Dr. Egon Dejonckheere
Egon does research in Emotion, Clinical Psychology and Abnormal Psychology. His most recent publication is 'The Bipolarity of Affect and Depressive Symptoms', featured in Journal of Personality and Social Psychology: Personality Processes and Individual Differences.
What is Ecological Momentary Assessment?
Ecological Momentary Assessment (EMA) is a research approach for measuring behavior, experience, and other psychological processes as they occur in people’s daily lives. Instead of asking individuals to recall and summarize their experiences over longer periods, EMA collects data repeatedly in real time or near real time, typically in natural environments such as home, work, or social settings.
EMA is best understood as part of the broader family of ambulatory assessment methods, which aim to capture real-world data outside the laboratory. What distinguishes EMA within this broader framework is its focus on sampling moments of experience and behavior in context, using structured self-reports that are distributed over time. In contrast to mobile sensing, it therefore relies on active participant involvement. Conceptually, EMA closely aligns with the Experience Sampling Method (ESM), which is why many researchers use the two terms interchangeably to describe the same approach.
A helpful way to conceptualize Ecological Momentary Assessment is as a high-frequency, structured diary strategy for everyday life. Rather than attempting to measure a person’s overall experience in a single assessment, EMA collects multiple observations across time and situations. These observations are then used to reconstruct patterns of behavior, emotion, and context as they naturally unfold.
EMA is best understood as part of the broader family of ambulatory assessment methods, which aim to capture real-world data outside the laboratory. What distinguishes EMA within this broader framework is its focus on sampling moments of experience and behavior in context, using structured self-reports that are distributed over time. In contrast to mobile sensing, it therefore relies on active participant involvement. Conceptually, EMA closely aligns with the Experience Sampling Method (ESM), which is why many researchers use the two terms interchangeably to describe the same approach.
A helpful way to conceptualize Ecological Momentary Assessment is as a high-frequency, structured diary strategy for everyday life. Rather than attempting to measure a person’s overall experience in a single assessment, EMA collects multiple observations across time and situations. These observations are then used to reconstruct patterns of behavior, emotion, and context as they naturally unfold.
Core strengths of EMA
Although Ecological Momentary Assessment (EMA) studies are always tailored to their specific research question, several defining qualities shape how the method operates and why it is so powerful.
Momentary assessments
A central feature of EMA is its emphasis on momentary measures. Data are collected at or very close to the time an experience occurs, rather than relying on retrospective recall. This distinction is critical because research shows that memory is not a neutral recording device. Instead, recall involves reconstruction processes that are influenced by heuristics, current mood, and beliefs, which can systematically distort reportsof past experience. EMA reduces this recall bias by capturing experiences before they are filtered through memory.
Ecologically Valid Data
A second defining characteristic is the focus on natural environments. EMA measurements are obtained while participants go about their daily lives, rather than in artificial laboratory settings. This ensures that the data reflect real-world conditions and avoids biases introduced by controlled environments, such as unnatural behavior or lab-specific effects. The external or ecological validity of the data obtained is substantially higher than that in experimental lab studies.
Repeated assessments
A third key feature is the use of repeated observations. EMA typically involves collecting many measurements per individual, sometimes dozens or even hundreds. These repeated observations serve two purposes. First, they may provide more reliable estimates of a person’s typical state. Second, and more importantly, they allow researchers to examine how experiences vary across situations and over time. This within-person variability is not treated as noise, but as meaningful information about psychological functioning and its temporal or situational effects.
Different types of sampling schemes
A fourth characteristic concerns the timing and control of assessments. In EMA, the timing of data collection is carefully designed to obtain a representative sample of experience. Assessments may be scheduled at fixed intervals, randomly within time windows, or triggered by (changes in) passive data streams. In other cases, participants decide when to report, on the basis of specific events they encounter.
Context-dependent focus
Finally, EMA is inherently context-sensitive. Assessments often include information about situational factors such as location, activity, social setting, or environmental conditions. This allows researchers to examine how behavior and experience are shaped by context, rather than treating them as isolated or purely internal phenomena.
Taken together, these characteristics make EMA a powerful method for capturing the complexity of real-world behavior. By combining momentary assessment, ecological validity, repeated observations, careful sampling, and contextual information, EMA provides a detailed and dynamic picture of how people function in daily life.
