Matching Your Research Question to the Right ESM Design

Experience Sampling Method (ESM)—also known as Ecological Momentary Assessment (EMA)—offers a way to collect snapshots of people’s lives as they unfold. Participants complete short surveys on their phones, often several times a day, across days or weeks. The result is intensive longitudinal data: fine-grained insights into how feelings, thoughts, behaviors, and other experiences change in real time (Trull & Ebner-Priemer, 2014; Hamaker & Wichers, 2017).

But not every research question calls for the same type of design. To get the most out of ESM, we need to match our question to the right approach.

Three Types of ESM Designs

1. Between-person: who differs from whom?

Suppose we ask: Do people who are generally happier than others also tend to be more productive than others?

This is a between-person question, where the focus is on comparing averages across individuals. It can be answered with:

• Cross-sectional surveys (many people, one time point),

• Panel or cohort studies (many people, a few time points),

• Or ESM (many people, many time points).

ESM isn’t strictly necessary here, but repeated measurements across time can sharpen estimates and reveal whether the relationship holds consistently (Voelkle et al., 2014; Schuurman, 2023).

2. Within-person: how does one person change?

Now imagine we ask: When a person feels happier than usual, are they also more productive than usual?

This is a within-person question, where we examine fluctuations inside an individual over time. Repeated measures are essential, and ESM is particularly well suited because it captures moment-to-moment changes as they happen (Hamaker, 2012).

Importantly, between-person and within-person findings often diverge. People who are generally happier than others may not show the same day-to-day link between their own happiness and productivity. That’s why it’s crucial to match design to question.

3. N-of-1: zooming in on individuals

Sometimes the focus is a single person. For example: When Alex feels unusually happy in the morning, does he tend to be more productive that same afternoon?

This is an N-of-1 design, in which repeated observations from one individual reveal their unique dynamics. With enough data, we can provide feedback tailored to that person’s patterns. N-of-1 approaches are especially powerful in clinical, organizational, or coaching contexts where personalized insights are the goal (Molenaar, 2004; Wright & Woods, 2020).

The trade-off is that power comes not from sample size but from the number of observations collected for that person.

What only ESM can capture

Because ESM gathers dense, real-time data, it can address questions that sparse designs miss:

• How long does a boost in happiness last before productivity returns to baseline?

• Does happiness tend to rise before productivity, or does productivity boost happiness later in the day?

• Are productivity levels relatively stable hour to hour, or highly variable?

These are questions about timing and dynamics—and they require the fine-grained structure of ESM.

Matching design to question

When planning a study, the design should follow the claim:

• If the goal is to understand who differs from whom, a cross-sectional or panel design may suffice, though ESM can add precision.

• If the goal is to understand how a person changes relative to themselves, repeated measures are required, and ESM is ideal.

• If the goal is to provide personalized insight, N-of-1 ESM designs are the right tool, provided there are enough repeated observations to detect patterns.

Avoiding common pitfalls

A frequent mistake is to interpret between-person results as if they reflected within-person processes. Finding that happier people report higher productivity does not prove that this person’s happiness at a given moment predicts their own productivity. To make such claims, we need within-person data.

Another challenge is under-sampling. If the process shifts within hours, sampling once per evening will miss the dynamics. Piloting schedules and checking compliance help ensure that the timing is right.

Starting Points for Deeper Reading

• Hamaker, E. L. (2012). Why researchers should think “within-person”: A paradigmatic rationale. In M. R. Mehl & T. S. Conner (Eds.), Handbook of Research Methods for Studying Daily Life (pp. 43–61). Guilford Press.

• Molenaar, P. C. (2004). A manifesto on psychology as idiographic science: Bringing the person back into scientific psychology, this time forever. Measurement, 2(4), 201–218. https://doi.org/10.1207/s15366359mea0204_1

• Schuurman, N. K. (2023). A “within/between problem” primer: About (not) separating within-person variance and between-person variance in psychology. PsyArXiv. https://doi.org/10.31234/osf.io/7zgkx

• Wright, A. G. C., & Woods, W. C. (2020). Personalized models of psychopathology. Annual Review of Clinical Psychology, 16, 49–74. https://doi.org/10.1146/annurev-clinpsy-102419-125032