ESM stands for the experience sampling method, a research technique where participants report what they’re feeling, doing, or thinking in real time throughout their day, rather than trying to remember it later. Instead of filling out a single survey at a doctor’s office asking “how have you felt this past week?”, ESM sends repeated prompts to your phone, asking you to describe your experience right now, in whatever setting you happen to be in.
How ESM Works in Practice
The basic idea is simple: a smartphone app sends you notifications at various points during the day, and each time you answer a short set of questions about your current state. You might rate your mood, note what you’re doing, log a symptom, or describe your surroundings. These check-ins are brief, often just a few questions, but they happen multiple times a day over days or weeks. The result is a detailed, real-time picture of someone’s lived experience that no single lab visit or end-of-week questionnaire could capture.
Researchers design ESM studies using different prompting strategies depending on what they want to learn. The most common is random time sampling, where notifications arrive at unpredictable moments throughout the day. This randomness is intentional: if prompts arrive at the same time every day, people start anticipating them and may change their behavior, like postponing a shower or skipping a trip to the store. Random signals catch people in the middle of ordinary life. A second approach, event sampling, asks participants to complete a survey whenever a specific event occurs, like a panic attack, a pain flare, or a conflict with someone. Researchers also consider factors like how many prompts per day, how many days the study lasts, and whether to add a separate morning questionnaire about sleep quality.
Why Real-Time Data Matters
Traditional surveys ask people to look back and summarize their experiences. “How anxious have you been this month?” or “How many times did you leave the house last week?” These kinds of questions seem straightforward, but they’re surprisingly unreliable. Human memory doesn’t average experiences evenly. Instead, people’s recollections tend to be dominated by the most intense moments (the peak) or the most recent ones (the end). If you had one terrible day in an otherwise decent week, your summary of the whole week will skew negative.
ESM sidesteps this problem entirely. Because each data point captures what’s happening right now, there’s no memory to distort. Researchers get a record of typical, everyday moments rather than a highlight reel of extremes. The data also has stronger ecological validity, a term that means the findings actually reflect real life rather than what happens under artificial lab or clinic conditions. When someone reports their mood while sitting in traffic or eating lunch with coworkers, that’s a fundamentally different kind of data than what you’d get from asking them to recall those moments a week later in a clinical office.
ESM vs. Ecological Momentary Assessment
You’ll often see ESM used interchangeably with another term: ecological momentary assessment, or EMA. The two methods share the same core principles, collecting data in real time, in real-world settings, with multiple assessments over time. But they come from different research traditions. ESM originated in psychology research in the 1980s, pioneered by Mihaly Csikszentmihalyi and Reed Larson, and focused primarily on self-reported subjective experiences like mood and attention. EMA, defined by Arthur Stone and Saul Shiffman in 1994, cast a wider net. It includes not just self-reports but also physiological monitoring (heart rate, breathing patterns) and behaviors tracked automatically by devices.
In practice, most researchers today treat ESM and EMA as overlapping terms. Some papers use “ESM/EMA” as a single umbrella label for any method that collects multiple real-time measurements in daily life.
Where ESM Is Used
Mental health research has been one of the biggest adopters of ESM. Conditions like depression, anxiety, and borderline personality disorder involve symptoms that fluctuate throughout the day and across situations, patterns that a once-a-week therapy check-in can easily miss. ESM lets researchers (and increasingly clinicians) see how someone’s mood shifts in response to social interactions, sleep, stress, or time of day. Studies can track event-specific triggers, like recording details every time a person experiences a panic attack or hears voices, building a granular map of when and where symptoms emerge.
Chronic pain is another area where ESM has proven valuable. A platform called SOMAScience, for example, uses an ESM-based smartphone app to capture momentary ratings of pain intensity, unpleasantness, interference with activities, mood, and pain location. This kind of data reveals how pain fluctuates across a day and what activities or emotional states make it better or worse, information that a single clinic visit asking “rate your pain from 1 to 10” simply cannot provide.
Physical activity research also relies on ESM. Studies combine accelerometer data (tracking movement) with real-time self-reports about mood and energy, allowing researchers to examine how exercise affects emotional states not just in general, but at specific moments throughout the day.
The Compliance Challenge
The biggest practical limitation of ESM is participant burden. Answering survey prompts multiple times a day for days or weeks is demanding, and not everyone keeps up. Compliance rates across studies vary widely. In adolescent research, average compliance has ranged from 51% to 92% depending on the study design. One large study of young people found an overall compliance rate of just 33.8%, though the picture was more nuanced than that single number suggests.
When researchers looked more closely, they identified four distinct patterns. About 15% of participants maintained high compliance throughout, averaging nearly 79% of prompts answered. Another group held steady at a medium level around 50%. A third group started strong but dropped off over time. And the largest group, 43% of participants, showed low engagement from the very beginning, likely because there was no direct incentive or they simply forgot to respond. These findings highlight that ESM studies need strategies to maintain motivation, whether through compensation, built-in reminders, or keeping surveys short enough that they don’t feel like a chore.
How ESM Data Gets Analyzed
ESM produces a particular kind of dataset: dozens or hundreds of observations per person, collected at irregular intervals, nested within days, nested within individuals. Traditional statistical methods like standard regression or repeated-measures analysis weren’t built for this structure. The go-to approach is multilevel modeling (also called hierarchical linear modeling), which accounts for the fact that observations within the same person are related to each other in ways that observations between different people are not.
Think of it as analysis that works on multiple levels simultaneously. At the lowest level, it examines how your mood varies from moment to moment. At the next level, it looks at how your daily patterns differ from someone else’s. A three-level model might examine individual moments nested within days nested within people. This approach handles the messy realities of ESM data gracefully: missing responses, uneven timing between prompts, and different numbers of observations per participant.