Panel Data Analysis using Stata 7

Internet Course

Contents

Course description
Course Organization
Course Programme
Request an Enrolment Form Now

Timberlake Consultants Ltd (TCL), a distributor of Stata, responding to the increasing demand for training in the areas of statistical methodology and critical appraisal, are organising a set of long distance learning courses. The courses are conducted in English.

TCL invite you to participate in the Internet course Panel Data Analysis using Stata 7.

The Principal Lecturer The principal lecture is:

Andrew Pickles, Professor of Epidemiological and Social Statistics, University of Manchester

The Course - The objective of the course is to provide participants with knowledge of both the statistical tools essential for empirical analysis of panel data and provides an opportunity for hands-on statistical analyses using the software package Stata.
The course will therefore:

  • Review the basic statistical concepts relevant to empirical data analysis
  • Provide a practical and systematic approach to statistical testing and analysis using medical and economic applications
  • Provide “hands-on” experience in the use of the software package Stata to analyse medical and economic data
  • Offer participants the opportunity to discuss their own application(s) with experienced researchers
  • Allow participants to follow the course at their own place, by selecting the time-table best suited to their individual schedules

Organisation of the Course – The course takes place over the Internet. Course participants are part of a closed e-mail list and receive course materials through this list. The list may also serve as a forum for discussion of ideas and problem solving. The course is scheduled to take place over 4 sessions. At the beginning of each session, participants receive the relevant material, in addition to answers to exercises from the previous session. A period of 1 week will be allowed between sessions (2 weeks between Sessions 2 and 3). During these breaks, participants are expected to go over the course materials and work through the exercises. Discussions between participants and instructor will also take place during these periods. One week is allowed after session 4 to provide participants with the ability to clarify any questions arising from this or the previous sessions.

Who should attend - The course, given in English, is aimed at

  • Social Scientists, Medical Researchers and economists working with longitudinal studies
  • Market Researchers

Cost - The cost of the course are:

£ 150 $ 225 Euros 240

The cost includes course materials and are inclusive of taxes. The number of delegates is restricted. Please register early to guarantee your place.

Agenda
(may be subject to change)

Session I: Introduction to Statistics and Change

Basic Statistics: mean, standard deviation, standard error

  • T-tests and regression
  • Regression models: intercepts, slopes, model formulae, residual plots and heteroscedasticity
  • Generalized Linear Models
  • Simple methods for analysing change: two-period panel data: baseline and outcome data from a randomised trial
    - cross-sectional estimate of treatment effect
    - longitudinal estimate of treatment effect
    - change score regression and analysis of covariance

Session II: Simple Subject Effects

  • Omitted Variables
  • Reshaping Data
  • The xtreg procedure
    - between-subjects estimator
    - within-subjects/fixed/conditional estimator
    - GLS random effects estimator
    - Hausman test
  • Residuals
  • xt for more complex panel structures
    - xtdes
    - xtsum
    - xttab
  • Applications of xtreg to more complex panel structures
  • Missing data: MCAR, MAR and non-ignorable missing data

Session III: More Complex Subject Effects and Robust Inference

  • Subject Effects
    - independent, exchangeable, autoregressive and unstructured correlations
    - robust parameter covariance estimation and ‘working correlation matrices’
    - robustness and missing data assumptions
  • Random Coefficient Models


Session IV: Methods for Discrete Responses

  • Subject Specific Models
    - within subjects using McNemar and conditional logistic regression.
    - Maximum Likelihood random effects using integrated likelihood
    - xtlogit, xtprobit, xtpois etc
    - GLLAMM Generalized Linear Latent and Mixed Models
  • Population Average Methods
    - Estimation using generalized estimating equations; xtgee

Top