ILD collection methods
This article provides an overview of the various methods that you can choose from to collect intensive longitudinal data (ILD). Familiarity with the terminology, technology, and targets associated with ILD collection methods enables you to more easily understand what other researchers are referring to, and whether their research and findings are relevant for what you are trying to learn about. However, gaining this familiarity can be challenging, because of the wide variety of methods and the various terminologies that are used to refer to these in different fields.
The current article takes a broad perspective by focusing on all data collection methods that are based on getting data from one or more cases (e.g., individuals or dyads), on many occasions. Hence, these include ambulatory methods by which you can measure people in their natural setting while they are living their daily life, but also data obtained in a lab setting. In this sense it deviates from many key accounts of ILD methodology, which tend to focus on ambulatory methods only (e.g., Conner & Lehman, 2012; Hektner et al., 2007; Mehl et al., 2024; Stone et al., 2007). The reason for preferring a broader perspective here is that the data that are obtained in the lab or in every day life both classify as ILD, and may have similar features in terms of temporal patterns, may be used to tackle similar research questions, and may require similar analysis techniques.
Below you can read more about data collection methods that can be used to obtain: 1) self-report data; 2) physiological data; 3) behavioral data; and 4) environmental data. To learn more about some of the terminology that exists in the ILD literature, you can also read the article about ILD umbrella terms.
1 Self-report data
Self-reports are a common way to obtain ILD in psychology and related fields. Self-reports may concern the participant’s affects, thoughts, attitudes, beliefs, and behaviors, but may also be used to ask about characteristics of the environment or context that the participant is in, for instance, whether one is inside or outside, and whether one is alone or with others. Moreover, self-reports can be used to obtain information about the occurrence of events such as smoking or drinking, having a panic attack, binging, or having an argument.
1.1 Daily diaries (DD)
These are self-reports that are based on asking participants to complete one report per day, often at the end of the day, concerning their experiences and behaviors during that day (Barta et al., 2012). This approach captures a daily summary perspective, making it suited for studying processes that unfold across days rather than moments.
1.2 Experience sampling methodology (ESM)
These are self-reports obtained at multiple time points throughout the day (Csikszentmihályi et al., 1977). The goal of ESM is “to capture daily life as it is directly perceived from one moment to the next” (p.6, Hektner et al., 2007). The focus can be on feelings, activities, thoughts, and self-appraisal, but also on the physical and social context that the participant is in. Typically, ESM data are obtained using random measurement occasions rather than fixed times, to prevent participants anticipating the next measurement occasion and adjusting their natural behavior because of this.
1.3 Event-contingent measurements
This ILD collection method is appropriate when you are interested in studying experiences and behaviors that only exist—or are only relevant—when a specific event is either taking place, or just occurred (Moskowitz & Sadikaj, 2012). Examples of events are behaviors like eating, smoking, social media use, or traveling. Other examples include social interactions, panic attacks, or being at a particular location. For some of these events it is possible to detect their occurrence automatically; alternatively, the participant is asked to indicate whenever an event has occurred. This may trigger either a single measurement occasion, or a series of densely spaced measurement occasions to allow you to investigate the temporal unfolding of the effect of the event (Schreuder et al., 2024). Ji et al. (in press) coined the term episode-contingent experience-sampling burst design for the latter.
1.4 Day reconstruction method (DRM)
This technique requires participants to think of their day as a continuous series of episodes and asks them to: a) give each episode a name (for example, “commuting to work”, or “at lunch with B”); and b) indicate the approximate times at which each episode began and ended. Participants are then asked to indicate for each episode what they were doing, where and with whom they were, and how they were feeling. DRM was inspired by both daily diaries and ESM, with the goal to obtain measurements of experiences during specific activities and circumstances; compared to ESM, it should be less expensive and less burdensome for participants, as well as better able to capture uncommon and brief events (Kahneman et al., 2004).
1.5 Intensity profile drawings
Like DRM, this method was also proposed as a way to shed light on the “blind spots” in between measurement occasions (Cloos et al., 2025). In this method participants are prompted at fixed time intervals to provide their momentary affective experiences, as typical in ESM. Additionally, participants are shown their current and previous affect rating plotted against time and asked to connect these two points with a line that represent the fluctuations in their affect between the two measurement occasions. Hence, this only requires recall since the previous measurement occasion, rather than for the entire day.
