Longitudinal data
Longitudinal research - in which the same people, companies, countries, or other units are measured at two or more time points - provide unique opportunities to investigate stability and change over time. It is therefore a valuable tool in many fields in the social and behavioural sciences.
Longitudinal research can be (roughly) divided into two categories: panel research, and intensive longitudinal research.
Panel research
Panel research consists of a relatively small number of repeated measures (say between 2 and 10), typically with relatively large intervals between the measurements. This is a more traditional form of longitudinal research, and is typical analyses of such data include latent growth curve modeling, latent transition modeling, and cross-lagged panel modeling.
Examples of our contributions to panel research:
Intensive longitudinal research
Intensive longitudinal research consists of many repeated measures (say more than 20, up to thousands), relatively dense in time. This is a more novel research approach, based on new measurement techniques such as experience sampling (ESM), ambulatory assessments, daily diaries, and ecological momentary assessments, which are based on using smart phones, activity trackers, and other wearable devices to capture life as it is lived. The vast amounts of data that result from this require the development of new statistical techniques.
Examples of our contributions to intensive longitudinal research:
- Studying the dynamics of processes
- R-package for multilevel hidden Markov models
- Power analysis for the RI-CLPM
- R-Package for the hysteretic threshold autoregressive model
Funding
- ERC consolidator grant (Ellen Hamaker)
- ZonMw grant (Rens van de Schoot)