-
Title
-
Ambulatory Assessment: Methods for Studying Everyday Life
-
Author
-
Conner, Tamlin S.
-
Mehl, Matthias R.
-
Research Area
-
Methods of Research
-
Topic
-
Research Methods ‐ Quantitative
-
Abstract
-
Ambulatory assessment is a class of methods that use mobile technology to understand people's biopsychosocial processes in natural settings, in real time, and on repeated occasions. In this essay, we discuss the rationale for ambulatory assessment including the benefits of measuring people in the real world (greater ecological validity, better understanding of people in contexts), in real time (avoidance of memory bias, greater sensitivity for capturing change), and over time (capturing within‐person patterns and temporal trends). Then, we review the latest ambulatory assessment techniques for measuring experiences, behaviors, and physiology in daily life. Experiences such as emotions, physical pain, and daily stressors can be tracked using daily diaries and smartphone‐based experience sampling. Behaviors such as activity, movement, location, and natural language use can be tracked using accelerometers, portable actigraphs, global positioning system (GPS) coordinates, and the electronically activated recorder (EAR). Physiological processes such as heart rate, blood pressure, and electrodermal activity can be measured using an array of ambulatory biosensors. Ambulatory assessment will continue to be revolutionized by smartphones, which are becoming integrated seamlessly into people's lives. Emerging trends include social sensing applications that make inferences about users' psychological processes based on multi‐channel information collected from smartphones, emergence of “big data collection” whereby ambulatory assessment data is gathered en masse from large populations, and the growing field of mobile health. These trends raise questions around the protection of participants' privacy and the synthesis of immense amounts of digital data. Ultimately, these developments will narrow the separation between science and everyday life as ambulatory assessment becomes an integrated part of people's mobile lives.
-
Related Essays
-
Social Epigenetics: Incorporating Epigenetic Effects as Social Cause and Consequence (Sociology), Douglas L. Anderton and Kathleen F. Arcaro
-
To Flop Is Human: Inventing Better Scientific Approaches to Anticipating Failure (Methods), Robert Boruch and Alan Ruby
-
Repeated Cross‐Sections in Survey Data (Methods), Henry E. Brady and Richard Johnston
-
The Evidence‐Based Practice Movement (Sociology), Edward W. Gondolf
-
Meta‐Analysis (Methods), Larry V. Hedges and Martyna Citkowicz
-
The Use of Geophysical Survey in Archaeology (Methods), Timothy J. Horsley
-
Network Research Experiments (Methods), Allen L. Linton and Betsy Sinclair
-
Longitudinal Data Analysis (Methods), Todd D. Little et al.
-
Data Mining (Methods), Gregg R. Murray and Anthony Scime
-
Remote Sensing with Satellite Technology (Archaeology), Sarah Parcak
-
Quasi‐Experiments (Methods), Charles S. Reichard
-
Digital Methods for Web Research (Methods), Richard Rogers
-
Virtual Worlds as Laboratories (Methods), Travis L. Ross et al.
-
Modeling Life Course Structure: The Triple Helix (Sociology), Tom Schuller
-
Content Analysis (Methods), Steven E. Stemler
-
Person‐Centered Analysis (Methods), Alexander von Eye and Wolfgang Wiedermann
-
Translational Sociology (Sociology), Elaine Wethington
-
Identifier
-
etrds0010
-
extracted text
-
Ambulatory Assessment: Methods
for Studying Everyday Life
TAMLIN S. CONNER and MATTHIAS R. MEHL
Abstract
Ambulatory assessment is a class of methods that use mobile technology to understand people’s biopsychosocial processes in natural settings, in real time, and on
repeated occasions. In this essay, we discuss the rationale for ambulatory assessment
including the benefits of measuring people in the real world (greater ecological
validity, better understanding of people in contexts), in real time (avoidance of
memory bias, greater sensitivity for capturing change), and over time (capturing
within-person patterns and temporal trends). Then, we review the latest ambulatory
assessment techniques for measuring experiences, behaviors, and physiology in
daily life. Experiences such as emotions, physical pain, and daily stressors can be
tracked using daily diaries and smartphone-based experience sampling. Behaviors
such as activity, movement, location, and natural language use can be tracked using
accelerometers, portable actigraphs, global positioning system (GPS) coordinates,
and the electronically activated recorder (EAR). Physiological processes such as
heart rate, blood pressure, and electrodermal activity can be measured using an
array of ambulatory biosensors. Ambulatory assessment will continue to be revolutionized by smartphones, which are becoming integrated seamlessly into people’s
lives. Emerging trends include social sensing applications that make inferences
about users’ psychological processes based on multi-channel information collected
from smartphones, emergence of “big data collection” whereby ambulatory
assessment data is gathered en masse from large populations, and the growing field
of mobile health. These trends raise questions around the protection of participants’
privacy and the synthesis of immense amounts of digital data. Ultimately, these
developments will narrow the separation between science and everyday life as
ambulatory assessment becomes an integrated part of people’s mobile lives.
INTRODUCTION: DEFINITION OF AND RATIONALE FOR
AMBULATORY ASSESSMENT
Laboratory-based methods have historically been the strength and pride
of the social and behavioral sciences. Most research method textbooks
emphasize how laboratory research helps control confounding variables
and thereby allows for the isolation of a causal factor of interest. Laboratory
Emerging Trends in the Social and Behavioral Sciences. Edited by Robert Scott and Stephen Kosslyn.
© 2015 John Wiley & Sons, Inc. ISBN 978-1-118-90077-2.
1
2
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
research has contributed enormously to our understanding of human
social behavior. Ultimately, though, laboratory research can exclusively
accumulate knowledge on what can happen—under isolated and controlled
circumstances—and it cannot speak to what does happen under the circumstances that people normally encounter in their everyday lives. Therefore, a
comprehensive science of human behavior also requires “thinking outside
the experimental box” and necessitates the study of humans in their natural
habitat—that is, the collection of data from individuals as they live their
lives in their daily environments.
This essay reviews existing methodologies for studying experiences,
behavior, and physiology in daily life (for a comprehensive review, see
Mehl & Conner, 2012). We here refer to these methodologies collectively as
ambulatory assessment methods. According to the Society for Ambulatory
Assessment, Ambulatory Assessment “comprises the use of field methods
to assess the ongoing behavior, physiology, experience and environmental
aspects of humans or nonhuman primates in naturalistic or unconstrained
settings. [It] designates an ecologically relevant assessment perspective that
aims at understanding biopsychosocial processes as they naturally unfold
in time and in context” (www.ambulatory-assessment.org). Other common
names to denote these methodologies include experience sampling methods,
diary methods, and ecological momentary assessment. At their core, these
methods allow researchers to study individuals (i) in their natural settings, (ii)
in real time (or close to real time), and (iii) on repeated occasions. Ambulatory
assessment derives its scientific rationale directly from these three measurement characteristics (i.e., “real-world,” “real-time,” “within-person”).
The “real-world” quality allows unprecedented access to what actually
happens in a person’s everyday life and the contexts that surround such
events (Mehl & Conner, 2012, Ch. 1). For example, ambulatory assessment
is a great way to track how much people are exercising every day and the
conditions that make exercise more or less likely. This real-world quality
also enables researchers to capitalize on the power of evocative daily events
that are difficult, if not unethical, to mimic in controlled laboratory settings
such as positive or negative daily interactions, disagreements with a friend,
issues with children, commuting, excessive smoking, or drinking. Moreover,
how people respond to a standardized laboratory stressor such as giving a
speech may differ from how they respond to stressful experiences in their
lives. Wilhelm and Grossman (2010) describe a participant who showed
rather minimal heart rate increases in response to a laboratory stress protocol
but quite dramatic heart rate increases later in the afternoon while watching
a soccer game at home. In addition, in a reverse pattern, patients often have
high blood pressure readings in the physician’s office but not in their home
environment—a phenomenon called the white coat hypertension. By testing
Ambulatory Assessment: Methods for Studying Everyday Life
3
behavior in real-world settings, researchers have greater confidence about
the ecological validity of findings—that is, whether an effect is representative
of how it operates in reality.
The “real-time” quality of ambulatory assessment enables researchers to
understand experiences as they occur, not how people represent them later
from memory. This quality is especially important for experiences that are
fleeting and often misremembered (e.g., emotions, and pain). A person’s
memory for how they “felt over the past week” or “the past month” is
influenced by a host of factors including salient experiences, mood at the
time of recall, personality traits, beliefs about the acceptability of certain
emotions, gender stereotypes, and cultural norms (Mehl & Conner, 2012,
Ch. 2). In fact, there is strong evidence that as the time delay between experience and reporting lengthens, self-reports become more stable, reflective,
belief-driven, and culturally homogenized rather than malleable, reflexive,
experience-driven, and individualized (Robinson & Clore, 2002). This distinction between real-time and recalled reporting can make a big difference
in understanding psychological processes. For one, ambulatory approaches
may be more sensitive for capturing change in emotional states in response
to interventions. In a trial of antidepressant medication treatment, changes in
depressive symptomology were detected earlier among patients randomly
assigned to track their symptoms each day for 30 days compared to patients
who reported their symptoms using standard one-week recall measures
(Lenderking et al., 2008). Other studies find stronger links between real-time
reports of emotion and atherosclerosis risk, immune system function,
and genetic vulnerability [reviewed by Conner and Barrett (2012)]. These
examples illustrate how-real time measurement may be more sensitive than
traditional memory-based measures under certain circumstances.
The “within-person” element of ambulatory assessment refers to the capacity
to identify patterns of behavior within the person across time (i.e., idiographic assessment) (Mehl & Conner, 2012, Ch. 3). This approach differs from
traditional cross-sectional research, that aims to identify patterns of behavior
between people (i.e., nomothetic assessment). Unlike cross sectional research
which observes people at a single time point, ambulatory assessment tracks
individuals’ experiences, behaviors, or physiology intensively over time,
yielding “intensive longitudinal data” (Walls & Schafer, 2006). That data can
be analyzed to uncover the temporal patterns and trends in each person’s
data, as well as commonalities and differences in these trends across people.
For example, in a classic study, Bolger and Schilling (1991) examined the
within-person relationships between daily stressors and anxiety for 339
people and found stronger within-person relationships for people higher
in Neuroticism. Ambulatory assessment has been used to uncover a wide
variety of within-person relationships, including the antecedents of heavy
4
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
drinking, triggers of smoking cravings, relationships between stress and
coping, and links between food consumption and mood.
FOUNDATIONAL RESEARCH: AMBULATORY ASSESSMENT
METHODS
AMBULATORY ASSESSMENT OF EXPERIENCE
One of the most common uses of ambulatory assessment is for studying
self-reported experiences in naturalistic settings. Experiences include current
mood states, levels of perceived stress, feelings of bodily pain or discomfort, and other aspects of daily life that can be self-reported. Thus, ambulatory assessment of experience inherently involves asking people to respond
to questions about their experiences as they go about their daily lives. The
Internet and, especially smartphones have revolutionized this methodology
(Miller, 2012) and now represent the main interface for ambulatory self-report
techniques. In general, there are two main ways of measuring experience
through ambulatory assessment—through Internet daily diaries and mobile
phone-based experience sampling.
Daily diary methods are probably the most common form of ambulatory
assessment. Daily diary methods are now done mainly through Internet surveys in which participants access the survey each day (or night) for a week
to three weeks to answer questions about their experiences that day. Diaries
are often used to assess experiences and events that occur on a daily basis
and can be easily recalled at the end of each day such as daily hassles, social
activities, emotions felt that day, and health behaviors such as alcohol use,
smoking, and food consumption. Although this approach is neither “real
time” nor necessarily ambulatory, it is considered “near to real time” and an
improvement upon asking people to recall their experiences across the entire
week or longer. Although daily diary studies are widely applied, they have
probably had the most impact on the psychology of romantic relationships,
the role daily stressors in mental and physical health, and the correlates and
consequences of daily health behaviors such as alcohol use (Mehl & Conner,
2012, Ch. 8). In addition, daily diary techniques are fairly easy to implement.
Researchers can design a simple Internet survey through any number of companies such as SurveyMonkey, Qualtrics, or SurveyGizmo. Participants then
access that survey during a specified time period, and the data can be downloaded at the end of the study.
A second approach to measuring experience is through mobile-phone-based
experience sampling. Experience sampling is the process of randomly signaling a person (typically 4–8 times per day) to report on their experiences at
Ambulatory Assessment: Methods for Studying Everyday Life
5
that moment [see suggested reading by Hektner, Schmidt, and Csikszentmihalyi (2007)]. This approach is used to study fleeting and ongoing subjective
experiences such as mood, pain, and stress because these experiences are
quick to decay in memory and are best assessed in true real time. Experience
sampling has probably had the most impact on the psychology of emotion.
Emotions are highly variable—they ebb and flow throughout the day in
response to changing internal and external events (Mehl & Conner, 2012,
Ch. 27). Experience sampling is ideally suited to capturing this changing
profile—an emotional “signature” that can reveal dynamic aspects of
functioning obscured by standard one-time surveys. Experience sampling
has revealed diurnal and weekly patterns in emotion, individual differences
in affective instability as a marker of psychopathology, differences in
the structure of emotional experience, divergences between experienced
versus remembered emotions, and covariation between emotions and
health-related factors. Experience sampling has also been used to identify
the emotional correlates of psychopathology including heightened affective
instability, anhedonia, and the emotional precursors to self-harm (Mehl &
Conner, 2012, Ch. 23).
Currently, there are three main approaches to experience sampling with
mobile phones. In each of these approaches, the trend is towards using people’s own phones to allow for seamless participation without the need for
an extra specialized device. One very simple approach is to send questions
via SMS text messaging. Texts can be scheduled and sent automatically
through most commercial SMS companies (e.g., www.message-media.com).
