Course Overview
Social scientists often grapple with phenomena that change over time. Criminologists, for instance, analyse crime trends and examine how criminogenic factors change over months or years. Psychologists explore the development of attitudes during childhood and adolescence, investigating how early experiences shape adult outcomes. Political scientists evaluate public policies, assessing whether they achieved their intended goals and how they affected different populations. What unites these diverse examples is the central role of time. When data are collected over both units (e.g., survey respondents, cities, pupils) and time (e.g., waves of data collection, months, years) — i.e., longitudinal data — they provide an opportunity to address fundamental questions in the social sciences. Panel data allow researchers to investigate temporal ordering, changes over time, developmental trajectories, reciprocal relationships, and sometimes even causal effects.
It will give me the confidence to know what options are out there for methods to help me conduct my research (attendee, 2025).
The course This course offers a comprehensive introduction to the analysis of longitudinal data. Different disciplines within the social sciences have developed distinct analytical frameworks to analyse panel data, and this course will provide an overview of some of these strategies. The focus will be on three core areas:
- Estimating latent trajectories to analyse change over time
- Addressing ‘reverse causality’ and modelling reciprocal relationships
- Leveraging panel data for causal inference within the potential outcomes framework.
Specific topics include an introduction to longitudinal data, multilevel models for change, latent growth curve models, cross-lagged panel models, difference-in-differences, and marginal structural models. For each topic, participants will learn the foundational principles and explore recent advancements in the field. The course adopts a hands-on approach, enabling participants to analyse panel data using R. Hands-on exercises and real-world applications will help solidify understanding of these methods. By the end of the course, participants will be well-equipped to analyse and interpret models using longitudinal data in their own research projects.
This was an amazing opportunity to learn from a leading expert (attendee, 2025).
Course Lead
Thiago R. Oliveira is a lecturer in quantitative criminology in the criminology department at the University of Manchester. He has a PhD in Social Research Methods from the London School of Economics and Political Science and has previously held positions at the University of Oxford and the University of Surrey. As a quantitative social scientist, he uses methods from data science, econometrics, and epidemiology to investigate policing's sometimes conflicting objectives of crime deterrence and public legitimacy. His research mostly draws on survey data and spatial data, and he is particularly interested in longitudinal data analysis and causal inference.
Course Objectives
During this course, attendees will:
- Acquire understanding of concepts, designs, and methods for longitudinal data analysis
- Use R to analyse longitudinal data and address a range of research questions that involve change over time
- Critically reflect upon the role of time in social science research
- Evaluate and fit models estimating latent trajectories under a multilevel and an SEM frameworks
- Understand the potentials and limitations of cross-lagged panel models in handling reverse causality
- Engage critically with the challenges and possibilities of leveraging panel data to address causal questions under the potential outcomes framework
Is there any preparatory work or pre-requisites?
No prior knowledge of longitudinal, multilevel, or causal inference methods is assumed for this course, but participants should have a solid understanding of linear regression models as a foundation for the topics covered. While prior experience with R is not strictly required, it is highly recommended to facilitate engagement with the course material. For those less familiar with R, please see the linked supplementary materials which introduce the basics of R and will help you to prepare for course attendance. Supplementary Materials For those not familiar with R
- Wickham, Hadley, Mine Çetinkaya-Rundel, and Garrett Grolemund. R for Data Science. 2nd edition. https://r4ds.hadley.nz
- Dauber, Daniel. R for Non-Programmers: A Guide for Social Scientists. https://r4np.com
For those looking for introductory textbooks on quantitative social science
- Imai, Kosuke. 2017. Quantitative Social Science: An Introduction. Princeton University Press.
For those looking to get ahead on the content of the course: Latent growth curve models under the multilevel modelling framework
- Steele, Fiona. 2008. "Multilevel models for longitudinal data." Journal of the Royal Statistical Society Series A: Statistics in Society 171 (1): 5-19.
- Curran, Patrick J. 2003. "Have multilevel models been structural equation models all along?." Multivariate Behavioral Research 38 (4): 529-569.
Latent growth curve models under the Structural Equation Modelling framework
- Bollen, Kenneth and Patrick Curran. 2008. Latent curve models: A structural equation perspective. Wiley.
