Cases
This article is about defining your cases of interest, that is, individual entities of a population from which you collect data for your study.
Depending on your research question, a case can be, for example, a person, dyad, household, or company. All cases together make up the population of interest. Generally, you draw a sample of multiple cases to make inferences about the population.
It is essential to clearly define what constitutes a case, as this underpins both the data collection process and the scope of any subsequent inferences.
In this article, you will find: 1) single case; 2) nested cases; and 3) multivariate versus nested cases—in the context of intensive longitudinal data.
1 Single case
In single-case designs, the individual case is treated as the primary focus of analysis using an [idiographic approach]. This often involves collecting many repeated measurements over time for that single case. The emphasis is on capturing dynamic, within-case variation—how processes unfold, interact, and change within one case across time, rather than comparing processes across individuals. The aim of the study is to make personalized inferences about a single case, which may be through observational or experimental designs.
Michael aims to detect early warning signs of relapse in a patient suffering from depression. Michael conducts an intensive longitudinal study using smartphone-based ecological momentary assessments (EMAs) and passive sensing data. Over the course of three months, the patient completes short daily surveys on mood, energy levels, and social activity, while the phone continuously collects data on sleep patterns, mobility, and communication frequency. Michael then uses time-series modeling and network analysis to identify dynamic patterns and potential early warning signals—such as increased mood variability or social withdrawal—that may precede a depressive relapse.
2 Nested cases
Cases are nested when individual cases (e.g., persons) are grouped within higher-level cases (e.g., a household), forming a multilevel structure. These structures are not merely organizational but often carry substantive meaning, as higher-level processes can influence processes occurring at the lower level, and vice versa. A typical intensive longitudinal design in Psychology involves observing a group of individuals over time. This allows researchers to apply multilevel analyses, which may provide insights into both within-person and between-person variation.
Carmen is interested in understanding why some individuals suffering from depression are more resilient than others. She designs a study that combines baseline questionnaires on emotion regulation, social support, and cognitive flexibility with ecological momentary assessments (EMAs) over a two-month period. Participants report on their daily mood, stressors, and coping strategies via a mobile app. Carmen then applies multilevel modeling to examine how within-person fluctuations relate to resilience, while also exploring between-person differences in the baseline trait-level variables. Her goal is to identify both stable characteristics and dynamic processes that distinguish more resilient individuals from those who struggle to recover.
3 Multivariate versus nested cases
It is also possible to observe multiple, multivariate outcomes of individual cases. Depending on the research goals, you may view them as multivariate outcomes of a single case, or as univariate outcomes of nested cases. For example, when studying the mood of two individuals that make up a dyad, the observed mood scores of both individuals can be the outcomes of the dyad, or the observed mood scores can be outcomes of single individuals nested in a dyad. The correct framing of the cases depends on your theoretical interests and analytical goals. In the single-case framing, the focus may be on the interaction between the individuals of the dyad which can be modeled using a single-level approach. The nested framing suggests a multilevel approach, which focuses on the variance decomposition within and between individuals, but does not incorporate interactions between individuals.
Sam wants to study how emotion patterns between adolescents and their parents influence each other. They conduct an intensive longitudinal study in which adolescents and their parents complete brief surveys three times a day for four weeks, reporting on their emotional states. Sam aims to understand how emotional contagion and regulation unfold within the family system throughout the day.
Kai wants to study how mood in a household with adolescents and their parents varies within the family context. They conduct an intensive longitudinal study in which adolescents and their parents complete brief surveys three times a day for four weeks. Using a multilevel modeling approach, Kai examines how much variance in average mood and emotion regulation is explained by household in relation to within-person variance, while also considering whether between-household differences, such as parenting style, explain differences in average mood and emotion regulation.
4 Takeaway
Clarifying what constitutes a case—whether single, nested, or multivariate—is essential for aligning your measurement, theory, and analysis.
5 Further reading
Citation
@online{berkhout2025,
author = {Berkhout, Sophie W.},
title = {Cases},
date = {2025-04-24},
langid = {en}
}