.2 of Module 5 provide an introduction to multilevel modelling.

How can I test my understanding of the material?

You will find quiz questions interspersed throughout the modules to allow you to assess your understanding of the material.

Before starting to work through the materials, we strongly encourage you to test your current understanding of statistics by taking our prerequisites quiz. On the basis of your score in this quiz, a recommendation will be given on which module to start with.

I do not have a copy of MLwiN. How can I do the practical exercises?

You can use our LEMMA MLwiN training version or the latest version of MLwiN (MLwiN 2.02 is not compatible with the training materials). There are several options, please see our MLwiN download page for further details.

I do not use MLwiN. Will I still find the materials useful?

Yes, much of the material can be used without software, or by users of software other than MLwiN.

Modules 1, 2 and 4 do not have practical exercises, and are therefore not tied to any particular software package.

Modules 3 (Multiple Regression) and Module 5 (Introduction to Multilevel Modelling) each consist of two integrated components: concepts and practicals. In ‘concepts’, we describe the statistical models using illustrative examples from a range of disciplines; this is done without reference to any software. The ‘MLwiN practical’ component goes through, in detail, the analysis of a particular data set with MLwiN using the modelling techniques described in the Concepts component of the module.

Users of other software packages could use the 30-day evaluation version or free-to-UK academics versions of MLwiN to output data in SPSS or Stata format. In the future, we hope to extend our materials to include practical instructions for other packages.

Do I have to pay to access the materials?

No. The materials can be accessed free-of-charge and downloaded by anyone. However, we do require users to register with us first.

During registration, we collect:

your name and contact information, including e-mail address.
information regarding your personal or professional interests, statistical and academic background, demographics, and experience with our products and services.

We also collect information about your use of and progress through the course, to help us in our research, and to help our funders evaluate the course.

More information is available in "Why do I need to register?", and in our site's Privacy Statement.

Why do I need to register?

We require you to register mostly so we can collect data to help us to conduct and publish research into the learning of multilevel methodology and applications. We hope to study learning trajectories:

use of our course, and
progress (measured by Quiz scores) in our course

...and relate it to our users' statistical and academic background.

However, your personal data will be held and processed in strict confidence by Centre for Multilevel Modelling researchers. We uphold UK data protection laws, and much of our long professional experience has been in the analysis of sensitive educational and health data. Data will be highly aggregated before any publication, so that individual persons cannot be identified. We also collect course users' demographic, academic background, and contact details on behalf of our project funders – the ESRC (through the NCRM - National Centre for Research Methods ). We pass this data onto them so that they can evaluate our course, particularly assessing its uptake. They may also contact you in order to ask you for feedback. You have no obligation to respond to their requests.

More information is available in our Privacy Statement.

Can I use the materials in my own teaching?

YES!

Learning materials on our site are free to use under a Creative Commons licence. We will be delighted for you to use these materials for your own (non-commercial) teaching, and ask that you cite us if you do so.

Your citation should include:

the module's author
the module number and name
the part of the module (Concepts, MLwiN Practical or Quiz)
"LEMMA VLE", the name of the project deliverable this was funded by
"Centre for Multilevel Modelling"
a link to our web site

For example:

Steele, F. (2008) Module 3: Multiple Regression MLwiN Practical. LEMMA VLE, Centre for Multilevel Modelling. Accessed at http://www.cmm.bris.ac.uk/lemma/course/view.php?id=13.

Touch base, give feedback, get support

Do please let us know that you're using our materials (email: lemma-help@bristol.ac.uk) – we're keen to support teachers of Multilevel Modeling, and feedback is always welcome!

Adapt the materials

Please also feel that you can adapt our materials to suit your needs. We've designed in a split between Concepts (generic information) and Practicals (examples using specific software and datasets from particular disciplines), so that there's less that you'll have to change.

Let us know (email: lemma-help@bristol.ac.uk) which Practical you intend to adapt, and we'll send you its original Microsoft Word file if required.

...then share them here?

We hope that you will create different versions of the Practicals that we can put on the site - publicising your expertise, and drawing the attention of more people to it as the site becomes useful to wider communities. We hope that you'll make versions that use:

examples and data from other disciplines, and
instructions and output from other software.

To ensure consistency, we may require some changes to your version before we can put it up on the site, and we do not guarantee that we will be able to host your version on our site.

Who designed the system?

The learning materials are kept in a course management system called Moodle. Moodle is an open source community project, and so has been designed by many people.

The curriculum design and key requirements for the site were by Fiona Steele, supported by colleagues, including: Jon Rasbash, Kelvyn Jones, Harvey Goldstein, Aileen Earle.

The individual learning materials are by the named authors at the start of each Module, including Fiona Steele, Tony Fielding and Rebecca Pillinger, and Jon Rasbash. Materials were edited by Fiona Steele, and sub-edited by Rebecca Pillinger.