- Bianconcini, Silvia, and Kenneth A. Bollen. 2018. "The latent variable-autoregressive latent trajectory model: A general framework for longitudinal data analysis." Structural Equation Modeling: A Multidisciplinary Journal 25(5): 791-808.
Cross-lagged panel models
- Allison, Paul D., Richard Williams, and Enrique Moral-Benito. 2017. "Maximum likelihood for cross-lagged panel models with fixed effects." Socius 3.
- Hamaker, Ellen L., Rebecca M. Kuiper, and Raoul PPP Grasman. 2015 "A critique of the cross-lagged panel model." Psychological methods 20 (1).
Difference-in-differences
- Imai, Kosuke, and In Song Kim. 2019. "When should we use unit fixed effects regression models for causal inference with longitudinal data?" American Journal of Political Science 63(2): 467-490.
- Imai, Kosuke, In Song Kim, and Erik H. Wang. 2023. "Matching methods for causal inference with time‐series cross‐sectional data." American Journal of Political Science 67(2): 587-605.
- Callaway, Brantly, and Pedro HC Sant’Anna. 2021. "Difference-in-differences with multiple time periods." Journal of Econometrics 225(2): 200-230.
Marginal structural models
- Robins, James M., Miguel Ángel Hernán, and Babette Brumback. 2000. “Marginal structural models and causal inference in epidemiology.” Epidemiology 11(5): 550-650.
- Xiang Zhou, and George Wodtke. 2020. “Residual balancing: A method for constructing weights for marginal structural models.” Political Analysis 28(4): 487-506.
Who should attend?
While PGRs and Early Career Researchers from across all disciplines are likely to be the main audience, the course may also be of interest to senior scholars, practice-based researchers, independent researchers, and scholars who are seeking to refresh and further develop their data analysis skills.
Course Timetable
This course will take place in-person Monday 6 July - Friday 10 July
1:30 - 5:00pm Monday
Introduction to longitudinal data Introduction to R programming
9:00 - 12:30pm Tuesday
Growth Curve Models: Introduction to multilevel models for longitudinal data Lab session: estimating growth curve models in R
1:30 - 5:00pm Tuesday
Latent Growth Curve Models: A Structural Equation Modelling approach Lab session: estimating latent growth curve models in R
9:00 - 12:30pm Wednesday
Cross-lagged panel models Lab session: estimating cross-lagged panel models in R
1:30 - 5:00pm Wednesday
Recent advancements in cross-lagged panel models: Random Intercepts CLPM and CLPM with fixed effects Lab session: estimating RI-CLPM and CLPM-FE in R
9:00 - 12:30pm Thursday
Introduction to causal inference with panel data Lab session: parallel trends and two-period difference-in-differences in R
1:30 - 5:00pm Thursday
Difference-in-differences estimators Lab session: estimating two-way fixed effects, staggered DiD, and matching with DiD in R
9:00 - 12:30pm Friday
Marginal structural models Lab session: calculating weights for MSM using inverse probability treatment weighting and residual balancing in R
What's included in the course?
Each full day includes a vegan buffet lunch served 12.30-1.30pm. There are morning and afternoon refreshment breaks with tea, coffee, water, and pastries/cakes.
The course includes a social programme - these are optional but free social events for everyone attending our summer school to meet attendees from other courses in a relaxed environment.
Accommodation and travel are not included in the course price. You will need to arrange any accommodation and travel separately.
Cost
- Full price: £900
- PGR/Reduced Rate: £600
As well as PGRs, reduced fees are available to those working within the voluntary, charity and community sector. We also have two bursary options available for those entitled to reduced fees. Please view more information on our main Summer School website to find out more and how to apply.
Book Your Place
Please purchase via our online store below before 15 June (payment by card only). If you any questions, or will have trouble purchasing by this date please get in touch with methods@manchester.ac.uk.
Any questions, please do not hesitate to contact us on methods@manchester.ac.uk
Credits:
Created with an image by Chaosamran_Studio - "Financial data document graph chart report statistic marketing research development planning management strategy analysis accounting. Financial business technology hologram concept. 3d rendering."