1.6 Conclusion
Most of the ILD collection methods that involve self-reports are based on using a digital device (e.g., a smart phone) to obtain responses from participants to particular questions, using categorizations (e.g., yes-no questions) or quantifications (e.g., a Likert scale, or slide ruler). Some researchers also offer participants open ended questions, allowing them to write or audio record their responses to these queries, which can then be used in qualitative research (McCombie et al., 2024).
A common critique of self-report methods is that they are subjective and that they may miss important aspects of the process of interest because these occur outside a person’s conscious experience. Moreover, because self-reports require a person to take time to answer the questions, techniques that require multiple measurement occasions throughout the day can be (somewhat) disruptive; this places practical constraints on the number and temporal spacing of measurement occasions. There are also concerns that the act of self-reporting actually intervenes with the very process a researcher tries to measure. For these reasons, other data collection methods discussed below may be of interest as an alternative for—or a complement to—self-reports in ILD research.
2 Physiological data
Physiological data come from measuring one or more aspects of bodily processes. They are typically regarded as objective, which is considered a major strength in comparison to the self-report measurements described above, which are inherently subjective. However, in the context of psychological research a major challenge lies in linking these objective signals to constructs that are substantively meaningful.
2.1 Heart rate, skin temperature, galvanic skin response, and respiration
Wearable devices—like smart watches and smart rings—can be used to measure various aspects of bodily processes with minimal interference with normal everyday life. Wearables can be used to collect data continuously over time, at frequent intervals, or based on a trigger (e.g., when the participant is at a certain location), depending on the process you are interested in. The major challenge with these data lies in how to link them to a psychological construct such as stress, or positive or negative affect: For instance, heart rate is notoriously sensitive to body movement and posture (Brouwer et al., 2018), and even when these are corrected for, it is still rather difficult to distinguish between cognitive or emotional arousal, and in the latter case, whether it concerns pleasant or unpleasant emotions (Yang et al., 2017). One option to get more insight is by using mental arousal as a trigger for self-reports in an event contingent design (Myrtek et al., 2005).
2.2 Biomarkers
This category of physiological data requires more intrusive or involved forms of data collection. An example is to use a small implant under the skin to measure a person’s glucose levels in daily life more or less continuously (e.g., every 5 minutes) for several days up to weeks (Funtanilla et al., 2019). There is also technology that allows you to measure a person’s alcohol consumption with a wearable device that measures the alcohol that defuses through their skin (Barnett et al., 2014). Another common approach is to measure momentary cortisol, which is considered an important indicator of stress in everyday life, from a person’s saliva by having the participant place a swab in their mouth until it is soaked (Schlotz, 2012). Note that the latter is somewhat involved and therefore typically only repeated a few times, in which case the obtained measurements do not qualify as ILD.
2.3 Electroencephalogram (EEG) and functional magnetic resonance imaging (fMRI)
Both EEG and fMRI can be used to measure brain activity and result in very dense ILD, with observations obtained at a timescale raging from milliseconds to seconds. Typically, such data are obtained in a lab setting, as it requires specialized equipment that tends to be immobile; participants are then measured while they are presented with specific tasks, to determine how their brain activity changes in response to this. Recently, there have been developments that allow for the measurement of brain activity outside the lab Uchitel et al. (2021). Although not part of the typical ambulatory data collection methods toolkit at this point, these technological innovations illustrate that there are few limitations when thinking about how to bring common lab techniques to everyday life outside the lab.
2.4 Conclusion
While the objectiveness of the data in this category is often considered a key strength, the downside of these measures is that it is challenging to link them to psychologically meaningful constructs. By combining them with other ILD collection methods such as self-reports and/or behavioral measurements, you can create a rich and diverse data collection plan that allows you to study specific psychological processes from various perspectives.
3 Behavioral data
Behavior can be measured through self-reports, but also through the use of wearables, video recordings, or phones. Some of these data collection methods only require the participant to wear a certain device, whereas other methods focus on or actually require interaction with a particular device.
3.1 Physical behavior
Physical activities, postures, and movement can be measured with wearable devices such as an accelerometer or actigraph (Giurgiu & Bussmann, 2024). Such devices typically record activity in (near) continuous time, thereby producing large quantities of ILD. One way in which the measurement of posture and movement plays a role in the study of psychological processes is that they are used in attempts to isolate the psychological component that is present in heart rate and respiration, but that are obscured by them. Researchers have also considered the relation between sedentary behavior (i.e., low-intensity behavior while being awake, such as watching tv, or sitting in a car) and various health factors, including cardiovascular disease (Lavie et al., 2019) and depression (Huang et al., 2020). The measures of physical behavior can also be combined with self-reports on for instance stress, to investigate whether stress is followed by less physical activity, and/or physical activity tends to be followed by lower self-reported stress levels (Wright et al., 2023).