Taken together, these characteristics make EMA a powerful method for capturing the complexity of real-world behavior. By combining momentary assessment, ecological validity, repeated observations, careful sampling, and contextual information, EMA provides a detailed and dynamic picture of how people function in daily life.
Potential limitations of EMA
While Ecological Momentary Assessment (EMA) may offer clear advantages in terms of ecological validity and temporal resolution, these strengths come with trade-offs. Compared to traditional retrospective surveys and laboratory studies, EMA introduces several methodological, practical, and analytical challenges that researchers need to consider carefully.
Participant burden, compliance and careless responding
One of the most immediate downsides of EMA is participant burden. Repeated assessments throughout the day can be intrusive and cognitively demanding, especially in intensive designs with high sampling frequencies. This burden affects both the quantity and quality ofthe data.
In terms of data quantity, higher burden typically leads to reduced EMA compliance or adherence over time, resulting in missing observations or even drop-out. While some missingness is to be expected, a more important question is whether this missingness is fully random. Participants may be less likely to respond in specific contexts, such as when they are busy, stressed, or engaged in certain activities. As a result, parts of daily life may be systematically underrepresented in the final dataset of experiences.
In terms of data quality, even when participants do respond, high burden can lead to more superficial or inattentive answers. Participants may rush through assessments to minimize disruption, increasing the likelihood of careless responding. This introduces additional noise or measurement error that is conceptually distinct from missing data, as responses are present but less reliable.
Together, these issues illustrate that participant burden in EMA does not only reduce how much data is collected, but also affects how trustworthy the obtained data are.
In terms of data quantity, higher burden typically leads to reduced EMA compliance or adherence over time, resulting in missing observations or even drop-out. While some missingness is to be expected, a more important question is whether this missingness is fully random. Participants may be less likely to respond in specific contexts, such as when they are busy, stressed, or engaged in certain activities. As a result, parts of daily life may be systematically underrepresented in the final dataset of experiences.
In terms of data quality, even when participants do respond, high burden can lead to more superficial or inattentive answers. Participants may rush through assessments to minimize disruption, increasing the likelihood of careless responding. This introduces additional noise or measurement error that is conceptually distinct from missing data, as responses are present but less reliable.
Together, these issues illustrate that participant burden in EMA does not only reduce how much data is collected, but also affects how trustworthy the obtained data are.
Measurement reactivity and interference
EMA assumes that repeated measurement captures processes as they naturally unfold, but the act of measuring itself can also influence those processes. Participants may become more aware of their emotions, behaviors, or symptoms simply because they are asked about them repeatedly. This measurement reactivity can lead to changes in the behavior or experience itself, for example a reduction in symptom intensity due to increased self-monitoring.
At the same time, it is difficult to demonstrate this effect conclusively, because it requires separating true change from change caused by the measurement. Studies across different domains, such as body image, pain, craving, emotions, and even suicidal ideation, generally find that these effects are small or non-existent at the group level. However, this does not mean they are irrelevant. For some individuals, repeated assessment may still meaningfully influence their experiences or behavior, which researchers should take into account.
A related concern is measurement interference, which can occur at both a behavioral and an affective level. Behavioral interference means that EMA prompts can interrupt what people are doing, such as conversations, work, or daily routines, and may therefore change their behavior prior to or in that moment. In contrast, affective interference refers to changes in how people feel as a result of the assessment itself, such as annoyance, irritation or fear of missing a notification. This distinction shows that EMA can influence both what people do and how they feel, meaning it does not only observe daily life, but may also subtly shape it.
At the same time, it is difficult to demonstrate this effect conclusively, because it requires separating true change from change caused by the measurement. Studies across different domains, such as body image, pain, craving, emotions, and even suicidal ideation, generally find that these effects are small or non-existent at the group level. However, this does not mean they are irrelevant. For some individuals, repeated assessment may still meaningfully influence their experiences or behavior, which researchers should take into account.
A related concern is measurement interference, which can occur at both a behavioral and an affective level. Behavioral interference means that EMA prompts can interrupt what people are doing, such as conversations, work, or daily routines, and may therefore change their behavior prior to or in that moment. In contrast, affective interference refers to changes in how people feel as a result of the assessment itself, such as annoyance, irritation or fear of missing a notification. This distinction shows that EMA can influence both what people do and how they feel, meaning it does not only observe daily life, but may also subtly shape it.