Participants reply to the questions contained in the text using numbers
on their keypad and the data can be downloaded from the SMS company
server at the end of the study. Although this approach is simple and does
not require a smartphone (mobile phone with Internet capability), it is also
the least flexible because of limited timing controls, lack of branching, and
restricted space for questions. A second approach is to send a hyperlink
to an online survey via SMS text messaging to participants with Internet
enabled smartphones. Here, participants receive a text message with a
hyperlink that directs them to a mobile-ready Internet survey. The survey
can be developed through any number of companies such as SurveyMonkey,
and the hyperlink to the online survey can be sent through a commercial
SMS company or a specialized service designed for experience sampling
(e.g., www.surveysignal.com). A third approach is to use application-based
smartphone tools. Here, participants download a smartphone app that
delivers a specialized survey to their smartphones. Although these apps and
surveys can be designed from the ground up, there are a growing number of
companies that provide assistance with survey development at a reasonable
charge (mEMA, iHabit, iForm, ISurvey, MovisensXS, and Qualtrics). There
6
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
are also a growing number of open-source development tools (Paco and
Funf). One issue with app-based experience sampling is that apps and
surveys are often designed for only one operating system (e.g., Apple’s iOS
or Google’s Android) that places limits on recruitment and participation.
However, increasingly, tools are being developed for both platforms.
AMBULATORY ASSESSMENT OF BEHAVIOR
Although self-reports are important to social science, oftentimes people
cannot or sometimes might not want to accurately report what they do. In
these circumstances, the direct—and ideally nonreactive—assessment of
real-world behavior is of high importance. For example, Mehl and colleagues
have developed the electronically activated recorder or EAR methodology
that allows for the relatively unobtrusive naturalistic observation of participants’ acoustic behavior in daily life (Mehl, Pennebaker, Crow, Dabbs,
& Price, 2001). The current EAR system, the “iEAR,” consists of a free iOS
app that runs on iPod touch and iPhone devices. Participants carry an iEAR
device on them as they go about their normal lives. The app periodically
records snippets of ambient sounds (e.g., 30 s every 12 min) thereby creating
a series of sound bites that, together, amount to acoustic logs of participants’
days as they naturally unfold. The ambient sound recordings are later
securely downloaded, reviewed by participants, and then coded for aspects
of participants’ momentary locations (e.g., in a public or private place),
activities (e.g., watching TV, and eating), interactions (e.g., along, in a group,
and on the phone), and emotional expressions (e.g., laughing and sighing).
Initial EAR research focused on the psychometric properties of naturalistically observed daily social behavior. This research showed (i) that a
broad spectrum of behaviors can be assessed reliably and with low levels
of reactivity from the sampled ambient sounds, (ii) that these behaviors
show large between-person variability and good temporal stability, and (iii)
that they have good convergent validity with theoretically related measures
(e.g., Big Five personality dimensions) (Mehl & Conner, 2012, Ch. 10). The
second generation studies, then, focused on the EAR’s potential to address
questions that are difficult to answer with other methods. For example, in
a cross-cultural study, Ramirez-Esparza and her colleagues used the EAR
method to study self-reported sociability in relation to observed sociability
in the United States and Mexico. They found that although American participants rated themselves significantly higher than Mexicans on the question “I
see myself as a person who is talkative,” they actually spent almost 10% less
time talking (Ramírez-Esparza, Mehl, Álvarez-Bermúdez, & Pennebaker,
2009). In a similar way, Mehl and his colleagues used the EAR method to
debunk the long-standing myth that women are by a factor more talkative
Ambulatory Assessment: Methods for Studying Everyday Life
7
than men (Mehl, Vazire, Ramirez-Esparza, Slatcher, & Pennebaker, 2007).
Using data from six studies, they showed that both sexes use on average
about 16,000 words per day. Together, these studies showed how the EAR
method can be used to study objective aspects of daily behavior and how it
can yield results that diverge from findings obtained with other methods.
A series of other creative ways for assessing behavior directly and unobtrusively in the real world have been developed. For example, time-lapse
photography has been used to study the flow of people and the use of
space in urban public places (Whyte, 1980). In modern studies, participants’
movement and location are tracked via actigraphy and GPS information. To
determine sleep patterns and circadian rhythms, studies have participants
wear small, rugged wrist watches that log body movements along with
day-night (i.e., light) patterns (Van de Water, Holmes, & Hurley, 2011). Multichannel activity monitoring devices provide more detailed information on
posture and motion through the placement of small accelerometer sensors on different body locations (e.g., arm, leg, or waist). Classification
algorithms then convert the raw sensor input into discrete posture (e.g.,
lying, and sitting, and standing) and motion (e.g., walking, cycling, and
driving) patterns. Importantly, validation studies have consistently found
critical discrepancies between self-reported and objective activity records
(Mehl & Conner, 2012, Ch. 13). Finally, location-tracking via either dedicated GPS devices or smartphones with GPS and Wi-Fi sensors are on the
way of becoming mainstream in the social sciences (Montoliu, Blom, &
Gatica-Perez, 2013; Wolf & Jacobs, 2010). Although these tools currently exist
as stand-alone assessment devices, in the future, they will be integrated into
mobile devices that people naturally carry with them which will allow more
seamless integrated assessment (Miller, 2012).
AMBULATORY ASSESSMENT OF PHYSIOLOGY
Finally, ambulatory assessment methods also exist for the sampling of
physiological activity in everyday life. An array of biosignals can now be
measured reliably via portable signal recording devices (e.g., electrocardiogram, blood pressure, electrodermal activity, and body temperature)
(Wilhelm & Grossman, 2010). Recently researchers have added ambulatory
assessment of hormones and other biomarkers to the list (Mehl & Conner,
2012, Ch. 11). As an example of research that implemented traditional ambulatory physiological monitoring, Lane, Zareba, Reis, Peterson, and Moss
(2011b) used experience sampling combined with ambulatory electrocardiography (a so-called Holter monitor) to show that daily emotions—even at
low intensities—triggered abnormal cardiac activity among patients with a
congenital heart abnormality. In a classic study on hormonal responses in
8
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
daily life, Smyth et al. (1998) combined experience sampling with momentary
assessment of cortisol. They found that momentary reports of current or
anticipated stress predicted increased cortisol secretion 20 min later.
Taken together, these two examples illustrate how ambulatory physiological monitoring has been used to link mundane and seemingly inconsequential experiences in our daily lives to objective physiological responses. The
development of novel ways to track what goes on underneath our skins as
we go about our lives is a rapidly advancing field and important advances
can be expected in the future (Kim et al., 2011).
CUTTING-EDGE RESEARCH AND FUTURE DIRECTIONS
As the mobile device revolution is unfolding around us, it is clear that ambulatory assessment will, over time, be revolutionized by it. Smartphones will
not just be devices for everyday communication but will also become devices
for large-scale scientific data collection and intervention (Kaplan & Stone,
2013; Yarkoni, 2012). They automatically store vast amounts of real-world
user interaction data and are equipped with an array of high quality sensors
to track the physical (e.g., location and position) and social (e.g., blue tooth
connections) context of these interactions. Finally, with add-on sensors, they
will be able monitor physiological parameters. In a visionary article, Miller
(2012) states, “the question is not whether smartphones will revolutionize
psychology but how, when, and where the revolution will happen” (p. 234).
SMARTPHONE SENSING
One flourishing research area at the intersection of the social and computer
sciences is the development of “smartphone sensing” applications. The idea
behind these applications is to make inferences about users’ emotions, behavior, environments, and life patterns through computational integration of the
data produced by (i) interactions with the user interface (e.g., timing and
duration of phone calls number text messages) and (ii) the multiple sensors embedded in smartphones (e.g., Bluetooth, GPS, and accelerometer).
For example, de Montjoye, Quoidbach, Robic, and Pentland (2013) recently
showed that the personality of smartphone users (e.g., extraversion and neuroticism) can be predicted with high levels of accuracy from information that
is routinely part of the data logs of mobile phone carriers (e.g., number of
interactions, number and diversity of contacts, response latency to events,
and distance traveled).
Lu et al. (2012) have applied this idea to automatic voice-based stress
detection via smartphones. Drawing on prior stress research, the so-called
StressSense app monitors ambient sounds for voices, performs speaker
Ambulatory Assessment: Methods for Studying Everyday Life
9
separation, and extracts stress-relevant voice parameters (e.g., speech rate,
pitch variability, and jitter). These parameters are then integrated into
stress-level estimates using machine learning algorithms that are trained
with the user’s galvanic skin response as the “ground truth” of how stressed
the user really is. The authors report high classification accuracy for both
outdoor and indoor environments. In a similar way, Rachuri et al. (2010) have
been developing a mobile phone application for the automatic recognition
of discrete emotions. Their “EmotionSense” app operates by extracting
voice parameters and comparing them against an internal “emotion prosody
library” that is derived from voice feature analysis of enacted target emotions
(happy sad, fearful, angry, neutral).
Finally, in an intriguing study, Lane et al. (2011a) report the development of
“BeWell” as a smartphone application to promote healthy lifestyles. The app
continuously monitors users’ physical activity (via the embedded accelerometer), sleep activity (via the accelerometer and recharging information), and
social activity (via ambient sounds containing voice). In a second step, it compares the estimated levels against established health recommendations (e.g.,
ideal value of 7 h of sleep). In a third step, the app feeds the results back
to the user intuitively on the display where it visualizes a person’s wellness
through an aquarium with swimming fish—the vitality of which reflects the
state of wellness. Because all computations are run directly on the phone, the
app is self-sufficient but currently absorbs a high amount of processing time
and battery life. As a proof-of-concept study, though, it shows a powerful
application of mobile-phone based social or “life-style” sensing.
“BIG DATA” COLLECTION
Future progress in this area is also tied into a rapidly decreasing per-person
cost thereby allowing data collection at large-scale levels. Already we
are beginning to see studies with “big data” from thousands of people.
For example, one group of researchers analyzed Geographical Positional
System signals from 100,000 mobile phone users over a 6-month period to
show reproducible regularities in their within-person movement patterns
(Gonzalez, Hidalgo, & Barabasi, 2008). Other research uses experience sampling tools made available to a wide audience in exchange for scientific use
of their anonymized data (e.g., Mappiness, Trackyourhappiness, EmotionSense, and Happathon). For example, data from 2250 users of TrackYourHappiness was used to show the conditions and contexts in which people report
being happier, such as when they are social and not “mind-wandering”
(Killingsworth & Gilbert, 2010). Likewise, Mappiness data from nearly 22,000
users in the United Kingdom found that people reported greater happiness
when they were located near natural environments as determined by GPS
10
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
location (MacKerron & Mourato, 2013). Other big data projects include the
Gallup-Healthways Well-Being Index, an American-based population-based
phone survey of over 1,000,000 people that includes a daily mood measure,
the World Well-Being Project, which analyzes language in Facebook and
Twitters posts to index differences in psychological states, and large-scale
mining of Twitter data (e.g., Golder & Macy, 2011)
MOBILE HEALTH: LARGE SCALE SAMPLING OF ELECTRONIC HEALTH INFORMATION
Large scale ambulatory assessment will also transform health research. The
E-Heart Study at the University of California San Francisco is a creative
new project that aims to collect ambulatory heart health data from one
million people. Tools in the study include mobile phone surveys, mobile
apps, and special sensors integrate with participants’ smartphones to provide real-time health recordings (heart rate, blood pressure, activity, sleep
quality, etc.). Their goal is to capitalize on big data to “develop strategies to
prevent and treat all aspects of heart disease” (https://www.health-eheartstudy.org/study). Other large mobile health projects include the National
Experience Sampling Project, which aims to collect ambulatory health data
at the population level. Future progress in this area will also benefit from
disposable wireless biometric patches that can be worn continuously.
KEY ISSUES GOING FORWARD IN AMBULATORY ASSESSMENT
Two of the most pressing issues going forward concern (i) the protection of
participants’ privacy and (ii) the synthesis of the immense amount of digital
data. There is little doubt that the Internet and, most importantly, online
social networking has already dramatically changed notions of privacy in
people’s personal lives. About a decade ago, it was ethically questionable to
“Google” someone before a date. Now, Facebook users readily post private
pictures of and intimate comments about their lives to hundreds of online
friends. For maximizing the capabilities of their mobile phones, people also
accept the complete, centralized tracking of their locations, browsing and
search history, and entertainment choices (e.g., iTunes, Netflix, YouTube, and
Kindle). These changes in how private information is shared are bound to
affect perceptions of what data is acceptable to collect for scientific purposes.
At the same time, though, these changes have profound implications for
the confidentiality of scientific data and the protection of participants’
privacy. King (2011) pointed out that it is de facto impossible to guarantee
anonymity by combining the three demographic variables date of birth,
gender, and zip code. In a similar way—and directly in the context of mobile
data collection—Kosinski, Stillwell, and Graepel (2013) recently showed
Ambulatory Assessment: Methods for Studying Everyday Life
11
that highly private and often stigmatizing characteristics such as sexual
orientation, ethnicity, and religious and political affiliation can be readily
predicted from only one type of digital data, “Likes” in users’ Facebook
profiles. The same was true for important health behaviors such as smoking,
drinking, and drug use. Combined with the scientifically desirable trend
towards data sharing and making (taxpayer-funded) data bases publically
available and advances in large-scale “big data” mining, it is clear that the
ambulatory assessment researchers, and the scientific community more
generally, have to develop new guidelines and methods of protecting the
privacy of human subjects.
Researchers will also need to develop better strategies for handling large
amounts of data. Ambulatory assessment data is already quite large and
requires specialized tools for treating the nested data structure such as multilevel modeling. However, big data will increase the size and complexity of
these data structures exponentially. Such data will require different analytic
approaches that likely draw on techniques from bioinformatics and computer science. Yarkoni (2012) calls this new approach “psychoinformatics,”
which can include tools such as network analysis, large-scale exploratory
data analysis, and a greater reliance on more flexible open-source statistical
software such as R. This requirement for greater statistical sophistication will
require new forms of training and greater collaboration among statisticians,
computer scientists, and psychological scientists.
