GESIS Training Courses

Scientific Coordination

Alisa Remizova

Administrative Coordination

Janina Götsche

Time Series Analysis for Modeling Intensive Longitudinal Data

Cologne / Unter Sachsenhausen 6-8
General Topics:
Course Level:
Software used:
Students: 220€
Academics: 330 €
Commercial: 660 €
Additional links
Lecturer(s): Noémi Schuurman

About the lecturer - Noémi Schuurman

Course description

In this course, participants will be introduced to time series analysis for modeling intensive longitudinal data in the context of the social sciences. Intensive longitudinal data are data that consist of many repeated measures (e.g., at least 25) for one or multiple cases (e.g., persons) and variables. In the context of psychology, these data are often collected through experience sampling or other forms of ecological momentary assessment. With these assessment techniques, participants are measured repeatedly throughout their daily lives, for example, through randomly prompted smartphone surveys multiple times a day. Participants report on, for example, their feelings, cognitions, and activities at the moment. During the workshop, we will discuss how the intensive longitudinal data can be analyzed with time series models, as well as the challenges and considerations in applying these techniques to the data. These techniques focus on modeling the dynamics in intensive longitudinal data and are suitable for research questions such as “Are these variables associated with themselves and each other (from variable a to b, b to a, or in both directions), over time?” On the first day, we will discuss models for single case analyses (e.g., modeling a single person at a time). On the second day, we will extend these techniques such that one can analyze multiple cases (e.g., persons) with multilevel models. These multilevel models will be estimated with Bayesian techniques using R. Mplus codes will also be available for interested participants. 
Organizational structure of the course
The workshop will include lectures and lab sessions carried out individually or in small groups and consisting of hands-on exercises with provided cases. During the workshop, you are encouraged to ask questions, discuss examples of modeling intensive longitudinal data you have seen, and bring your research problems and questions.

Target group

Participants will find the course useful if:
  • They want to take the first steps in learning how to model intensive longitudinal data for the social sciences, for example, data collected with experience sampling techniques.

  • Learning objectives

    By the end of the course, participants will:
  • Attain basic knowledge of the basics of time series analysis and intensive longitudinal data, including basic descriptives concepts such as stationarity, ARIMA models, dynamic networks;
  • Attain a thorough understanding of the interpretation of univariate and multivariate single-subject autoregressive models applied to intensive longitudinal data in the social and behavioral sciences, including important caveats;
  • Be aware of the within/between problem when modeling multiple subject data, techniques to counter the problem, related important concepts such as ergodicity;
  • Attain a thorough understanding of multilevel extensions of univariate and multivariate autoregressive models applied to intensive longitudinal data in the social and behavioral sciences;
  • Get practical experience with the above approaches in R.

  • Prerequisites

  • A solid grasp of linear regression modeling;
  • A solid grasp of basic hypothesis testing;
  • Experience with working in R (e.g., importing and manipulating data, performing analyses);
  • Knowledge of Bayesian modeling will help get the most out of this course but is not required.
  • To refresh or gain knowledge of R or Bayesian modeling, you can attend an online introductory workshop on R or an online introductory workshop on Bayesian statistics that GESIS offers before the current workshop.
    Software and hardware requirements
    The lab meetings of this workshop consist of exercises in R There will also be Mplus codes/exercises for most of the models available for those interested, but please note that GESIS does not provide an Mplus license. You are free to use this software if you have it.
  • R is free and open source.
  • Mplus is not free and requires a license. Mplus is typically considered a bit more user-friendly, particularly if you are not already familiar with R. Mplus uses more defaults that hide away more complicated options/decisions when we move to more complicated models, in particular, the ones that require Bayesian analysis.
    To prepare to use either one or both software programs, we recommend doing the following before the workshop starts.

    For R:
    Install R and Rstudio and JAGS (needed for Bayesian analyses) before coming to the course;
    For Mplus:
    If you have a license for Mplus v8.1 or higher, you can use it for the alternative Mplus exercises.