Hilary Browne was the web-developer - doing most of the visual design, all of the CSS and HTML hacking, and key parts of the interaction design.

Sacha Brostoff did most of the navigation design, interaction design and usability testing, and a little of the visual design.

Chris Charlton is the system administrator: doing all the php, database, software customisation and web-server magic that put the course on the internet, bent Moodle to our needs, and kept it working.

Contact us

Unfortunately we do not have resources to answer queries about the content of the materials.

However, go to our LEMMA-help enquiry form if you have questions about or problems with using the system, bug reports, or comments on materials.

Or email us on lemma-help@bristol.ac.uk to let us know you're teaching with them, or to let us know you would like to put materials on the site.

Why can't we answer questions about the content of the materials?

The materials are very detailed and designed for self-learning. In particular, you can test your understanding of the materials by answering the quiz questions in each lesson of a module. You can retake the quiz as many times as you need to, to understand.

If these materials don't work for you, it may be that brushing up your statistical foundations will help. There are links to prerequisites throughout the course. If this doesn't help, you may need some face-to-face time with someone knowledgeable in multilevel methods. There are workshops several times a year in the UK, and some of them are workshops given by us.

How do I get to the course?

Click:

After you have opened the page, you can add it to your web-browser's Favourites or Bookmarks.

When I try to open the worksheets for the LEMMA course I get an error message

Q: I'm taking the LEMMA Course and until yesterday everything was working perfectly, but today I just couldn't open the MLwiN Datafiles. Every time I try I get this message: "Run-time error '5': Invalid procedure call or argument". I've already install the latest version but it didn't work. I tried several times uninstalling and reinstalling the software, but it didn't work… Any ideas of what can I do? Many thanks in advance.

A: Has anything on your system changed between when you were able to open the worksheets and now? Examples might be other software being installed or removed. It might be worth manually clearing the MLwiN installation directory after removing it, before attempting to reinstall it. This is the directory where you chose to install it (usually C:\Program Files\MLwiN...). If you have an account with administrative privileges on that machine it might also be worth checking to see if you have the same problems when running as that user.
- As I don't have administrator rights in my computer, it is a pain each time I need to install a software. Finally I could open the files. As I realized I can open the old example files, what I did was to save the files in my computer and then change the extension from '.wsz' to '.ws' I know it is not the best, but it worked.

When I double click on one of the worksheets for the LEMMA course, it opens with notepad not MLwiN

When I click on the icon for 5.1.wsz in Module 5.1 I seem to get a notepad file that has a lot of blank characters, not a data file.

First note that the datasets for the LEMMA course will not open with MLwiN version 2.02: you will need the latest version, or the teaching version (which you can download for free here). Make sure that you have one of these versions of MLwiN. If you do have the appropriate version but are experiencing this problem, then first try reinstalling your copy of MLwiN. If this fails, you can try downloading again and then reinstalling. If that doesn't work or if you cannot do this, there are two options: a work around or a fix.

Option 1: Work around
When you click on the link to the datafile, choose Save to computer instead of Open. Then start MLwiN and open the data file from MLwiN's File menu.

Option 2: Fix
The fix is to reassociate the data file's .wsz extension with MLwiN (so that Windows opens them in MLwiN if you double click them, or if you choose Open instead of Save to computer when downloading), which is pretty easy and takes just a minute.
Again, when you click on the link to the datafile in your browser choose Save to computer rather than Open, and save for example to My Documents. Then open up My Documents or whichever folder you have saved the file into in Windows Explorer - so you can see the data file's icon. If the cause of the problem with opening the files is the situation this fix is designed for, then it should not have the MLwiN logo on it. (If the data file does have the MLwiN logo then there is probably a different cause of the problem and this fix will probably not work).
The next part of these instructions is written correctly for Windows XP, but probably something similar will work for Windows Vista. Right click on the data file (i.e. click with the right mouse button, not the left one that people usually use), and choose the menu option Open With, then in the submenu select Choose Program. You should get a new window with a list of programs and a tick box labelled "Always use the selected program to open this kind of file" WHICH YOU SHOULD TICK, then select MLwiN, then click OK. If MLwiN is not in the list of programs, click the Browse button and locate and open the "mlwin.exe" file. It would normally be in the folder C:\Program Files\MLwiN v2\ or C:\Program Files\MLwiN... (Note that the icon for the program should be pale blue and white, not yellow and dark blue: the program with the yellow and dark blue icon is version 2.02 which will not open the LEMMA datafiles).
After doing this, the data file on your computer should now have the MLwiN logo, and when you double click it should open up in MLwiN, and when you open files directly from your browser, they should open in MLwiN too.
- I didn't reinstall, but I tried the work around (didn't work), and the fix (which did).