CONCLUSION
The separation between science and everyday life will become narrower with
each decade as ambulatory assessment becomes integrated seamlessly into
people’s lives. Although ambulatory assessment will continue to complement rather than replace controlled laboratory science, it will begin to play
a larger role in science than it has in the past especially as findings from
population-based data-sets begin to bear fruit. Issues of privacy and data
management notwithstanding, the future of ambulatory assessment future
will be a dynamic, collective, and collaborative process.
REFERENCES
Bolger, N., & Schilling, E. A. (1991). Personality and the problems of everyday life:
The role of neuroticism in exposure and reactivity to daily stressors. Journal of Personality, 59, 355–386.
Conner, T. S., & Barrett, L. F. (2012). Trends in ambulatory self-reports: The role
of momentary experience in psychosomatic medicine. Psychosomatic Medicine, 74,
327–337. doi:10.1097/PSY.0b013e3182546f18
12
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
de Montjoye, Y.-A., Quoidbach, J., Robic, F., & Pentland, A. S. (2013). Predicting personality using novel mobile phone-based metrics. Social Computing,
Behavioral-Cultural Modeling and Prediction, 7812, 48–55.
Golder, S. A., & Macy, M. W. (2011). Diurnal and seasonal mood vary with work,
sleep, and daylength across diverse cultures. Science, 333, 1878–1881. doi:10.1126/
science.1202775
Gonzalez, M. C., Hidalgo, C. A., & Barabasi, A. L. (2008). Understanding individual
human mobility patterns. Nature, 453, 779–782. doi:10.1038/nature06958
Kaplan, R. M., & Stone, A. A. (2013). Bringing the laboratory and clinic to the community: Mobile technologies for health promotion and disease prevention. Annual
Review of Psychology, 64, 471–498. doi:10.1146/annurev-psych-113011-143736
Killingsworth, M. A., & Gilbert, D. T. (2010). A wandering mind is an unhappy mind.
Science, 330, 932. doi:10.1126/science.1192439
Kim, D.-H., Lu, N., Ma, R., Kim, Y.-S., Kim, R.-H., Wang, S., … Islam, A. (2011).
Epidermal electronics. Science, 333(6044), 838–843. doi:10.1016/j.wneu.2011.
10.001
King, G. (2011). Ensuring the data-rich future of the social sciences. Science, 331(6018),
719–721. doi:10.1126/science.1197872
Kosinski, M., Stillwell, D., & Graepel, T. (2013). Private traits and attributes are
predictable from digital records of human behavior. Proceedings of the National
Academy of Sciences of the United States of America. doi:10.1073/pnas.1218772110
Lane, N. D., Mohammod, M., Lin, M., Yang, X., Lu, H., Ali, S., … Campbell, A.
(2011a). Bewell: A smartphone application to monitor, model and promote wellbeing.
Paper presented at the 5th International Conference on Pervasive Computing
Technologies for Healthcare (PervasiveHealth2011).
Lane, R. D., Zareba, W., Reis, H. T., Peterson, D. R., & Moss, A. J. (2011b). Changes
in ventricular repolarization duration during typical daily emotion in patients
with long QT syndrome. Psychosomatic Medicine, 73(1), 98–105. doi:10.1097/
PSY.0b013e318203310a
Lenderking, W. R., Hu, M., Tennen, H., Cappelleri, J. C., Petrie, C. D., & Rush, A. J.
(2008). Daily process methodology for measuring earlier antidepressant response.
Contemporary Clinical Trials, 29, 867–877. doi:10.1016/j.cct.2008.05.012
Lu, H., Frauendorfer, D., Rabbi, M., Mast, M. S., Chittaranjan, G. T., Campbell, A.
T., … Choudhury, T. (2012). StressSense: detecting stress in unconstrained acoustic
environments using smartphones. Paper presented at the Proceedings of the 2012
ACM Conference on Ubiquitous Computing.
MacKerron, G., & Mourato, S. (2013). Happiness is greater in natural environments.
Global Environmental Change, 23(5), 992–1000.
Mehl, M. R., & Conner, T. S. (2012). Handbook of research methods for studying daily life.
New York, NY: Guilford Press.
Mehl, M. R., Pennebaker, J. W., Crow, D. M., Dabbs, J., & Price, J. H. (2001). The
electronically activated recorder (EAR): A device for sampling naturalistic daily
activities and conversations. Behavior Research Methods, Instruments, and Computers,
33(4), 517–523.
Ambulatory Assessment: Methods for Studying Everyday Life
13
Mehl, M. R., Vazire, S., Ramirez-Esparza, N., Slatcher, R. B., & Pennebaker, J. W.
(2007). Are women really more talkative than men? Science, 317(5834), 82.
doi:10.1126/science.1139940
Miller, G. (2012). The smartphone psychology manifesto. Perspectives on Psychological
Science, 7(3), 221–237. doi:10.1177/1745691612441215
Montoliu, R., Blom, J., & Gatica-Perez, D. (2013). Discovering places of interest
in everyday life from smartphone data. Multimedia Tools and Applications, 62(1),
179–207. doi:10.1007/s11042-011-0982-z
Rachuri, K. K., Musolesi, M., Mascolo, C., Rentfrow, P. J., Longworth, C., & Aucinas,
A. (2010). EmotionSense: A mobile phones based adaptive platform for experimental social
psychology research. Paper presented at the Proceedings of the 12th ACM international conference on Ubiquitous computing.
Ramírez-Esparza, N., Mehl, M., Álvarez-Bermúdez, J., & Pennebaker, J. (2009). Are
Mexicans more or less sociable than Americans? Insights from a naturalistic
observation study. Journal of Research in Personality, 43(1), 1–7. doi:10.1016/
j.jrp.2008.09.002
Robinson, M. D., & Clore, G. L. (2002). Belief and feeling: Evidence for an
accessibility model of emotional self-report. Psychological Bulletin, 128, 934–960.
doi:10.1037/0033-2909.128.6.934
Smyth, J., Ockenfels, M. C., Porter, L., Kirschbaum, C., Hellhammer, D. H., & Stone,
A. A. (1998). Stressors and mood measured on a momentary basis are associated with salivary cortisol secretion. Psychoneuroendocrinology, 23(4), 353–370.
doi:10.1016/s0306-4530(98)00008-0
Van de Water, A. T., Holmes, A., & Hurley, D. A. (2011). Objective measurements
of sleep for non-laboratory settings as alternatives to polysomnography—A
systematic review. Journal of Sleep Research, 20(1 Pt 2), 183–200. doi:10.1111/
j.1365-2869.2009.00814.x
Walls, T. A., & Schafer, J. L. (2006). Models for intensive longitudinal data. New York,
NY: Oxford University Press.
Whyte, W. H. (1980). The social life of small urban spaces. Washington, D.C.: The Conservation Foundation.
Wilhelm, F. H., & Grossman, P. (2010). Emotions beyond the laboratory: Theoretical fundaments, study design, and analytic strategies for advanced ambulatory
assessment. Biological Psychology. doi:10.1016/j.biopsycho.2010.01.017
Wolf, P. S., & Jacobs, W. J. (2010). GPS technology and human psychological research:
A methodological proposal. Journal of Methods and Measurement in the Social Sciences, 1(1), 1–7.
Yarkoni, T. (2012). Psychoinformatics: New horizons at the interface of the psychological and computing sciences. Current Directions in Psychological Science, 21(6),
391–397. doi:10.1177/0963721412457362
FURTHER READING
Bolger, N., & Laurenceau, J.-P. (2013). Intensive longitudinal methods: An introduction
to diary and experience sampling research. New York, NY: Guilford Press.
14
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
Hektner, J. M., Schmidt, J. A., & Csikszentmihalyi, M. (2007). Experience sampling
method: Measuring the quality of everyday life. Thousand Oaks, CA: SAGE Publications.
Mehl, M. R., & Conner, T. S. (2012). Handbook of research methods for studying daily life.
New York, NY: Guilford Press.
Miller, G. (2012). The smartphone psychology manifesto. Perspectives on Psychological
Science, 7(3), 221–237. doi:10.1177/1745691612441215
Stone, A., Shiffman, S., Atienza, A., & Nebeling, L. (Eds.) (2007). The science of real-time
data capture: Self-reports in health research. New York, NY: Oxford University Press.
Yarkoni, T. (2012). Psychoinformatics: New horizons at the interface of the psychological and computing sciences. Current Directions in Psychological Science, 21(6),
391–397. doi:10.1177/0963721412457362
TAMLIN S. CONNER SHORT BIOGRAPHY
Tamlin S. Conner, PhD is Senior Lecturer in Psychology at the University
of Otago in New Zealand. She received her doctorate in social psychology
from Boston College and completed postdoctoral training in health and
personality psychology at the University of Connecticut Health Center. She
has published numerous articles on the theory and practice of experience
sampling, is a leading expert on ambulatory self-report techniques, and
conducts research on well-being, emotions, and the science of self-report.
Recently, she coedited the Handbook of Research Methods for Studying Daily
Life (2012; Guilford Press) with Matthias Mehl. She is a founding member and current secretary of the Society for Ambulatory Assessment.
http://www.otago.ac.nz/psychology/staff/tamlinconner.html
MATTHIAS R. MEHL SHORT BIOGRAPHY
Matthias R. Mehl, PhD, is Associate Professor of Psychology at the
University of Arizona. He received his doctorate in social and personality
psychology from the University of Texas at Austin. Over the past decade, he
developed the electronically activated recorder (EAR) as a novel methodology for the unobtrusive naturalistic observation of daily life. He has given
workshops and published numerous articles on novel methods for studying
daily life and recently coedited the Handbook of Research Methods for Studying
Daily Life (2012; Guilford Press) with Tamlin Conner. He is a founding
member and the current Vice President of the Society for Ambulatory
Assessment. In 2011, the Association for Psychological Science identified
him as a “Rising Star.” http://dingo.sbs.arizona.edu/∼mehl/
Ambulatory Assessment: Methods for Studying Everyday Life
15
RELATED ESSAYS
Social Epigenetics: Incorporating Epigenetic Effects as Social Cause and
Consequence (Sociology), Douglas L. Anderton and Kathleen F. Arcaro
To Flop Is Human: Inventing Better Scientific Approaches to Anticipating
Failure (Methods), Robert Boruch and Alan Ruby
Repeated Cross-Sections in Survey Data (Methods), Henry E. Brady and
Richard Johnston
The Evidence-Based Practice Movement (Sociology), Edward W. Gondolf
Meta-Analysis (Methods), Larry V. Hedges and Martyna Citkowicz
The Use of Geophysical Survey in Archaeology (Methods), Timothy J.
Horsley
Network Research Experiments (Methods), Allen L. Linton and Betsy Sinclair
Longitudinal Data Analysis (Methods), Todd D. Little et al.
Data Mining (Methods), Gregg R. Murray and Anthony Scime
Remote Sensing with Satellite Technology (Archaeology), Sarah Parcak
Quasi-Experiments (Methods), Charles S. Reichard
Digital Methods for Web Research (Methods), Richard Rogers
Virtual Worlds as Laboratories (Methods), Travis L. Ross et al.
Modeling Life Course Structure: The Triple Helix (Sociology), Tom Schuller
Content Analysis (Methods), Steven E. Stemler
Person-Centered Analysis (Methods), Alexander von Eye and Wolfgang
Wiedermann
Translational Sociology (Sociology), Elaine Wethington
-
Ambulatory Assessment: Methods
for Studying Everyday Life
TAMLIN S. CONNER and MATTHIAS R. MEHL
Abstract
Ambulatory assessment is a class of methods that use mobile technology to understand people’s biopsychosocial processes in natural settings, in real time, and on
repeated occasions. In this essay, we discuss the rationale for ambulatory assessment
including the benefits of measuring people in the real world (greater ecological
validity, better understanding of people in contexts), in real time (avoidance of
memory bias, greater sensitivity for capturing change), and over time (capturing
within-person patterns and temporal trends). Then, we review the latest ambulatory
assessment techniques for measuring experiences, behaviors, and physiology in
daily life. Experiences such as emotions, physical pain, and daily stressors can be
tracked using daily diaries and smartphone-based experience sampling. Behaviors
such as activity, movement, location, and natural language use can be tracked using
accelerometers, portable actigraphs, global positioning system (GPS) coordinates,
and the electronically activated recorder (EAR). Physiological processes such as
heart rate, blood pressure, and electrodermal activity can be measured using an
array of ambulatory biosensors. Ambulatory assessment will continue to be revolutionized by smartphones, which are becoming integrated seamlessly into people’s
lives. Emerging trends include social sensing applications that make inferences
about users’ psychological processes based on multi-channel information collected
from smartphones, emergence of “big data collection” whereby ambulatory
assessment data is gathered en masse from large populations, and the growing field
of mobile health. These trends raise questions around the protection of participants’
privacy and the synthesis of immense amounts of digital data. Ultimately, these
developments will narrow the separation between science and everyday life as
ambulatory assessment becomes an integrated part of people’s mobile lives.
INTRODUCTION: DEFINITION OF AND RATIONALE FOR
AMBULATORY ASSESSMENT
Laboratory-based methods have historically been the strength and pride
of the social and behavioral sciences. Most research method textbooks
emphasize how laboratory research helps control confounding variables
and thereby allows for the isolation of a causal factor of interest. Laboratory
Emerging Trends in the Social and Behavioral Sciences. Edited by Robert Scott and Stephen Kosslyn.
© 2015 John Wiley & Sons, Inc. ISBN 978-1-118-90077-2.
