MEAFA workshop on Multilevel/mixed Models Using Stata, 10 - 12 February 2010

Announcements

13 Jan 2010: In order to satisfy the high demand, we created an identical workshop that will run during 8-10 Feb 2010. The content and presenters for this workshop are the exactly same (as described below), as follows: Mon 08 Feb: Working Efficiently with Stata (Demetris Christodoulou), Tue 09 - Wed 10 Feb: Mixed models (Bobby Gutierrez). This workshop is now open for reservations.

11 Dec 2009: The demand for the workshop has been overwhelming and all places have been reserved within three days of the initial announcement.

Brief description of multilevel/mixed models

Multilevel/mixed models contain both fixed effects analogous to the coefficients in standard regression models and random effects not directly estimated but instead summarized through the unique elements of their variance-covariance matrix. Mixed models may contain more than one level of nested random effects, and hence these models are also referred to as multilevel or hierarchical models, particularly in the social sciences. Stata&rsquo s approach to linear mixed models is to assign random effects to independent panels where a hierarchy of nested panels can be defined for handling nested random effects.

Description of the 3-day workshop

The first day of the workshop provides an overall introduction to Stata 11 and demonstrates how to work efficiently with reproducible and tractable routines. The first day is of interest to those who are not familiar with Stata or want become more efficient in their work. The next two days engage with the theory and application of multilevel/mixed models.The course will be interactive, use real data, and offer ample opportunity for specific research questions and for working exercises to enforce what is learned. Detailed notes will be provided outlining all theory and applications.

  • Day 1 (Feb 10): Working efficiently with Stata 11
    This day assumes no previous knowledge of Stata. It will introduce the environment of Stata 11 and ways to customise Stata. It will discuss basic data management techniques, how to use logs and work with reproducible and tractable do-files, running basic statistical and estimation commands, producing tables and making graphs.
    Presenter: Demetris Christodoulou, MEAFA General Convenor.

  • Days 2 and 3 (Feb 11-12): Multilevel/mixed Models Using Stata
    The course on mixed models assumes basic knowledge of Stata and of standard linear regression. The first part will look at the classic random-intercept linear model. Several approaches will be discussed for fitting this model, along with the associated benefits and assumptions of each approach. The second part will focus on random coefficients and the various covariance structures that can be imposed with multiple random-effects terms. The theme of the third part can best be described as tricks of the trade, covering various methods for fitting more complex models including crossed-effects models, growth curve models, and models with complex and grouped constraints on covariance structures. The fourth part will consist of predictions, model diagnostics, and other postestimation tasks. During the first four parts, the discussion will be confined to linear mixed models for continuous responses. The fifth part will focus on models for other types of responses in particular binary and count responses. During this part of the course, you will learn that most of what is discussed for linear mixed models can be applied equally to mixed models with noncontinuous responses.
    Presenter: Roberto G. Gutierrez., StataCorp Director of Statistics

Enrollment and Fees

Those interested only in learning how to work efficiently with Stata 11 may attend only the first day. Those interested only in mixed models and already have basic knowledge of Stata may attend only the last two days. Those interested in mixed models but do not have experience with Stata should attend all three days.

Fees vary on the days attended (prices exclude GST):

  • Attend only Day 1 on Working efficiently with Stata 11: $500
  • Attend only Days 2 & 3 on Multilevel/mixed Models Using Stata: $1300
  • Anttend all three days: $1600

Fees include extensive course material, a detailed guide to Stata 11, data sets, lectures, use of computing facilities, temporary use of Stata 11 licenses, full catering and opportunity to network with fellow researchers.

Numbers are limited and places are reserved on a first-come first-served basis, and to avoid disappointement and secure a place you must fill-in the Reservation Form. Successful attendees will be notified shortly after reservation and invoices will be issued accordingly. MEAFA maintains a no refund policy following payment. For more information on enrollment and fees contact meafa@econ.usyd.edu.au.

Discounts

You may qualify for one of the following discounts:

  • 40% discount for a restricted number of non-employed full-time PhD students
  • 20% discount for additional attendees from the same organisation or academic institution

Venue

The workshop will take place at the The University of Sydney Business School computer labs, Building H69 ground floor, cnr Codrington & Rose streets, the University of Sydney (see interactive map). You do not need to bring your own laptop. PCs and Stata 11 licenses for Microsoft Windows will be provided.

Programme

Day 1: Monday 8 Feb & Wednesday 10 Feb 2010
Working efficiently with Stata 11

08:40

Welcome tea and coffee

09:00-10:30

Introduction to Stata 11 Environment

Stata environemnt; configuration and special features; updates; personalised system; directory management; obtain help and perform search; online sources; Stata syntax

10:30-10:45

Morning break

10:45-12:15

Data Formats and Data Handling

Import, export, load and save datasets; simulation; review and document the dataset; ordering of dataset; display format

12:15-13:15

Lunch

13:15-14:45

Data Structures and Types of Variables

Categorical data; continuous data; append and merge other datasets; numerical, string and date/time variables; missing data; generate variables; dummy variables; special purpose variables

14:45-15:00

Afternoon break

15:00-16:30

Data Management and Output Management

logs for output; do-files; basic statistical and estimation commands; tables and graphs; stored results

16:30-17:00

User-Specific Questions


Day 2: Tuesday 10 Feb & Thursday 11 Feb 2010
Multilevel/mixed Models Using Stata, Part A

08:40

Welcome tea and coffee

09:00-10:30

What constitutes a linear mixed model?

Introduction to linear mixed models; the random-intercept model; the within estimator versus the GLS estimator; the Hausman test

10:30-10:45

Morning break

10:45-12:15

MLE, xtmixed, xtreg

Maximum likelihood and restricted maximum likelihood; using xtmixed and xtreg for the random-intercept model;

12:15-13:15

Lunch

13:15-14:45

The random coefficient model

Adding random coefficients; specifying models hierarchically; covariance structures for random effects; growth curves

14:45-15:00

Afternoon break

15:00-16:30

Random coefficient problems and introduction to multilevel models

Linear transformations of covariates in a random-effects setting; likelihood ratio (LR) tests; introduction to multiple-level models

16:30-17:00

User-Specific Questions


Day 3: Wednesday 10 Feb & Friday 12 Feb 2010
Multilevel/mixed Models Using Stata, Part B

08:40

Welcome tea and coffee

09:00-10:30

Multilevel models

Multilevel crossed-effects models; using Stata&rsquo s &ldquo R.&rdquo factor notation for mixed models; complex and grouped constraints on variance components; heteroskedastic residual errors; alternate residual-error structures

10:30-10:45

Morning break

10:45-12:15

Creating Programs

Creating a program from a do-file; when to use a program; including a program in a do-file; the command clear and its use in programming; creating an ado file from a program

12:15-13:15

Lunch

13:15-14:45

Best linear unbiased predictions

Best linear unbiased predictions (BLUPs); residuals; fit diagnostics; diagnostic plots; Cataloging and comparing mixed-models results in Stata

14:45-15:00

Afternoon break

15:00-16:30

Binary and count responses

Straighforward extension to multilevel/mixed models with binary and count responses; estimation via adaptive Gaussian quadrature; model building using the Laplacian approximation; predictions and other postestimation tasks

16:30-17:00

User-Specific Questions