Maria Pampaka holds a joint position, as a Senior Lecturer of Education and Social Statistics, at the University of Manchester. My substantive research interest lies within educational research and in particular in the association between teaching practices and students’ learning outcomes focusing on STEM and recently on issues around gender (in)equality. Methodologically, my expertise and interests lie within evaluation and measurement, and advanced quantitative methods, including longitudinal data analysis and dealing with missing data.

For more details:



Course content: 

The course will introduce longitudinal data analysis and provide students with the practical skills of performing longitudinal analysis with a focus on data management and analysing micro panel data.
The module will begin by introducing various longitudinal designs and discussing the advantages (and limitations) of panel data. We will then go into analytical considerations, starting from descriptive analysis and then regression techniques for modelling change in both continuous and binary outcomes. We will cover, among others, fixed and random effects models, hierarchical regression (multilevel modelling) and growth curve modelling. We will further explore and introduce modelling of event occurrence and duration, transitions as well as age, period and cohort analysis and related issues. One session will additionally look into more complex issues including accelerated longitudinal designs, establishing measurement invariance in longitudinal studies and resolving missing data issues. Every session will be accompanied by a practical computer-based session with analytical examples and exercises drawing on a variety of datasets.

The focus of the module is applied, but some mathematics will be used to formalise theoretical concepts.


Course objectives:

The course will introduce students to the methodological and analytical skills that will enable them to address questions about the measurement and explanation of change. By the end of the course the students should be able to:

  • Understand the concepts, designs and terms of longitudinal research
  • Understand and critically evaluate publications with applications of longitudinal data analysis
  • Apply a range of different techniques for longitudinal data analysis
  • Be able to choose a design, an appropriate mode and method of analysis for their own research questions

The course is appropriate for postgraduate level students with experience of regression models.


Course prerequisites:

  • Final year undergraduate level knowledge of linear regression methods (OLS regression)
  • Some experience of multilevel modelling and logistic regression techniques is also useful.
  • Familiarity and ideally some experience with using stata for analysis.

Day by day overview:


Day 1: Basics – Longitudinal Studies and Data 

  • Types of longitudinal data
  • Overview of longitudinal studies
  • Rationale for Longitudinal studies
  • Advantages and challenges.
  • Dimensions of change
  • Handling panel data/ data management [Practical] 


Day 2: Describing panel data 

  • Within and between variation
  • Transitions
  • Using weights
  • Graphs for trends and change over time


Day 3:  Modelling panel data – Introduction 

  • Overview of models depending on outcome (dependent) variables
  • Different types of independent variables
  • Modelling panel data: key notation
  • Causality
  • Challenges of panel data modelling (Statistical dependencies, Unobservables, Measurement error bias)


Day 4: Modelling continuous outcomes

  • Pooled Ordinary Least Square
  • Fixed and Random effects regression
  • Conditional and Marginal models
  • Comparing the different regression models
  • Multilevel models


Day 5: Modelling categorical outcomes 

  • Logistic Regression – Recap
  • Modelling Binary outcomes in Panel Data
  • – Within-group (FE) estimation
  • – Random effects estimation
  • – Choosing between FE and RE estimation


Day 6:  Modelling change

  • Analysis of change (using change scores and impact functions)
  • Dynamic models using lagged variables
  • Growth Curve Modelling (overview of SEM and multilevel approach)


Day 7: Exploring and Modeling event occurrence (and duration)

  • Analysing transitions and durations: Survival/Event History Analysis
  • Kaplan-Meier Estimates
  • Cox Proportional Hazards Model


Day 8: Modelling transitions and Age, Period, Cohort analysis 

  • Introduction to Multistate Modeling
  • Introduction to age, period, cohort analysis
  • Exploring age, period, cohort effects


Day 9: Further complexities with longitudinal designs

  • Accelerated Longitudinal Designs
  • Measurement over time: establishing measurement invariance
  • Non response and attrition: Multiple imputation for dealing with missing data


Day 10: Contemporary issues with Longitudinal Analysis 

  • Conceptual debates
  • Modelling debates
  • Student presentations