1
2
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
research has contributed enormously to our understanding of human
social behavior. Ultimately, though, laboratory research can exclusively
accumulate knowledge on what can happen—under isolated and controlled
circumstances—and it cannot speak to what does happen under the circumstances that people normally encounter in their everyday lives. Therefore, a
comprehensive science of human behavior also requires “thinking outside
the experimental box” and necessitates the study of humans in their natural
habitat—that is, the collection of data from individuals as they live their
lives in their daily environments.
This essay reviews existing methodologies for studying experiences,
behavior, and physiology in daily life (for a comprehensive review, see
Mehl & Conner, 2012). We here refer to these methodologies collectively as
ambulatory assessment methods. According to the Society for Ambulatory
Assessment, Ambulatory Assessment “comprises the use of field methods
to assess the ongoing behavior, physiology, experience and environmental
aspects of humans or nonhuman primates in naturalistic or unconstrained
settings. [It] designates an ecologically relevant assessment perspective that
aims at understanding biopsychosocial processes as they naturally unfold
in time and in context” (www.ambulatory-assessment.org). Other common
names to denote these methodologies include experience sampling methods,
diary methods, and ecological momentary assessment. At their core, these
methods allow researchers to study individuals (i) in their natural settings, (ii)
in real time (or close to real time), and (iii) on repeated occasions. Ambulatory
assessment derives its scientific rationale directly from these three measurement characteristics (i.e., “real-world,” “real-time,” “within-person”).
The “real-world” quality allows unprecedented access to what actually
happens in a person’s everyday life and the contexts that surround such
events (Mehl & Conner, 2012, Ch. 1). For example, ambulatory assessment
is a great way to track how much people are exercising every day and the
conditions that make exercise more or less likely. This real-world quality
also enables researchers to capitalize on the power of evocative daily events
that are difficult, if not unethical, to mimic in controlled laboratory settings
such as positive or negative daily interactions, disagreements with a friend,
issues with children, commuting, excessive smoking, or drinking. Moreover,
how people respond to a standardized laboratory stressor such as giving a
speech may differ from how they respond to stressful experiences in their
lives. Wilhelm and Grossman (2010) describe a participant who showed
rather minimal heart rate increases in response to a laboratory stress protocol
but quite dramatic heart rate increases later in the afternoon while watching
a soccer game at home. In addition, in a reverse pattern, patients often have
high blood pressure readings in the physician’s office but not in their home
environment—a phenomenon called the white coat hypertension. By testing
Ambulatory Assessment: Methods for Studying Everyday Life
3
behavior in real-world settings, researchers have greater confidence about
the ecological validity of findings—that is, whether an effect is representative
of how it operates in reality.
The “real-time” quality of ambulatory assessment enables researchers to
understand experiences as they occur, not how people represent them later
from memory. This quality is especially important for experiences that are
fleeting and often misremembered (e.g., emotions, and pain). A person’s
memory for how they “felt over the past week” or “the past month” is
influenced by a host of factors including salient experiences, mood at the
time of recall, personality traits, beliefs about the acceptability of certain
emotions, gender stereotypes, and cultural norms (Mehl & Conner, 2012,
Ch. 2). In fact, there is strong evidence that as the time delay between experience and reporting lengthens, self-reports become more stable, reflective,
belief-driven, and culturally homogenized rather than malleable, reflexive,
experience-driven, and individualized (Robinson & Clore, 2002). This distinction between real-time and recalled reporting can make a big difference
in understanding psychological processes. For one, ambulatory approaches
may be more sensitive for capturing change in emotional states in response
to interventions. In a trial of antidepressant medication treatment, changes in
depressive symptomology were detected earlier among patients randomly
assigned to track their symptoms each day for 30 days compared to patients
who reported their symptoms using standard one-week recall measures
(Lenderking et al., 2008). Other studies find stronger links between real-time
reports of emotion and atherosclerosis risk, immune system function,
and genetic vulnerability [reviewed by Conner and Barrett (2012)]. These
examples illustrate how-real time measurement may be more sensitive than
traditional memory-based measures under certain circumstances.
The “within-person” element of ambulatory assessment refers to the capacity
to identify patterns of behavior within the person across time (i.e., idiographic assessment) (Mehl & Conner, 2012, Ch. 3). This approach differs from
traditional cross-sectional research, that aims to identify patterns of behavior
between people (i.e., nomothetic assessment). Unlike cross sectional research
which observes people at a single time point, ambulatory assessment tracks
individuals’ experiences, behaviors, or physiology intensively over time,
yielding “intensive longitudinal data” (Walls & Schafer, 2006). That data can
be analyzed to uncover the temporal patterns and trends in each person’s
data, as well as commonalities and differences in these trends across people.
For example, in a classic study, Bolger and Schilling (1991) examined the
within-person relationships between daily stressors and anxiety for 339
people and found stronger within-person relationships for people higher
in Neuroticism. Ambulatory assessment has been used to uncover a wide
variety of within-person relationships, including the antecedents of heavy
4
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
drinking, triggers of smoking cravings, relationships between stress and
coping, and links between food consumption and mood.
FOUNDATIONAL RESEARCH: AMBULATORY ASSESSMENT
METHODS
AMBULATORY ASSESSMENT OF EXPERIENCE
One of the most common uses of ambulatory assessment is for studying
self-reported experiences in naturalistic settings. Experiences include current
mood states, levels of perceived stress, feelings of bodily pain or discomfort, and other aspects of daily life that can be self-reported. Thus, ambulatory assessment of experience inherently involves asking people to respond
to questions about their experiences as they go about their daily lives. The
Internet and, especially smartphones have revolutionized this methodology
(Miller, 2012) and now represent the main interface for ambulatory self-report
techniques. In general, there are two main ways of measuring experience
through ambulatory assessment—through Internet daily diaries and mobile
phone-based experience sampling.
Daily diary methods are probably the most common form of ambulatory
assessment. Daily diary methods are now done mainly through Internet surveys in which participants access the survey each day (or night) for a week
to three weeks to answer questions about their experiences that day. Diaries
are often used to assess experiences and events that occur on a daily basis
and can be easily recalled at the end of each day such as daily hassles, social
activities, emotions felt that day, and health behaviors such as alcohol use,
smoking, and food consumption. Although this approach is neither “real
time” nor necessarily ambulatory, it is considered “near to real time” and an
improvement upon asking people to recall their experiences across the entire
week or longer. Although daily diary studies are widely applied, they have
probably had the most impact on the psychology of romantic relationships,
the role daily stressors in mental and physical health, and the correlates and
consequences of daily health behaviors such as alcohol use (Mehl & Conner,
2012, Ch. 8). In addition, daily diary techniques are fairly easy to implement.
Researchers can design a simple Internet survey through any number of companies such as SurveyMonkey, Qualtrics, or SurveyGizmo. Participants then
access that survey during a specified time period, and the data can be downloaded at the end of the study.
A second approach to measuring experience is through mobile-phone-based
experience sampling. Experience sampling is the process of randomly signaling a person (typically 4–8 times per day) to report on their experiences at
Ambulatory Assessment: Methods for Studying Everyday Life
5
that moment [see suggested reading by Hektner, Schmidt, and Csikszentmihalyi (2007)]. This approach is used to study fleeting and ongoing subjective
experiences such as mood, pain, and stress because these experiences are
quick to decay in memory and are best assessed in true real time. Experience
sampling has probably had the most impact on the psychology of emotion.
Emotions are highly variable—they ebb and flow throughout the day in
response to changing internal and external events (Mehl & Conner, 2012,
Ch. 27). Experience sampling is ideally suited to capturing this changing
profile—an emotional “signature” that can reveal dynamic aspects of
functioning obscured by standard one-time surveys. Experience sampling
has revealed diurnal and weekly patterns in emotion, individual differences
in affective instability as a marker of psychopathology, differences in
the structure of emotional experience, divergences between experienced
versus remembered emotions, and covariation between emotions and
health-related factors. Experience sampling has also been used to identify
the emotional correlates of psychopathology including heightened affective
instability, anhedonia, and the emotional precursors to self-harm (Mehl &
Conner, 2012, Ch. 23).
Currently, there are three main approaches to experience sampling with
mobile phones. In each of these approaches, the trend is towards using people’s own phones to allow for seamless participation without the need for
an extra specialized device. One very simple approach is to send questions
via SMS text messaging. Texts can be scheduled and sent automatically
through most commercial SMS companies (e.g., www.message-media.com).
Participants reply to the questions contained in the text using numbers
on their keypad and the data can be downloaded from the SMS company
server at the end of the study. Although this approach is simple and does
not require a smartphone (mobile phone with Internet capability), it is also
the least flexible because of limited timing controls, lack of branching, and
restricted space for questions. A second approach is to send a hyperlink
to an online survey via SMS text messaging to participants with Internet
enabled smartphones. Here, participants receive a text message with a
hyperlink that directs them to a mobile-ready Internet survey. The survey
can be developed through any number of companies such as SurveyMonkey,
and the hyperlink to the online survey can be sent through a commercial
SMS company or a specialized service designed for experience sampling
(e.g., www.surveysignal.com). A third approach is to use application-based
smartphone tools. Here, participants download a smartphone app that
delivers a specialized survey to their smartphones. Although these apps and
surveys can be designed from the ground up, there are a growing number of
companies that provide assistance with survey development at a reasonable
charge (mEMA, iHabit, iForm, ISurvey, MovisensXS, and Qualtrics). There
6
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
are also a growing number of open-source development tools (Paco and
Funf). One issue with app-based experience sampling is that apps and
surveys are often designed for only one operating system (e.g., Apple’s iOS
or Google’s Android) that places limits on recruitment and participation.
However, increasingly, tools are being developed for both platforms.
AMBULATORY ASSESSMENT OF BEHAVIOR
Although self-reports are important to social science, oftentimes people
cannot or sometimes might not want to accurately report what they do. In
these circumstances, the direct—and ideally nonreactive—assessment of
real-world behavior is of high importance. For example, Mehl and colleagues
have developed the electronically activated recorder or EAR methodology
that allows for the relatively unobtrusive naturalistic observation of participants’ acoustic behavior in daily life (Mehl, Pennebaker, Crow, Dabbs,
& Price, 2001). The current EAR system, the “iEAR,” consists of a free iOS
app that runs on iPod touch and iPhone devices. Participants carry an iEAR
device on them as they go about their normal lives. The app periodically
records snippets of ambient sounds (e.g., 30 s every 12 min) thereby creating
a series of sound bites that, together, amount to acoustic logs of participants’
days as they naturally unfold. The ambient sound recordings are later
securely downloaded, reviewed by participants, and then coded for aspects
of participants’ momentary locations (e.g., in a public or private place),
activities (e.g., watching TV, and eating), interactions (e.g., along, in a group,
and on the phone), and emotional expressions (e.g., laughing and sighing).
Initial EAR research focused on the psychometric properties of naturalistically observed daily social behavior. This research showed (i) that a
broad spectrum of behaviors can be assessed reliably and with low levels
of reactivity from the sampled ambient sounds, (ii) that these behaviors
show large between-person variability and good temporal stability, and (iii)
that they have good convergent validity with theoretically related measures
(e.g., Big Five personality dimensions) (Mehl & Conner, 2012, Ch. 10). The
second generation studies, then, focused on the EAR’s potential to address
questions that are difficult to answer with other methods. For example, in
a cross-cultural study, Ramirez-Esparza and her colleagues used the EAR
method to study self-reported sociability in relation to observed sociability
in the United States and Mexico. They found that although American participants rated themselves significantly higher than Mexicans on the question “I
see myself as a person who is talkative,” they actually spent almost 10% less
time talking (Ramírez-Esparza, Mehl, Álvarez-Bermúdez, & Pennebaker,
2009). In a similar way, Mehl and his colleagues used the EAR method to
debunk the long-standing myth that women are by a factor more talkative
Ambulatory Assessment: Methods for Studying Everyday Life
7
than men (Mehl, Vazire, Ramirez-Esparza, Slatcher, & Pennebaker, 2007).
Using data from six studies, they showed that both sexes use on average
about 16,000 words per day. Together, these studies showed how the EAR
method can be used to study objective aspects of daily behavior and how it
can yield results that diverge from findings obtained with other methods.
A series of other creative ways for assessing behavior directly and unobtrusively in the real world have been developed. For example, time-lapse
photography has been used to study the flow of people and the use of
space in urban public places (Whyte, 1980). In modern studies, participants’
movement and location are tracked via actigraphy and GPS information. To
determine sleep patterns and circadian rhythms, studies have participants
wear small, rugged wrist watches that log body movements along with
day-night (i.e., light) patterns (Van de Water, Holmes, & Hurley, 2011). Multichannel activity monitoring devices provide more detailed information on
posture and motion through the placement of small accelerometer sensors on different body locations (e.g., arm, leg, or waist). Classification
algorithms then convert the raw sensor input into discrete posture (e.g.,
lying, and sitting, and standing) and motion (e.g., walking, cycling, and
driving) patterns. Importantly, validation studies have consistently found
critical discrepancies between self-reported and objective activity records
(Mehl & Conner, 2012, Ch. 13). Finally, location-tracking via either dedicated GPS devices or smartphones with GPS and Wi-Fi sensors are on the
way of becoming mainstream in the social sciences (Montoliu, Blom, &
Gatica-Perez, 2013; Wolf & Jacobs, 2010). Although these tools currently exist
as stand-alone assessment devices, in the future, they will be integrated into
mobile devices that people naturally carry with them which will allow more
seamless integrated assessment (Miller, 2012).