3.2 Video recordings
Video recordings easily result in ILD that can be used to categorize and quantify verbal and nonverbal behaviors in the lab, but also in natural settings. The challenge with such data is to convert the raw recordings into meaningful data; this is typically done either by an expert rater (Sadler et al., 2009), or through automatic coding (Flemotomos et al., 2022; Li et al., 2023). Another approach is to have participants watch a video recording of themselves and rate their own affective state as they recall while watching the video (Gottman & Levenson, 1985). Video recordings can be used in a lab setting, but have also been used frequently during home visits to obtain observations from the natural setting. Other scenarios in which video recordings have been used are in a therapeutic setting (Thomas et al., 2014), or in student-teacher interactions taking place in a classroom setting (Pennings et al., 2018).
3.3 Phone or app use
An increasingly important source of ILD is the mobile phone. Data concerning phone use in daily life may include information about timing and duration of phone calls, content of text messages, and timing and duration of social media use (Müller et al., 2024). This form of data collection can easily result in huge quantities of data, which poses challenges for how to extract meaningful information from them (Brinberg et al., 2021; Reeves et al., 2020). To link raw data consisting of frequent screenshots to affective experiences of individuals, it is possible to use either automated or human ratings of emotional content in images; but while these have been shown to align well (Rocklin et al., 2023), they may miss important idiographic meaning and experiences. For instance, for most people a photo of a puppy may induce positive feelings, but it may also provoke discomfort in someone with a fear of dogs. The collection of phone and app use data may also be combined with self-reports about affective experiences, which allows you to investigate whether individuals for instance tend to feel more or less positive while (or just after) using specific apps (Minich & Moreno, 2024).
3.4 Cognitive performance
A common example of behavioral ILD are the strings of reaction times and accuracies on trials of a decision-making task administered repeatedly in a lab setting. Such ILD are often obtained with the intention to measure cognitive processes; typically, such data are then summarized per block of trials or per condition, ignoring the temporal dynamics in these data. To determine the day-to-day dynamics of cognitive performance, you can obtain and summarize such data at a daily frequency (e.g., Schmiedek et al., 2010). Moreover, some researchers have started to include small cognitive tasks into their measurement procedures throughout daily life, which allows for a more detailed perspective on everyday variation in cognitive performance (Schmitter-Edgecombe et al., 2020).
3.5 Speech
Another behavior that is of interest to psychological researchers is a person’s speech. You can collect speech data in a controlled setting, such as a lab or a visit to the doctor or therapist, but also in everyday life with a device like the electronically activated recorder (EAR); this is a wearable device that obtains audio snippets throughout the day (Mehl et al., 2001; Mehl, 2017). To derive information about a person’s psychological well-being, you may consider aspects such as rate, pauses and pitch variability, which may be indicators of depression (Wadle et al., 2024).
3.6 Conclusion
Behavior can be measured in many ways, including through self-reports; the methods described in this section are considered objective, rather than subjective. The data collection methods may differ in the degree to which the participant has to actively engage with a device (for the purpose of measurement alone), and in the timescale the measurements tend to tap into. Which of these techniques is most appropriate, will depend on the research question you are trying to answer. As with the other objective techniques, a specific challenge lies in how to link the raw data to particular psychological constructs, such as the affective state of a person is in, or the degree of connection between two individuals.
4 Environmental data
The final category of ILD collection methods concerns the environment a person is in. While such information is sometimes also obtained through self-report, the focus here is on data collection methods that do not require active engagement from the participant.
4.1 Location
Examples of location data are those obtained with the global position system (GPS) or WiFi (Lautenbach et al., 2024). Location tracking can produce dense ILD, with updates ever few seconds, depending on the device used and the study design. You can use GPS data to determine the context of a person in terms of green space, noise level, or air pollution. You can also use location information to make predictions: For instance, staying more at home (over the course of a week) than usually may indicate an increased risk of a suicide attempt in high-risk adolescents (Auerbach et al., 2025). You may also use location as a trigger to start a series of self-report questions in an event-contingent design (Ebner-Priemer & Santangelo, 2024), or to trigger an intervention in real time using a just-in-time adaptive intervention (Nahum-Shani et al., 2018).