Response shifts
Another related issue is response shifts, such as reconceptualization and recalibration. Reconceptualization refers to changes in how participants understand the construct being measured. For example, what someone considers “stress” or “social company” may evolve over the course of the study as they reflect on these experiences more often. Recalibration, in contrast, refers to changes in how participants use the response scale. Participants may start to use the scale more strictly or more leniently over time, even if their underlying experience has not changed.
These processes mean that changes in EMA data do not always reflect true changes in the underlying construct, but may instead reflect shifts in interpretation or scale use. This makes it potentially more challenging to interpret within-person variability over time, as observed trends can partly be driven by how participants report their experiences rather than by the experiences themselves.
These processes mean that changes in EMA data do not always reflect true changes in the underlying construct, but may instead reflect shifts in interpretation or scale use. This makes it potentially more challenging to interpret within-person variability over time, as observed trends can partly be driven by how participants report their experiences rather than by the experiences themselves.
Privacy and ethical considerations
Because EMA often involves collecting data in real-world contexts, sometimes in combination with mobile sensing, it raises important privacy concerns. Participants may be asked to report sensitive information multiple times per day (e.g., How depressed are you right now?), and continuous or context-aware data collection can increase perceived intrusiveness. Ensuring informed consent, data security, and processing transparency is therefore more demanding than in many traditional study designs.
No causal inference
Compared to laboratory studies, EMA sacrifices experimental control in favor of ecological validity. Researchers have limited control over participants’ environments, timing of events, and contextual influences. This makes it more difficult to establish causal relationships, as confounding variables cannot be controlled as tightly as in lab settings. While temporal ordering in EMA can support stronger inferences than cross-sectional data (e.g., Granger causality), EMA is confined to an observational or correlational analysis level and does not fully substitute for experimental manipulation.
Analytical complexity
The intensive longitudinal data generated by EMA are inherently multilevel, time-dependent, and often unbalanced due to missing observations. Analyzing such data requires specialized statistical approaches, such as multilevel models, time series analyses, or dynamic structural equation models. These methods are more complex than the standard techniques used for cross-sectional survey data and require stronger statistical expertise. In addition, interpreting results can be challenging, particularly when disentangling within-person and between-person effects.
What research questions can EMA answer?
Ecological Momentary Assessment (EMA) is particularly valuable because it enables a range of research questions that cannot be addressed using traditional methods (e.g., cross-sectional surveys or lab-based studies). Although the method has been applied across diverse domains such as the science of emotion, psychopathology, pain, social media use, diet and nutrition, and personality, the underlying research questions it addresses can typically be reduced to one of the following core types.
Examining psychological constructs over time
A first type of questions concerns how behavior and experience unfold over time. Because EMA collects repeated measurements, it allows researchers to describe trajectories and patterns within individuals, such as how symptoms change throughout the day or how behaviors evolve across longer periods. These analyses provide insight into the natural development of psychological processes.
For example, in the domain of emotion, EMA can reveal diurnal patterns such as a gradual increase in irritability or anxiety over the course of the day, followed by recovery in the evening. In the domain of personality, EMA can show that characteristics typically considered stable, such as extraversion, fluctuate substantially within individuals across situations and moments, revealing that these traits are less stable in daily life than traditional cross-sectional trait measures would suggest.
For example, in the domain of emotion, EMA can reveal diurnal patterns such as a gradual increase in irritability or anxiety over the course of the day, followed by recovery in the evening. In the domain of personality, EMA can show that characteristics typically considered stable, such as extraversion, fluctuate substantially within individuals across situations and moments, revealing that these traits are less stable in daily life than traditional cross-sectional trait measures would suggest.
Examining temporal interrelations between psychological constructs
A second series of questions focuses on the relations between constructs in everyday settings. EMA makes it possible to examine how different variables co-occur within specific moments. For example, researchers can investigate whether certain emotional states tend to arise in particular situations, or how contextual factors are linked to behavior. In the domain of psychopathology, for instance, EMA data may show that moments characterized by social interaction are consistently accompanied by lower levels of negative affect within individuals, whereas moments spent alone are more often associated with increased depression or rumination.
Relatedly, researchers can examine temporal sequences and mechanisms. EMA allows researchers to examine what happens before and after specific events, such as episodes of stress, symptom onset, or behavioral lapses. By preserving the temporal order of observations, EMA enables the study of predictive chains and short-term processes, often referred to as microprocesses. For example, in pain research, EMA can reveal that increases in momentary stress are followed by elevated pain symptoms at the next assessment, which in turn predicts subsequent avoidance behavior or decreases in activity, thereby mapping out a short-term process that unfolds across hours within daily life.