AMBULATORY ASSESSMENT OF PHYSIOLOGY
Finally, ambulatory assessment methods also exist for the sampling of
physiological activity in everyday life. An array of biosignals can now be
measured reliably via portable signal recording devices (e.g., electrocardiogram, blood pressure, electrodermal activity, and body temperature)
(Wilhelm & Grossman, 2010). Recently researchers have added ambulatory
assessment of hormones and other biomarkers to the list (Mehl & Conner,
2012, Ch. 11). As an example of research that implemented traditional ambulatory physiological monitoring, Lane, Zareba, Reis, Peterson, and Moss
(2011b) used experience sampling combined with ambulatory electrocardiography (a so-called Holter monitor) to show that daily emotions—even at
low intensities—triggered abnormal cardiac activity among patients with a
congenital heart abnormality. In a classic study on hormonal responses in
8
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
daily life, Smyth et al. (1998) combined experience sampling with momentary
assessment of cortisol. They found that momentary reports of current or
anticipated stress predicted increased cortisol secretion 20 min later.
Taken together, these two examples illustrate how ambulatory physiological monitoring has been used to link mundane and seemingly inconsequential experiences in our daily lives to objective physiological responses. The
development of novel ways to track what goes on underneath our skins as
we go about our lives is a rapidly advancing field and important advances
can be expected in the future (Kim et al., 2011).
CUTTING-EDGE RESEARCH AND FUTURE DIRECTIONS
As the mobile device revolution is unfolding around us, it is clear that ambulatory assessment will, over time, be revolutionized by it. Smartphones will
not just be devices for everyday communication but will also become devices
for large-scale scientific data collection and intervention (Kaplan & Stone,
2013; Yarkoni, 2012). They automatically store vast amounts of real-world
user interaction data and are equipped with an array of high quality sensors
to track the physical (e.g., location and position) and social (e.g., blue tooth
connections) context of these interactions. Finally, with add-on sensors, they
will be able monitor physiological parameters. In a visionary article, Miller
(2012) states, “the question is not whether smartphones will revolutionize
psychology but how, when, and where the revolution will happen” (p. 234).
SMARTPHONE SENSING
One flourishing research area at the intersection of the social and computer
sciences is the development of “smartphone sensing” applications. The idea
behind these applications is to make inferences about users’ emotions, behavior, environments, and life patterns through computational integration of the
data produced by (i) interactions with the user interface (e.g., timing and
duration of phone calls number text messages) and (ii) the multiple sensors embedded in smartphones (e.g., Bluetooth, GPS, and accelerometer).
For example, de Montjoye, Quoidbach, Robic, and Pentland (2013) recently
showed that the personality of smartphone users (e.g., extraversion and neuroticism) can be predicted with high levels of accuracy from information that
is routinely part of the data logs of mobile phone carriers (e.g., number of
interactions, number and diversity of contacts, response latency to events,
and distance traveled).
Lu et al. (2012) have applied this idea to automatic voice-based stress
detection via smartphones. Drawing on prior stress research, the so-called
StressSense app monitors ambient sounds for voices, performs speaker
Ambulatory Assessment: Methods for Studying Everyday Life
9
separation, and extracts stress-relevant voice parameters (e.g., speech rate,
pitch variability, and jitter). These parameters are then integrated into
stress-level estimates using machine learning algorithms that are trained
with the user’s galvanic skin response as the “ground truth” of how stressed
the user really is. The authors report high classification accuracy for both
outdoor and indoor environments. In a similar way, Rachuri et al. (2010) have
been developing a mobile phone application for the automatic recognition
of discrete emotions. Their “EmotionSense” app operates by extracting
voice parameters and comparing them against an internal “emotion prosody
library” that is derived from voice feature analysis of enacted target emotions
(happy sad, fearful, angry, neutral).
Finally, in an intriguing study, Lane et al. (2011a) report the development of
“BeWell” as a smartphone application to promote healthy lifestyles. The app
continuously monitors users’ physical activity (via the embedded accelerometer), sleep activity (via the accelerometer and recharging information), and
social activity (via ambient sounds containing voice). In a second step, it compares the estimated levels against established health recommendations (e.g.,
ideal value of 7 h of sleep). In a third step, the app feeds the results back
to the user intuitively on the display where it visualizes a person’s wellness
through an aquarium with swimming fish—the vitality of which reflects the
state of wellness. Because all computations are run directly on the phone, the
app is self-sufficient but currently absorbs a high amount of processing time
and battery life. As a proof-of-concept study, though, it shows a powerful
application of mobile-phone based social or “life-style” sensing.
“BIG DATA” COLLECTION
Future progress in this area is also tied into a rapidly decreasing per-person
cost thereby allowing data collection at large-scale levels. Already we
are beginning to see studies with “big data” from thousands of people.
For example, one group of researchers analyzed Geographical Positional
System signals from 100,000 mobile phone users over a 6-month period to
show reproducible regularities in their within-person movement patterns
(Gonzalez, Hidalgo, & Barabasi, 2008). Other research uses experience sampling tools made available to a wide audience in exchange for scientific use
of their anonymized data (e.g., Mappiness, Trackyourhappiness, EmotionSense, and Happathon). For example, data from 2250 users of TrackYourHappiness was used to show the conditions and contexts in which people report
being happier, such as when they are social and not “mind-wandering”
(Killingsworth & Gilbert, 2010). Likewise, Mappiness data from nearly 22,000
users in the United Kingdom found that people reported greater happiness
when they were located near natural environments as determined by GPS
10
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
location (MacKerron & Mourato, 2013). Other big data projects include the
Gallup-Healthways Well-Being Index, an American-based population-based
phone survey of over 1,000,000 people that includes a daily mood measure,
the World Well-Being Project, which analyzes language in Facebook and
Twitters posts to index differences in psychological states, and large-scale
mining of Twitter data (e.g., Golder & Macy, 2011)
MOBILE HEALTH: LARGE SCALE SAMPLING OF ELECTRONIC HEALTH INFORMATION
Large scale ambulatory assessment will also transform health research. The
E-Heart Study at the University of California San Francisco is a creative
new project that aims to collect ambulatory heart health data from one
million people. Tools in the study include mobile phone surveys, mobile
apps, and special sensors integrate with participants’ smartphones to provide real-time health recordings (heart rate, blood pressure, activity, sleep
quality, etc.). Their goal is to capitalize on big data to “develop strategies to
prevent and treat all aspects of heart disease” (https://www.health-eheartstudy.org/study). Other large mobile health projects include the National
Experience Sampling Project, which aims to collect ambulatory health data
at the population level. Future progress in this area will also benefit from
disposable wireless biometric patches that can be worn continuously.
KEY ISSUES GOING FORWARD IN AMBULATORY ASSESSMENT
Two of the most pressing issues going forward concern (i) the protection of
participants’ privacy and (ii) the synthesis of the immense amount of digital
data. There is little doubt that the Internet and, most importantly, online
social networking has already dramatically changed notions of privacy in
people’s personal lives. About a decade ago, it was ethically questionable to
“Google” someone before a date. Now, Facebook users readily post private
pictures of and intimate comments about their lives to hundreds of online
friends. For maximizing the capabilities of their mobile phones, people also
accept the complete, centralized tracking of their locations, browsing and
search history, and entertainment choices (e.g., iTunes, Netflix, YouTube, and
Kindle). These changes in how private information is shared are bound to
affect perceptions of what data is acceptable to collect for scientific purposes.
At the same time, though, these changes have profound implications for
the confidentiality of scientific data and the protection of participants’
privacy. King (2011) pointed out that it is de facto impossible to guarantee
anonymity by combining the three demographic variables date of birth,
gender, and zip code. In a similar way—and directly in the context of mobile
data collection—Kosinski, Stillwell, and Graepel (2013) recently showed
Ambulatory Assessment: Methods for Studying Everyday Life
11
that highly private and often stigmatizing characteristics such as sexual
orientation, ethnicity, and religious and political affiliation can be readily
predicted from only one type of digital data, “Likes” in users’ Facebook
profiles. The same was true for important health behaviors such as smoking,
drinking, and drug use. Combined with the scientifically desirable trend
towards data sharing and making (taxpayer-funded) data bases publically
available and advances in large-scale “big data” mining, it is clear that the
ambulatory assessment researchers, and the scientific community more
generally, have to develop new guidelines and methods of protecting the
privacy of human subjects.
Researchers will also need to develop better strategies for handling large
amounts of data. Ambulatory assessment data is already quite large and
requires specialized tools for treating the nested data structure such as multilevel modeling. However, big data will increase the size and complexity of
these data structures exponentially. Such data will require different analytic
approaches that likely draw on techniques from bioinformatics and computer science. Yarkoni (2012) calls this new approach “psychoinformatics,”
which can include tools such as network analysis, large-scale exploratory
data analysis, and a greater reliance on more flexible open-source statistical
software such as R. This requirement for greater statistical sophistication will
require new forms of training and greater collaboration among statisticians,
computer scientists, and psychological scientists.
CONCLUSION
The separation between science and everyday life will become narrower with
each decade as ambulatory assessment becomes integrated seamlessly into
people’s lives. Although ambulatory assessment will continue to complement rather than replace controlled laboratory science, it will begin to play
a larger role in science than it has in the past especially as findings from
population-based data-sets begin to bear fruit. Issues of privacy and data
management notwithstanding, the future of ambulatory assessment future
will be a dynamic, collective, and collaborative process.
REFERENCES
Bolger, N., & Schilling, E. A. (1991). Personality and the problems of everyday life:
The role of neuroticism in exposure and reactivity to daily stressors. Journal of Personality, 59, 355–386.
Conner, T. S., & Barrett, L. F. (2012). Trends in ambulatory self-reports: The role
of momentary experience in psychosomatic medicine. Psychosomatic Medicine, 74,
327–337. doi:10.1097/PSY.0b013e3182546f18
12
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
de Montjoye, Y.-A., Quoidbach, J., Robic, F., & Pentland, A. S. (2013). Predicting personality using novel mobile phone-based metrics. Social Computing,
Behavioral-Cultural Modeling and Prediction, 7812, 48–55.
Golder, S. A., & Macy, M. W. (2011). Diurnal and seasonal mood vary with work,
sleep, and daylength across diverse cultures. Science, 333, 1878–1881. doi:10.1126/
science.1202775
Gonzalez, M. C., Hidalgo, C. A., & Barabasi, A. L. (2008). Understanding individual
human mobility patterns. Nature, 453, 779–782. doi:10.1038/nature06958
Kaplan, R. M., & Stone, A. A. (2013). Bringing the laboratory and clinic to the community: Mobile technologies for health promotion and disease prevention. Annual
Review of Psychology, 64, 471–498. doi:10.1146/annurev-psych-113011-143736
Killingsworth, M. A., & Gilbert, D. T. (2010). A wandering mind is an unhappy mind.
Science, 330, 932. doi:10.1126/science.1192439
Kim, D.-H., Lu, N., Ma, R., Kim, Y.-S., Kim, R.-H., Wang, S., … Islam, A. (2011).
Epidermal electronics. Science, 333(6044), 838–843. doi:10.1016/j.wneu.2011.
10.001
King, G. (2011). Ensuring the data-rich future of the social sciences. Science, 331(6018),
719–721. doi:10.1126/science.1197872
Kosinski, M., Stillwell, D., & Graepel, T. (2013). Private traits and attributes are
predictable from digital records of human behavior. Proceedings of the National
Academy of Sciences of the United States of America. doi:10.1073/pnas.1218772110
Lane, N. D., Mohammod, M., Lin, M., Yang, X., Lu, H., Ali, S., … Campbell, A.
(2011a). Bewell: A smartphone application to monitor, model and promote wellbeing.
Paper presented at the 5th International Conference on Pervasive Computing
Technologies for Healthcare (PervasiveHealth2011).
Lane, R. D., Zareba, W., Reis, H. T., Peterson, D. R., & Moss, A. J. (2011b). Changes
in ventricular repolarization duration during typical daily emotion in patients
with long QT syndrome. Psychosomatic Medicine, 73(1), 98–105. doi:10.1097/
PSY.0b013e318203310a
Lenderking, W. R., Hu, M., Tennen, H., Cappelleri, J. C., Petrie, C. D., & Rush, A. J.
(2008). Daily process methodology for measuring earlier antidepressant response.
Contemporary Clinical Trials, 29, 867–877. doi:10.1016/j.cct.2008.05.012
Lu, H., Frauendorfer, D., Rabbi, M., Mast, M. S., Chittaranjan, G. T., Campbell, A.
T., … Choudhury, T. (2012). StressSense: detecting stress in unconstrained acoustic
environments using smartphones. Paper presented at the Proceedings of the 2012
ACM Conference on Ubiquitous Computing.
MacKerron, G., & Mourato, S. (2013). Happiness is greater in natural environments.
Global Environmental Change, 23(5), 992–1000.
Mehl, M. R., & Conner, T. S. (2012). Handbook of research methods for studying daily life.
New York, NY: Guilford Press.
Mehl, M. R., Pennebaker, J. W., Crow, D. M., Dabbs, J., & Price, J. H. (2001). The
electronically activated recorder (EAR): A device for sampling naturalistic daily
activities and conversations. Behavior Research Methods, Instruments, and Computers,
33(4), 517–523.
Ambulatory Assessment: Methods for Studying Everyday Life
13
Mehl, M. R., Vazire, S., Ramirez-Esparza, N., Slatcher, R. B., & Pennebaker, J. W.
(2007). Are women really more talkative than men? Science, 317(5834), 82.
doi:10.1126/science.1139940
Miller, G. (2012). The smartphone psychology manifesto. Perspectives on Psychological
Science, 7(3), 221–237. doi:10.1177/1745691612441215
Montoliu, R., Blom, J., & Gatica-Perez, D. (2013). Discovering places of interest
in everyday life from smartphone data. Multimedia Tools and Applications, 62(1),
179–207. doi:10.1007/s11042-011-0982-z
Rachuri, K. K., Musolesi, M., Mascolo, C., Rentfrow, P. J., Longworth, C., & Aucinas,
A. (2010). EmotionSense: A mobile phones based adaptive platform for experimental social
psychology research. Paper presented at the Proceedings of the 12th ACM international conference on Ubiquitous computing.