4.2 Light exposure
Exposure to light has been an important variable in studies on cicadian rhythms (Adamsson et al., 2017), sleep patterns (Didikoglu et al., 2023), and insomnia (Zeitzer et al., 2011). There are various devices that participants can wear, which can be used to record ambulatory light exposure in real-life settings. By combining such measures with self-reports, you can investigate the dynamics between ambient light and momentary subjective experiences. For instance, Shankman et al. (2025) found that light exposure and physical activity (measured via actigraphs) were both independently associated with higher levels of momentary positive affect.
4.3 Ambient sound
Another environmental feature that has been considered of interest in ILD research is sound from the environment, as this can provide valuable information about the situation and context a person is in. For example, it can be used to determine whether a person is inside or outside, and whether a person is exposed to sounds that may serve as distractors or stressors. When you measure ambient sound, you will also capture sounds related to behaviors of the participant, for instance, whether they are talking, making music, or doing construction work.
4.4 Conclusion
Different aspects of the environment may be of interest for different reasons. For instance, you may want to know how stressful the environment is in terms of business and noise; or you may be interested in how often people are exposed to or engaged in conversations; or you may be interested in the amount of light that people are exposed to. Psychologists will typically use this information to relate it to psychological constructs such as concentration, fatigue, or depression. This means that environmental data will often be combined with other ILD based on self-reports, physiological and/or behavioral measurements.
5 Think more about
When choosing a data collection method, you have to consider how the construct you are interested in can be best operationalized and measured, what timescale you believe the process of interest is operating on and how to tap into this with a specific temporal design, and what a particular ILD collection method combined with a specific temporal design implies in terms of participant burden. Some of these consideration—like whether measurements are subjective or objective, how densely they need to be spaced in time, and whether the process needs to be measured in the lab or in real life—were already touch upon in the descriptions above.
In addition to knowing about the various ILD collection methods described above, it is also useful to familiarize yourself with some of the commonly encountered umbrella terms from the ILD literature. These terms—like ecological momentary assessment and passive sensing—refer to selections of the ILD collection methods discussed above. Many of these terms are also used interchangeably, depending on the discipline you are in, which makes it at times somewhat challenging to read outside one’s own discipline. It is therefore also advisable to always be very clear on what it is that you are referring to with a specific term.
6 Takeaway
There are many methods that you can use to obtain ambulatory, real-life measurements from people in their natural habitat. While these techniques differ in whether they rely on active or passive sensing, and whether they focus on a person’s behaviors, emotions, cognitions, physiology, and/or context and environment, the shared goal is to get information about processes as they take place in every day life, thereby maximizing the ecological validity of these measurements. There are many names that you encounter in the ILD literature; some of the names are synonyms stemming from different research traditions, whereas other names only partly overlap. It is useful to have some knowledge about how terminology relates to each other, as it helps to see what others are referring to.
In addition to the real-life, ambulatory measurements methods, there are also ILD collection methods that are used to study processes taking place in a lab setting. Various of these techniques overlap with the ambulatory measurement methods (e.g., video and speech recordings, physiological measurements). To capture both categories of measurement techniques and to place the focus on methods that result in many repeated measurement per case, MATILDA uses the term ILD collection methods.
You should not consider the collection methods presented in this article exhaustive of all currently available ILD collection methods. Although the intention is to cover the most prominent techniques here, there may be various promising alternatives not covered here. Furthermore, ongoing advancements in data collection and preprocessing techniques, and innovative combinations of existing methods will continue to offer researchers new opportunities for investigating processes by using ILD.
7 Further reading
We have collected various topics for you to read more about below.
- ILD umbrella terms
- [Ecological momentary assessment versus ecological retrospective assessment]
- [Active versus passive sensing]
- [Retrospective, ecological retrospective, and ecological momentary assessment]
- [Remembering versus experiencing self]
- [Designing a measurement instrument to target between-person or within-person variance]
- [Measurement burst designs]
- [Event contingent bursts]
Acknowledgments
This work was supported by the European Research Council (ERC) Consolidator Grant awarded to E. L. Hamaker (ERC-2019-COG-865468).
References
Citation
@article{hamaker2026,
author = {Hamaker, Ellen L. and Hoekstra, Ria H. A.},
title = {ILD Collection Methods},
journal = {MATILDA Preprints},
number = {2026-06-05},
date = {2026-06-05},
url = {https://matilda.fss.uu.nl/articles/ild-collection-methods.html},
langid = {en}
}