Relatedly, researchers can examine temporal sequences and mechanisms. EMA allows researchers to examine what happens before and after specific events, such as episodes of stress, symptom onset, or behavioral lapses. By preserving the temporal order of observations, EMA enables the study of predictive chains and short-term processes, often referred to as microprocesses. For example, in pain research, EMA can reveal that increases in momentary stress are followed by elevated pain symptoms at the next assessment, which in turn predicts subsequent avoidance behavior or decreases in activity, thereby mapping out a short-term process that unfolds across hours within daily life.
Examining individual differences in temporal variability patterns
A final class of questions addresses individual differences in patterns of variability. While EMA emphasizes within-person dynamics, it also reveals that individuals differ in how these dynamics unfold. For example, in psychopathology research, some individuals may show strong coupling between stress and mood, while others show weaker associations. In the domain of diet and nutrition, EMA can show that some individuals exhibit pronounced fluctuations in unhealthy snacking across the day, with clear peaks in the evening, whereas others maintain relatively stable eating patterns.
EMA thus supports both generalizable findings and more personalized insights. These patterns are typically modeled using multilevel analyses, which explicitly account for the nested structure of repeated observations within individuals and allow researchers to quantify both within-person dynamics and between-person differences simultaneously.
EMA thus supports both generalizable findings and more personalized insights. These patterns are typically modeled using multilevel analyses, which explicitly account for the nested structure of repeated observations within individuals and allow researchers to quantify both within-person dynamics and between-person differences simultaneously.
Roots and history of EMA
Ecological Momentary Assessment (EMA) emerged from a broader effort within psychology and behavioral medicine to study phenomena in real-world settings rather than relying solely on laboratory experiments or retrospective reports.
Traditional methods posed clear limitations. Laboratory studies, while offering experimental control, often failed to capture behavior as it occurs in everyday life. Retrospective self-reports, although easy to administer, were shown to be affected by memory decay, reconstruction processes, and cognitive biases, leading to systematic inaccuracies.
In response to these limitations, researchers began developing methods that directly sample experience and behavior in natural contexts. EMA drew on several earlier traditions, including daily diary studies, self-monitoring approaches in behavioral research, and Experience Sampling Methodology (ESM). The term was coined in the early 1990s and was formally introduced in the work of Arthur Stone and Saul Shiffman, who used EMA within health and clinical research contexts.
Early EMA studies relied on paper diaries and signaling devices such as pagers or programmable watches. While these methods captured the essential logic of momentary assessment, they were limited by issues such as delayed reporting (i.e., backlogging) and incomplete data (e.g., lost diary entries).
Technological advances have since transformed EMA. The introduction of electronic diaries, handheld devices, and more recently smartphones and wearable sensors has enabled more precise timing, automated prompting, and integration of multiple data streams. These developments have made EMA more accurate, scalable, and widely applicable across different research domains.
Traditional methods posed clear limitations. Laboratory studies, while offering experimental control, often failed to capture behavior as it occurs in everyday life. Retrospective self-reports, although easy to administer, were shown to be affected by memory decay, reconstruction processes, and cognitive biases, leading to systematic inaccuracies.
In response to these limitations, researchers began developing methods that directly sample experience and behavior in natural contexts. EMA drew on several earlier traditions, including daily diary studies, self-monitoring approaches in behavioral research, and Experience Sampling Methodology (ESM). The term was coined in the early 1990s and was formally introduced in the work of Arthur Stone and Saul Shiffman, who used EMA within health and clinical research contexts.
Early EMA studies relied on paper diaries and signaling devices such as pagers or programmable watches. While these methods captured the essential logic of momentary assessment, they were limited by issues such as delayed reporting (i.e., backlogging) and incomplete data (e.g., lost diary entries).
Technological advances have since transformed EMA. The introduction of electronic diaries, handheld devices, and more recently smartphones and wearable sensors has enabled more precise timing, automated prompting, and integration of multiple data streams. These developments have made EMA more accurate, scalable, and widely applicable across different research domains.
How to conduct an Ecological Momentary Assessment (EMA) study?
Conducting an Ecological Momentary Assessment (EMA) study involves several key steps. First, define a clear research question and identify the processes you want to capture in daily life. Next, design brief, momentary self-reports and choose an appropriate study duration, assessment frequency and sampling strategy (e.g., time-based or event-based notifications). After setting up the study in a mobile platform, participants may be onboarded during an intake and instructed on how to respond in the EMA app in real time.