Ramírez-Esparza, N., Mehl, M., Álvarez-Bermúdez, J., & Pennebaker, J. (2009). Are
Mexicans more or less sociable than Americans? Insights from a naturalistic
observation study. Journal of Research in Personality, 43(1), 1–7. doi:10.1016/
j.jrp.2008.09.002
Robinson, M. D., & Clore, G. L. (2002). Belief and feeling: Evidence for an
accessibility model of emotional self-report. Psychological Bulletin, 128, 934–960.
doi:10.1037/0033-2909.128.6.934
Smyth, J., Ockenfels, M. C., Porter, L., Kirschbaum, C., Hellhammer, D. H., & Stone,
A. A. (1998). Stressors and mood measured on a momentary basis are associated with salivary cortisol secretion. Psychoneuroendocrinology, 23(4), 353–370.
doi:10.1016/s0306-4530(98)00008-0
Van de Water, A. T., Holmes, A., & Hurley, D. A. (2011). Objective measurements
of sleep for non-laboratory settings as alternatives to polysomnography—A
systematic review. Journal of Sleep Research, 20(1 Pt 2), 183–200. doi:10.1111/
j.1365-2869.2009.00814.x
Walls, T. A., & Schafer, J. L. (2006). Models for intensive longitudinal data. New York,
NY: Oxford University Press.
Whyte, W. H. (1980). The social life of small urban spaces. Washington, D.C.: The Conservation Foundation.
Wilhelm, F. H., & Grossman, P. (2010). Emotions beyond the laboratory: Theoretical fundaments, study design, and analytic strategies for advanced ambulatory
assessment. Biological Psychology. doi:10.1016/j.biopsycho.2010.01.017
Wolf, P. S., & Jacobs, W. J. (2010). GPS technology and human psychological research:
A methodological proposal. Journal of Methods and Measurement in the Social Sciences, 1(1), 1–7.
Yarkoni, T. (2012). Psychoinformatics: New horizons at the interface of the psychological and computing sciences. Current Directions in Psychological Science, 21(6),
391–397. doi:10.1177/0963721412457362
FURTHER READING
Bolger, N., & Laurenceau, J.-P. (2013). Intensive longitudinal methods: An introduction
to diary and experience sampling research. New York, NY: Guilford Press.
14
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
Hektner, J. M., Schmidt, J. A., & Csikszentmihalyi, M. (2007). Experience sampling
method: Measuring the quality of everyday life. Thousand Oaks, CA: SAGE Publications.
Mehl, M. R., & Conner, T. S. (2012). Handbook of research methods for studying daily life.
New York, NY: Guilford Press.
Miller, G. (2012). The smartphone psychology manifesto. Perspectives on Psychological
Science, 7(3), 221–237. doi:10.1177/1745691612441215
Stone, A., Shiffman, S., Atienza, A., & Nebeling, L. (Eds.) (2007). The science of real-time
data capture: Self-reports in health research. New York, NY: Oxford University Press.
Yarkoni, T. (2012). Psychoinformatics: New horizons at the interface of the psychological and computing sciences. Current Directions in Psychological Science, 21(6),
391–397. doi:10.1177/0963721412457362
TAMLIN S. CONNER SHORT BIOGRAPHY
Tamlin S. Conner, PhD is Senior Lecturer in Psychology at the University
of Otago in New Zealand. She received her doctorate in social psychology
from Boston College and completed postdoctoral training in health and
personality psychology at the University of Connecticut Health Center. She
has published numerous articles on the theory and practice of experience
sampling, is a leading expert on ambulatory self-report techniques, and
conducts research on well-being, emotions, and the science of self-report.
Recently, she coedited the Handbook of Research Methods for Studying Daily
Life (2012; Guilford Press) with Matthias Mehl. She is a founding member and current secretary of the Society for Ambulatory Assessment.
http://www.otago.ac.nz/psychology/staff/tamlinconner.html
MATTHIAS R. MEHL SHORT BIOGRAPHY
Matthias R. Mehl, PhD, is Associate Professor of Psychology at the
University of Arizona. He received his doctorate in social and personality
psychology from the University of Texas at Austin. Over the past decade, he
developed the electronically activated recorder (EAR) as a novel methodology for the unobtrusive naturalistic observation of daily life. He has given
workshops and published numerous articles on novel methods for studying
daily life and recently coedited the Handbook of Research Methods for Studying
Daily Life (2012; Guilford Press) with Tamlin Conner. He is a founding
member and the current Vice President of the Society for Ambulatory
Assessment. In 2011, the Association for Psychological Science identified
him as a “Rising Star.” http://dingo.sbs.arizona.edu/∼mehl/
Ambulatory Assessment: Methods for Studying Everyday Life
15
RELATED ESSAYS
Social Epigenetics: Incorporating Epigenetic Effects as Social Cause and
Consequence (Sociology), Douglas L. Anderton and Kathleen F. Arcaro
To Flop Is Human: Inventing Better Scientific Approaches to Anticipating
Failure (Methods), Robert Boruch and Alan Ruby
Repeated Cross-Sections in Survey Data (Methods), Henry E. Brady and
Richard Johnston
The Evidence-Based Practice Movement (Sociology), Edward W. Gondolf
Meta-Analysis (Methods), Larry V. Hedges and Martyna Citkowicz
The Use of Geophysical Survey in Archaeology (Methods), Timothy J.
Horsley
Network Research Experiments (Methods), Allen L. Linton and Betsy Sinclair
Longitudinal Data Analysis (Methods), Todd D. Little et al.
Data Mining (Methods), Gregg R. Murray and Anthony Scime
Remote Sensing with Satellite Technology (Archaeology), Sarah Parcak
Quasi-Experiments (Methods), Charles S. Reichard
Digital Methods for Web Research (Methods), Richard Rogers
Virtual Worlds as Laboratories (Methods), Travis L. Ross et al.
Modeling Life Course Structure: The Triple Helix (Sociology), Tom Schuller
Content Analysis (Methods), Steven E. Stemler
Person-Centered Analysis (Methods), Alexander von Eye and Wolfgang
Wiedermann
Translational Sociology (Sociology), Elaine Wethington
Ambulatory Assessment: Methods
for Studying Everyday Life
TAMLIN S. CONNER and MATTHIAS R. MEHL
Abstract
Ambulatory assessment is a class of methods that use mobile technology to understand people’s biopsychosocial processes in natural settings, in real time, and on
repeated occasions. In this essay, we discuss the rationale for ambulatory assessment
including the benefits of measuring people in the real world (greater ecological
validity, better understanding of people in contexts), in real time (avoidance of
memory bias, greater sensitivity for capturing change), and over time (capturing
within-person patterns and temporal trends). Then, we review the latest ambulatory
assessment techniques for measuring experiences, behaviors, and physiology in
daily life. Experiences such as emotions, physical pain, and daily stressors can be
tracked using daily diaries and smartphone-based experience sampling. Behaviors
such as activity, movement, location, and natural language use can be tracked using
accelerometers, portable actigraphs, global positioning system (GPS) coordinates,
and the electronically activated recorder (EAR). Physiological processes such as
heart rate, blood pressure, and electrodermal activity can be measured using an
array of ambulatory biosensors. Ambulatory assessment will continue to be revolutionized by smartphones, which are becoming integrated seamlessly into people’s
lives. Emerging trends include social sensing applications that make inferences
about users’ psychological processes based on multi-channel information collected
from smartphones, emergence of “big data collection” whereby ambulatory
assessment data is gathered en masse from large populations, and the growing field
of mobile health. These trends raise questions around the protection of participants’
privacy and the synthesis of immense amounts of digital data. Ultimately, these
developments will narrow the separation between science and everyday life as
ambulatory assessment becomes an integrated part of people’s mobile lives.
INTRODUCTION: DEFINITION OF AND RATIONALE FOR
AMBULATORY ASSESSMENT
Laboratory-based methods have historically been the strength and pride
of the social and behavioral sciences. Most research method textbooks
emphasize how laboratory research helps control confounding variables
and thereby allows for the isolation of a causal factor of interest. Laboratory
Emerging Trends in the Social and Behavioral Sciences. Edited by Robert Scott and Stephen Kosslyn.
© 2015 John Wiley & Sons, Inc. ISBN 978-1-118-90077-2.
1
2
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
research has contributed enormously to our understanding of human
social behavior. Ultimately, though, laboratory research can exclusively
accumulate knowledge on what can happen—under isolated and controlled
circumstances—and it cannot speak to what does happen under the circumstances that people normally encounter in their everyday lives. Therefore, a
comprehensive science of human behavior also requires “thinking outside
the experimental box” and necessitates the study of humans in their natural
habitat—that is, the collection of data from individuals as they live their
lives in their daily environments.
This essay reviews existing methodologies for studying experiences,
behavior, and physiology in daily life (for a comprehensive review, see
Mehl & Conner, 2012). We here refer to these methodologies collectively as
ambulatory assessment methods. According to the Society for Ambulatory
Assessment, Ambulatory Assessment “comprises the use of field methods
to assess the ongoing behavior, physiology, experience and environmental
aspects of humans or nonhuman primates in naturalistic or unconstrained
settings. [It] designates an ecologically relevant assessment perspective that
aims at understanding biopsychosocial processes as they naturally unfold
in time and in context” (www.ambulatory-assessment.org). Other common
names to denote these methodologies include experience sampling methods,
diary methods, and ecological momentary assessment. At their core, these
methods allow researchers to study individuals (i) in their natural settings, (ii)
in real time (or close to real time), and (iii) on repeated occasions. Ambulatory
assessment derives its scientific rationale directly from these three measurement characteristics (i.e., “real-world,” “real-time,” “within-person”).
The “real-world” quality allows unprecedented access to what actually
happens in a person’s everyday life and the contexts that surround such
events (Mehl & Conner, 2012, Ch. 1). For example, ambulatory assessment
is a great way to track how much people are exercising every day and the
conditions that make exercise more or less likely. This real-world quality
also enables researchers to capitalize on the power of evocative daily events
that are difficult, if not unethical, to mimic in controlled laboratory settings
such as positive or negative daily interactions, disagreements with a friend,
issues with children, commuting, excessive smoking, or drinking. Moreover,
how people respond to a standardized laboratory stressor such as giving a
speech may differ from how they respond to stressful experiences in their
lives. Wilhelm and Grossman (2010) describe a participant who showed
rather minimal heart rate increases in response to a laboratory stress protocol
but quite dramatic heart rate increases later in the afternoon while watching
a soccer game at home. In addition, in a reverse pattern, patients often have
high blood pressure readings in the physician’s office but not in their home
environment—a phenomenon called the white coat hypertension. By testing
Ambulatory Assessment: Methods for Studying Everyday Life
3
behavior in real-world settings, researchers have greater confidence about
the ecological validity of findings—that is, whether an effect is representative
of how it operates in reality.
The “real-time” quality of ambulatory assessment enables researchers to
understand experiences as they occur, not how people represent them later
from memory. This quality is especially important for experiences that are
fleeting and often misremembered (e.g., emotions, and pain). A person’s
memory for how they “felt over the past week” or “the past month” is
influenced by a host of factors including salient experiences, mood at the
time of recall, personality traits, beliefs about the acceptability of certain
emotions, gender stereotypes, and cultural norms (Mehl & Conner, 2012,
Ch. 2). In fact, there is strong evidence that as the time delay between experience and reporting lengthens, self-reports become more stable, reflective,
belief-driven, and culturally homogenized rather than malleable, reflexive,
experience-driven, and individualized (Robinson & Clore, 2002). This distinction between real-time and recalled reporting can make a big difference
in understanding psychological processes. For one, ambulatory approaches
may be more sensitive for capturing change in emotional states in response
to interventions. In a trial of antidepressant medication treatment, changes in
depressive symptomology were detected earlier among patients randomly
assigned to track their symptoms each day for 30 days compared to patients
who reported their symptoms using standard one-week recall measures
(Lenderking et al., 2008). Other studies find stronger links between real-time
reports of emotion and atherosclerosis risk, immune system function,
and genetic vulnerability [reviewed by Conner and Barrett (2012)]. These
examples illustrate how-real time measurement may be more sensitive than
traditional memory-based measures under certain circumstances.
The “within-person” element of ambulatory assessment refers to the capacity
to identify patterns of behavior within the person across time (i.e., idiographic assessment) (Mehl & Conner, 2012, Ch. 3). This approach differs from
traditional cross-sectional research, that aims to identify patterns of behavior
between people (i.e., nomothetic assessment). Unlike cross sectional research
which observes people at a single time point, ambulatory assessment tracks
individuals’ experiences, behaviors, or physiology intensively over time,
yielding “intensive longitudinal data” (Walls & Schafer, 2006). That data can
be analyzed to uncover the temporal patterns and trends in each person’s
data, as well as commonalities and differences in these trends across people.
For example, in a classic study, Bolger and Schilling (1991) examined the
within-person relationships between daily stressors and anxiety for 339
people and found stronger within-person relationships for people higher
in Neuroticism. Ambulatory assessment has been used to uncover a wide
variety of within-person relationships, including the antecedents of heavy
4
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
drinking, triggers of smoking cravings, relationships between stress and
coping, and links between food consumption and mood.
FOUNDATIONAL RESEARCH: AMBULATORY ASSESSMENT
METHODS
AMBULATORY ASSESSMENT OF EXPERIENCE
One of the most common uses of ambulatory assessment is for studying
self-reported experiences in naturalistic settings. Experiences include current
mood states, levels of perceived stress, feelings of bodily pain or discomfort, and other aspects of daily life that can be self-reported. Thus, ambulatory assessment of experience inherently involves asking people to respond
to questions about their experiences as they go about their daily lives. The
Internet and, especially smartphones have revolutionized this methodology
(Miller, 2012) and now represent the main interface for ambulatory self-report
techniques. In general, there are two main ways of measuring experience
through ambulatory assessment—through Internet daily diaries and mobile
phone-based experience sampling.
Daily diary methods are probably the most common form of ambulatory
assessment. Daily diary methods are now done mainly through Internet surveys in which participants access the survey each day (or night) for a week
to three weeks to answer questions about their experiences that day. Diaries
are often used to assess experiences and events that occur on a daily basis
and can be easily recalled at the end of each day such as daily hassles, social
activities, emotions felt that day, and health behaviors such as alcohol use,
smoking, and food consumption. Although this approach is neither “real
time” nor necessarily ambulatory, it is considered “near to real time” and an
improvement upon asking people to recall their experiences across the entire
week or longer. Although daily diary studies are widely applied, they have
probably had the most impact on the psychology of romantic relationships,
the role daily stressors in mental and physical health, and the correlates and
consequences of daily health behaviors such as alcohol use (Mehl & Conner,
2012, Ch. 8). In addition, daily diary techniques are fairly easy to implement.
Researchers can design a simple Internet survey through any number of companies such as SurveyMonkey, Qualtrics, or SurveyGizmo. Participants then
access that survey during a specified time period, and the data can be downloaded at the end of the study.
A second approach to measuring experience is through mobile-phone-based
experience sampling. Experience sampling is the process of randomly signaling a person (typically 4–8 times per day) to report on their experiences at
Ambulatory Assessment: Methods for Studying Everyday Life
5
that moment [see suggested reading by Hektner, Schmidt, and Csikszentmihalyi (2007)]. This approach is used to study fleeting and ongoing subjective
experiences such as mood, pain, and stress because these experiences are
quick to decay in memory and are best assessed in true real time. Experience
sampling has probably had the most impact on the psychology of emotion.
Emotions are highly variable—they ebb and flow throughout the day in
response to changing internal and external events (Mehl & Conner, 2012,
Ch. 27). Experience sampling is ideally suited to capturing this changing
profile—an emotional “signature” that can reveal dynamic aspects of
functioning obscured by standard one-time surveys. Experience sampling
has revealed diurnal and weekly patterns in emotion, individual differences
in affective instability as a marker of psychopathology, differences in
the structure of emotional experience, divergences between experienced
versus remembered emotions, and covariation between emotions and
health-related factors. Experience sampling has also been used to identify
the emotional correlates of psychopathology including heightened affective
instability, anhedonia, and the emotional precursors to self-harm (Mehl &
Conner, 2012, Ch. 23).
Currently, there are three main approaches to experience sampling with
mobile phones. In each of these approaches, the trend is towards using people’s own phones to allow for seamless participation without the need for
an extra specialized device. One very simple approach is to send questions
via SMS text messaging. Texts can be scheduled and sent automatically
through most commercial SMS companies (e.g., www.message-media.com).
Participants reply to the questions contained in the text using numbers
on their keypad and the data can be downloaded from the SMS company
server at the end of the study. Although this approach is simple and does
not require a smartphone (mobile phone with Internet capability), it is also
the least flexible because of limited timing controls, lack of branching, and
restricted space for questions. A second approach is to send a hyperlink
to an online survey via SMS text messaging to participants with Internet
enabled smartphones. Here, participants receive a text message with a
hyperlink that directs them to a mobile-ready Internet survey. The survey
can be developed through any number of companies such as SurveyMonkey,
and the hyperlink to the online survey can be sent through a commercial
SMS company or a specialized service designed for experience sampling
(e.g., www.surveysignal.com). A third approach is to use application-based
smartphone tools. Here, participants download a smartphone app that
delivers a specialized survey to their smartphones. Although these apps and
surveys can be designed from the ground up, there are a growing number of
companies that provide assistance with survey development at a reasonable
charge (mEMA, iHabit, iForm, ISurvey, MovisensXS, and Qualtrics). There
6
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
are also a growing number of open-source development tools (Paco and
Funf). One issue with app-based experience sampling is that apps and
surveys are often designed for only one operating system (e.g., Apple’s iOS
or Google’s Android) that places limits on recruitment and participation.
However, increasingly, tools are being developed for both platforms.
AMBULATORY ASSESSMENT OF BEHAVIOR
Although self-reports are important to social science, oftentimes people
cannot or sometimes might not want to accurately report what they do. In
these circumstances, the direct—and ideally nonreactive—assessment of
real-world behavior is of high importance. For example, Mehl and colleagues
have developed the electronically activated recorder or EAR methodology
that allows for the relatively unobtrusive naturalistic observation of participants’ acoustic behavior in daily life (Mehl, Pennebaker, Crow, Dabbs,
& Price, 2001). The current EAR system, the “iEAR,” consists of a free iOS
app that runs on iPod touch and iPhone devices. Participants carry an iEAR
device on them as they go about their normal lives. The app periodically
records snippets of ambient sounds (e.g., 30 s every 12 min) thereby creating
a series of sound bites that, together, amount to acoustic logs of participants’
days as they naturally unfold. The ambient sound recordings are later
securely downloaded, reviewed by participants, and then coded for aspects
of participants’ momentary locations (e.g., in a public or private place),
activities (e.g., watching TV, and eating), interactions (e.g., along, in a group,
and on the phone), and emotional expressions (e.g., laughing and sighing).
Initial EAR research focused on the psychometric properties of naturalistically observed daily social behavior. This research showed (i) that a
broad spectrum of behaviors can be assessed reliably and with low levels
of reactivity from the sampled ambient sounds, (ii) that these behaviors
show large between-person variability and good temporal stability, and (iii)
that they have good convergent validity with theoretically related measures
(e.g., Big Five personality dimensions) (Mehl & Conner, 2012, Ch. 10). The
second generation studies, then, focused on the EAR’s potential to address
questions that are difficult to answer with other methods. For example, in
a cross-cultural study, Ramirez-Esparza and her colleagues used the EAR
method to study self-reported sociability in relation to observed sociability
in the United States and Mexico. They found that although American participants rated themselves significantly higher than Mexicans on the question “I
see myself as a person who is talkative,” they actually spent almost 10% less
time talking (Ramírez-Esparza, Mehl, Álvarez-Bermúdez, & Pennebaker,
2009). In a similar way, Mehl and his colleagues used the EAR method to
debunk the long-standing myth that women are by a factor more talkative
Ambulatory Assessment: Methods for Studying Everyday Life
7
than men (Mehl, Vazire, Ramirez-Esparza, Slatcher, & Pennebaker, 2007).
Using data from six studies, they showed that both sexes use on average
about 16,000 words per day. Together, these studies showed how the EAR
method can be used to study objective aspects of daily behavior and how it
can yield results that diverge from findings obtained with other methods.
A series of other creative ways for assessing behavior directly and unobtrusively in the real world have been developed. For example, time-lapse
photography has been used to study the flow of people and the use of
space in urban public places (Whyte, 1980). In modern studies, participants’
movement and location are tracked via actigraphy and GPS information. To
determine sleep patterns and circadian rhythms, studies have participants
wear small, rugged wrist watches that log body movements along with
day-night (i.e., light) patterns (Van de Water, Holmes, & Hurley, 2011). Multichannel activity monitoring devices provide more detailed information on
posture and motion through the placement of small accelerometer sensors on different body locations (e.g., arm, leg, or waist). Classification
algorithms then convert the raw sensor input into discrete posture (e.g.,
lying, and sitting, and standing) and motion (e.g., walking, cycling, and
driving) patterns. Importantly, validation studies have consistently found
critical discrepancies between self-reported and objective activity records
(Mehl & Conner, 2012, Ch. 13). Finally, location-tracking via either dedicated GPS devices or smartphones with GPS and Wi-Fi sensors are on the
way of becoming mainstream in the social sciences (Montoliu, Blom, &
Gatica-Perez, 2013; Wolf & Jacobs, 2010). Although these tools currently exist
as stand-alone assessment devices, in the future, they will be integrated into
mobile devices that people naturally carry with them which will allow more
seamless integrated assessment (Miller, 2012).
AMBULATORY ASSESSMENT OF PHYSIOLOGY
Finally, ambulatory assessment methods also exist for the sampling of
physiological activity in everyday life. An array of biosignals can now be
measured reliably via portable signal recording devices (e.g., electrocardiogram, blood pressure, electrodermal activity, and body temperature)
(Wilhelm & Grossman, 2010). Recently researchers have added ambulatory
assessment of hormones and other biomarkers to the list (Mehl & Conner,
2012, Ch. 11). As an example of research that implemented traditional ambulatory physiological monitoring, Lane, Zareba, Reis, Peterson, and Moss
(2011b) used experience sampling combined with ambulatory electrocardiography (a so-called Holter monitor) to show that daily emotions—even at
low intensities—triggered abnormal cardiac activity among patients with a
congenital heart abnormality. In a classic study on hormonal responses in
8
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
daily life, Smyth et al. (1998) combined experience sampling with momentary
assessment of cortisol. They found that momentary reports of current or
anticipated stress predicted increased cortisol secretion 20 min later.
Taken together, these two examples illustrate how ambulatory physiological monitoring has been used to link mundane and seemingly inconsequential experiences in our daily lives to objective physiological responses. The
development of novel ways to track what goes on underneath our skins as
we go about our lives is a rapidly advancing field and important advances
can be expected in the future (Kim et al., 2011).
CUTTING-EDGE RESEARCH AND FUTURE DIRECTIONS
As the mobile device revolution is unfolding around us, it is clear that ambulatory assessment will, over time, be revolutionized by it. Smartphones will
not just be devices for everyday communication but will also become devices
for large-scale scientific data collection and intervention (Kaplan & Stone,
2013; Yarkoni, 2012). They automatically store vast amounts of real-world
user interaction data and are equipped with an array of high quality sensors
to track the physical (e.g., location and position) and social (e.g., blue tooth
connections) context of these interactions. Finally, with add-on sensors, they
will be able monitor physiological parameters. In a visionary article, Miller
(2012) states, “the question is not whether smartphones will revolutionize
psychology but how, when, and where the revolution will happen” (p. 234).
SMARTPHONE SENSING
One flourishing research area at the intersection of the social and computer
sciences is the development of “smartphone sensing” applications. The idea
behind these applications is to make inferences about users’ emotions, behavior, environments, and life patterns through computational integration of the
data produced by (i) interactions with the user interface (e.g., timing and
duration of phone calls number text messages) and (ii) the multiple sensors embedded in smartphones (e.g., Bluetooth, GPS, and accelerometer).
For example, de Montjoye, Quoidbach, Robic, and Pentland (2013) recently
showed that the personality of smartphone users (e.g., extraversion and neuroticism) can be predicted with high levels of accuracy from information that
is routinely part of the data logs of mobile phone carriers (e.g., number of
interactions, number and diversity of contacts, response latency to events,
and distance traveled).
Lu et al. (2012) have applied this idea to automatic voice-based stress
detection via smartphones. Drawing on prior stress research, the so-called
StressSense app monitors ambient sounds for voices, performs speaker
Ambulatory Assessment: Methods for Studying Everyday Life
9
separation, and extracts stress-relevant voice parameters (e.g., speech rate,
pitch variability, and jitter). These parameters are then integrated into
stress-level estimates using machine learning algorithms that are trained
with the user’s galvanic skin response as the “ground truth” of how stressed
the user really is. The authors report high classification accuracy for both
outdoor and indoor environments. In a similar way, Rachuri et al. (2010) have
been developing a mobile phone application for the automatic recognition
of discrete emotions. Their “EmotionSense” app operates by extracting
voice parameters and comparing them against an internal “emotion prosody
library” that is derived from voice feature analysis of enacted target emotions
(happy sad, fearful, angry, neutral).
Finally, in an intriguing study, Lane et al. (2011a) report the development of
“BeWell” as a smartphone application to promote healthy lifestyles. The app
continuously monitors users’ physical activity (via the embedded accelerometer), sleep activity (via the accelerometer and recharging information), and
social activity (via ambient sounds containing voice). In a second step, it compares the estimated levels against established health recommendations (e.g.,
ideal value of 7 h of sleep). In a third step, the app feeds the results back
to the user intuitively on the display where it visualizes a person’s wellness
through an aquarium with swimming fish—the vitality of which reflects the
state of wellness. Because all computations are run directly on the phone, the
app is self-sufficient but currently absorbs a high amount of processing time
and battery life. As a proof-of-concept study, though, it shows a powerful
application of mobile-phone based social or “life-style” sensing.
“BIG DATA” COLLECTION
Future progress in this area is also tied into a rapidly decreasing per-person
cost thereby allowing data collection at large-scale levels. Already we
are beginning to see studies with “big data” from thousands of people.
For example, one group of researchers analyzed Geographical Positional
System signals from 100,000 mobile phone users over a 6-month period to
show reproducible regularities in their within-person movement patterns
(Gonzalez, Hidalgo, & Barabasi, 2008). Other research uses experience sampling tools made available to a wide audience in exchange for scientific use
of their anonymized data (e.g., Mappiness, Trackyourhappiness, EmotionSense, and Happathon). For example, data from 2250 users of TrackYourHappiness was used to show the conditions and contexts in which people report
being happier, such as when they are social and not “mind-wandering”
(Killingsworth & Gilbert, 2010). Likewise, Mappiness data from nearly 22,000
users in the United Kingdom found that people reported greater happiness
when they were located near natural environments as determined by GPS
10
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
location (MacKerron & Mourato, 2013). Other big data projects include the
Gallup-Healthways Well-Being Index, an American-based population-based
phone survey of over 1,000,000 people that includes a daily mood measure,
the World Well-Being Project, which analyzes language in Facebook and
Twitters posts to index differences in psychological states, and large-scale
mining of Twitter data (e.g., Golder & Macy, 2011)
MOBILE HEALTH: LARGE SCALE SAMPLING OF ELECTRONIC HEALTH INFORMATION
Large scale ambulatory assessment will also transform health research. The
E-Heart Study at the University of California San Francisco is a creative
new project that aims to collect ambulatory heart health data from one
million people. Tools in the study include mobile phone surveys, mobile
apps, and special sensors integrate with participants’ smartphones to provide real-time health recordings (heart rate, blood pressure, activity, sleep
quality, etc.). Their goal is to capitalize on big data to “develop strategies to
prevent and treat all aspects of heart disease” (https://www.health-eheartstudy.org/study). Other large mobile health projects include the National
Experience Sampling Project, which aims to collect ambulatory health data
at the population level. Future progress in this area will also benefit from
disposable wireless biometric patches that can be worn continuously.