A crucial step is piloting the study. Pilots allow you to test whether the assesment frequency, questionnaire length, and timing are feasible in real life, and whether participants understand the items as intended. They also help identify technical issues and reduce the risk of low compliance or poor data quality in the full study.
Once data collection starts, monitor compliance and data quality, and finally analyze the data using methods that account for its multilevel and time-dependent structure. If you are planning to run an EMA study, using a dedicated platform can help streamline setup, data collection, and secure data storage, allowing you to focus on your research questions rather than the technical overhead.
For a more detailed walkthrough of each stage and the associated considerations, access our step-by-step guide to running an EMA study.
A crucial step is piloting the study. Pilots allow you to test whether the assesment frequency, questionnaire length, and timing are feasible in real life, and whether participants understand the items as intended. They also help identify technical issues and reduce the risk of low compliance or poor data quality in the full study.
Once data collection starts, monitor compliance and data quality, and finally analyze the data using methods that account for its multilevel and time-dependent structure. If you are planning to run an EMA study, using a dedicated platform can help streamline setup, data collection, and secure data storage, allowing you to focus on your research questions rather than the technical overhead.
For a more detailed walkthrough of each stage and the associated considerations, access our step-by-step guide to running an EMA study.
Dr. Egon Dejonckheere
Egon Dejonckheere is an Assistant Professor at Tilburg University, affiliated with KU Leuven, specializing in ambulatory assessment and intensive longitudinal research. His work combines methodological innovation to improve these methods with applications in mental health and emotion dynamics, and includes dozens of smartphone-based studies in daily life.
FAQ
01.
What is the difference between EMA and ESM?
Ecological Momentary Assessment (EMA) and the Experience Sampling Method (ESM) refer to the same research approach: collecting repeated, real-time self-reports in daily life. The terms originate from different research traditions, but are now used interchangeably in most scientific contexts.
02.
Why is EMA more accurate than traditional surveys?
EMA is more accurate than traditional surveys because it measures experiences in real time rather than relying on memory. This reduces recall bias, which occurs when people reconstruct past experiences using heuristics, current mood, or beliefs.
03.
Why is EMA better than traditional lab research?
EMA is better than traditional lab research because it measures experiences in real time in everyday life, increasing ecological validity. It allows researchers to capture how psychological processes unfold in people’s actual lives, rather than relying on artificial settings or single-point measurements.
04.
What sampling methods are used in EMA?
The main EMA sampling methods are time-contingent (assessments scheduled at fixed or random times) and event-contingent (reports triggered by specific events or experiences, including those detected via sensors such as changes in location or physiological signals). Choosing between these approaches depends on your research question, as each comes with its own advantages and limitations in terms of burden, precision, and ecological validity.
05.
What are the main disadvantages of EMA?
EMA offers high ecological validity and detailed temporal data, but it comes with trade-offs. The main disadvantages include participant burden, missing data, careless responding, measurement reactivity, reduced experimental control, privacy concerns, and analytical complexity. These challenges make EMA studies more demanding to design, run, and interpret compared to traditional surveys or laboratory studies.
06.
What is measurement reactivity in EMA?
Measurement reactivity refers to the idea that repeatedly asking participants about their experiences can change those experiences. For example, frequent questions about mood or symptoms may increase self-awareness and lead to changes in behavior or emotional intensity. While research generally finds small effects at the group level, reactivity can still be meaningful for individual participants.
07.
What are reconceptualization and recalibration in EMA?
Reconceptualization and recalibration are types of response shifts. Reconceptualization occurs when participants change how they understand a concept, such as “stress” or "social company”, over time. Recalibration refers to changes in how participants use response scales, for example becoming stricter or more lenient. These shifts make it harder to determine whether observed changes reflect real changes or changes in reporting.
08.
What kind of research questions can EMA answer?
EMA can answer three core types of research questions in daily life: how a single psychological variable changes over time, how different variables relate to each other over time, and how individuals differ in their patterns of variability and dynamics in these variables.
09.
Why use smartphones instead of paper diaries in EMA studies?
Smartphones with an EMA app improve data quality by enabling real-time prompts, automatic time-stamping, and compliance tracking. This reduces backfilling and missing data, and allows integration with sensors and wearable devices. In some cases, however, paper-and-pencil EMA may still be a viable option.