KEY ISSUES GOING FORWARD IN AMBULATORY ASSESSMENT
Two of the most pressing issues going forward concern (i) the protection of
participants’ privacy and (ii) the synthesis of the immense amount of digital
data. There is little doubt that the Internet and, most importantly, online
social networking has already dramatically changed notions of privacy in
people’s personal lives. About a decade ago, it was ethically questionable to
“Google” someone before a date. Now, Facebook users readily post private
pictures of and intimate comments about their lives to hundreds of online
friends. For maximizing the capabilities of their mobile phones, people also
accept the complete, centralized tracking of their locations, browsing and
search history, and entertainment choices (e.g., iTunes, Netflix, YouTube, and
Kindle). These changes in how private information is shared are bound to
affect perceptions of what data is acceptable to collect for scientific purposes.
At the same time, though, these changes have profound implications for
the confidentiality of scientific data and the protection of participants’
privacy. King (2011) pointed out that it is de facto impossible to guarantee
anonymity by combining the three demographic variables date of birth,
gender, and zip code. In a similar way—and directly in the context of mobile
data collection—Kosinski, Stillwell, and Graepel (2013) recently showed
Ambulatory Assessment: Methods for Studying Everyday Life
11
that highly private and often stigmatizing characteristics such as sexual
orientation, ethnicity, and religious and political affiliation can be readily
predicted from only one type of digital data, “Likes” in users’ Facebook
profiles. The same was true for important health behaviors such as smoking,
drinking, and drug use. Combined with the scientifically desirable trend
towards data sharing and making (taxpayer-funded) data bases publically
available and advances in large-scale “big data” mining, it is clear that the
ambulatory assessment researchers, and the scientific community more
generally, have to develop new guidelines and methods of protecting the
privacy of human subjects.
Researchers will also need to develop better strategies for handling large
amounts of data. Ambulatory assessment data is already quite large and
requires specialized tools for treating the nested data structure such as multilevel modeling. However, big data will increase the size and complexity of
these data structures exponentially. Such data will require different analytic
approaches that likely draw on techniques from bioinformatics and computer science. Yarkoni (2012) calls this new approach “psychoinformatics,”
which can include tools such as network analysis, large-scale exploratory
data analysis, and a greater reliance on more flexible open-source statistical
software such as R. This requirement for greater statistical sophistication will
require new forms of training and greater collaboration among statisticians,
computer scientists, and psychological scientists.
CONCLUSION
The separation between science and everyday life will become narrower with
each decade as ambulatory assessment becomes integrated seamlessly into
people’s lives. Although ambulatory assessment will continue to complement rather than replace controlled laboratory science, it will begin to play
a larger role in science than it has in the past especially as findings from
population-based data-sets begin to bear fruit. Issues of privacy and data
management notwithstanding, the future of ambulatory assessment future
will be a dynamic, collective, and collaborative process.
REFERENCES
Bolger, N., & Schilling, E. A. (1991). Personality and the problems of everyday life:
The role of neuroticism in exposure and reactivity to daily stressors. Journal of Personality, 59, 355–386.
Conner, T. S., & Barrett, L. F. (2012). Trends in ambulatory self-reports: The role
of momentary experience in psychosomatic medicine. Psychosomatic Medicine, 74,
327–337. doi:10.1097/PSY.0b013e3182546f18
12
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
de Montjoye, Y.-A., Quoidbach, J., Robic, F., & Pentland, A. S. (2013). Predicting personality using novel mobile phone-based metrics. Social Computing,
Behavioral-Cultural Modeling and Prediction, 7812, 48–55.
Golder, S. A., & Macy, M. W. (2011). Diurnal and seasonal mood vary with work,
sleep, and daylength across diverse cultures. Science, 333, 1878–1881. doi:10.1126/
science.1202775
Gonzalez, M. C., Hidalgo, C. A., & Barabasi, A. L. (2008). Understanding individual
human mobility patterns. Nature, 453, 779–782. doi:10.1038/nature06958
Kaplan, R. M., & Stone, A. A. (2013). Bringing the laboratory and clinic to the community: Mobile technologies for health promotion and disease prevention. Annual
Review of Psychology, 64, 471–498. doi:10.1146/annurev-psych-113011-143736
Killingsworth, M. A., & Gilbert, D. T. (2010). A wandering mind is an unhappy mind.
Science, 330, 932. doi:10.1126/science.1192439
Kim, D.-H., Lu, N., Ma, R., Kim, Y.-S., Kim, R.-H., Wang, S., … Islam, A. (2011).
Epidermal electronics. Science, 333(6044), 838–843. doi:10.1016/j.wneu.2011.
10.001
King, G. (2011). Ensuring the data-rich future of the social sciences. Science, 331(6018),
719–721. doi:10.1126/science.1197872
Kosinski, M., Stillwell, D., & Graepel, T. (2013). Private traits and attributes are
predictable from digital records of human behavior. Proceedings of the National
Academy of Sciences of the United States of America. doi:10.1073/pnas.1218772110
Lane, N. D., Mohammod, M., Lin, M., Yang, X., Lu, H., Ali, S., … Campbell, A.
(2011a). Bewell: A smartphone application to monitor, model and promote wellbeing.
Paper presented at the 5th International Conference on Pervasive Computing
Technologies for Healthcare (PervasiveHealth2011).
Lane, R. D., Zareba, W., Reis, H. T., Peterson, D. R., & Moss, A. J. (2011b). Changes
in ventricular repolarization duration during typical daily emotion in patients
with long QT syndrome. Psychosomatic Medicine, 73(1), 98–105. doi:10.1097/
PSY.0b013e318203310a
Lenderking, W. R., Hu, M., Tennen, H., Cappelleri, J. C., Petrie, C. D., & Rush, A. J.
(2008). Daily process methodology for measuring earlier antidepressant response.
Contemporary Clinical Trials, 29, 867–877. doi:10.1016/j.cct.2008.05.012
Lu, H., Frauendorfer, D., Rabbi, M., Mast, M. S., Chittaranjan, G. T., Campbell, A.
T., … Choudhury, T. (2012). StressSense: detecting stress in unconstrained acoustic
environments using smartphones. Paper presented at the Proceedings of the 2012
ACM Conference on Ubiquitous Computing.
MacKerron, G., & Mourato, S. (2013). Happiness is greater in natural environments.
Global Environmental Change, 23(5), 992–1000.
Mehl, M. R., & Conner, T. S. (2012). Handbook of research methods for studying daily life.
New York, NY: Guilford Press.
Mehl, M. R., Pennebaker, J. W., Crow, D. M., Dabbs, J., & Price, J. H. (2001). The
electronically activated recorder (EAR): A device for sampling naturalistic daily
activities and conversations. Behavior Research Methods, Instruments, and Computers,
33(4), 517–523.
Ambulatory Assessment: Methods for Studying Everyday Life
13
Mehl, M. R., Vazire, S., Ramirez-Esparza, N., Slatcher, R. B., & Pennebaker, J. W.
(2007). Are women really more talkative than men? Science, 317(5834), 82.
doi:10.1126/science.1139940
Miller, G. (2012). The smartphone psychology manifesto. Perspectives on Psychological
Science, 7(3), 221–237. doi:10.1177/1745691612441215
Montoliu, R., Blom, J., & Gatica-Perez, D. (2013). Discovering places of interest
in everyday life from smartphone data. Multimedia Tools and Applications, 62(1),
179–207. doi:10.1007/s11042-011-0982-z
Rachuri, K. K., Musolesi, M., Mascolo, C., Rentfrow, P. J., Longworth, C., & Aucinas,
A. (2010). EmotionSense: A mobile phones based adaptive platform for experimental social
psychology research. Paper presented at the Proceedings of the 12th ACM international conference on Ubiquitous computing.
Ramírez-Esparza, N., Mehl, M., Álvarez-Bermúdez, J., & Pennebaker, J. (2009). Are
Mexicans more or less sociable than Americans? Insights from a naturalistic
observation study. Journal of Research in Personality, 43(1), 1–7. doi:10.1016/
j.jrp.2008.09.002
Robinson, M. D., & Clore, G. L. (2002). Belief and feeling: Evidence for an
accessibility model of emotional self-report. Psychological Bulletin, 128, 934–960.
doi:10.1037/0033-2909.128.6.934
Smyth, J., Ockenfels, M. C., Porter, L., Kirschbaum, C., Hellhammer, D. H., & Stone,
A. A. (1998). Stressors and mood measured on a momentary basis are associated with salivary cortisol secretion. Psychoneuroendocrinology, 23(4), 353–370.
doi:10.1016/s0306-4530(98)00008-0
Van de Water, A. T., Holmes, A., & Hurley, D. A. (2011). Objective measurements
of sleep for non-laboratory settings as alternatives to polysomnography—A
systematic review. Journal of Sleep Research, 20(1 Pt 2), 183–200. doi:10.1111/
j.1365-2869.2009.00814.x
Walls, T. A., & Schafer, J. L. (2006). Models for intensive longitudinal data. New York,
NY: Oxford University Press.
Whyte, W. H. (1980). The social life of small urban spaces. Washington, D.C.: The Conservation Foundation.
Wilhelm, F. H., & Grossman, P. (2010). Emotions beyond the laboratory: Theoretical fundaments, study design, and analytic strategies for advanced ambulatory
assessment. Biological Psychology. doi:10.1016/j.biopsycho.2010.01.017
Wolf, P. S., & Jacobs, W. J. (2010). GPS technology and human psychological research:
A methodological proposal. Journal of Methods and Measurement in the Social Sciences, 1(1), 1–7.
Yarkoni, T. (2012). Psychoinformatics: New horizons at the interface of the psychological and computing sciences. Current Directions in Psychological Science, 21(6),
391–397. doi:10.1177/0963721412457362
FURTHER READING
Bolger, N., & Laurenceau, J.-P. (2013). Intensive longitudinal methods: An introduction
to diary and experience sampling research. New York, NY: Guilford Press.
14
EMERGING TRENDS IN THE SOCIAL AND BEHAVIORAL SCIENCES
Hektner, J. M., Schmidt, J. A., & Csikszentmihalyi, M. (2007). Experience sampling
method: Measuring the quality of everyday life. Thousand Oaks, CA: SAGE Publications.
Mehl, M. R., & Conner, T. S. (2012). Handbook of research methods for studying daily life.
New York, NY: Guilford Press.
Miller, G. (2012). The smartphone psychology manifesto. Perspectives on Psychological
Science, 7(3), 221–237. doi:10.1177/1745691612441215
Stone, A., Shiffman, S., Atienza, A., & Nebeling, L. (Eds.) (2007). The science of real-time
data capture: Self-reports in health research. New York, NY: Oxford University Press.
Yarkoni, T. (2012). Psychoinformatics: New horizons at the interface of the psychological and computing sciences. Current Directions in Psychological Science, 21(6),
391–397. doi:10.1177/0963721412457362
TAMLIN S. CONNER SHORT BIOGRAPHY
Tamlin S. Conner, PhD is Senior Lecturer in Psychology at the University
of Otago in New Zealand. She received her doctorate in social psychology
from Boston College and completed postdoctoral training in health and
personality psychology at the University of Connecticut Health Center. She
has published numerous articles on the theory and practice of experience
sampling, is a leading expert on ambulatory self-report techniques, and
conducts research on well-being, emotions, and the science of self-report.
Recently, she coedited the Handbook of Research Methods for Studying Daily
Life (2012; Guilford Press) with Matthias Mehl. She is a founding member and current secretary of the Society for Ambulatory Assessment.
http://www.otago.ac.nz/psychology/staff/tamlinconner.html
MATTHIAS R. MEHL SHORT BIOGRAPHY
Matthias R. Mehl, PhD, is Associate Professor of Psychology at the
University of Arizona. He received his doctorate in social and personality
psychology from the University of Texas at Austin. Over the past decade, he
developed the electronically activated recorder (EAR) as a novel methodology for the unobtrusive naturalistic observation of daily life. He has given
workshops and published numerous articles on novel methods for studying
daily life and recently coedited the Handbook of Research Methods for Studying
Daily Life (2012; Guilford Press) with Tamlin Conner. He is a founding
member and the current Vice President of the Society for Ambulatory
Assessment. In 2011, the Association for Psychological Science identified
him as a “Rising Star.” http://dingo.sbs.arizona.edu/∼mehl/
Ambulatory Assessment: Methods for Studying Everyday Life
15
RELATED ESSAYS
Social Epigenetics: Incorporating Epigenetic Effects as Social Cause and
Consequence (Sociology), Douglas L. Anderton and Kathleen F. Arcaro
To Flop Is Human: Inventing Better Scientific Approaches to Anticipating
Failure (Methods), Robert Boruch and Alan Ruby
Repeated Cross-Sections in Survey Data (Methods), Henry E. Brady and
Richard Johnston
The Evidence-Based Practice Movement (Sociology), Edward W. Gondolf
Meta-Analysis (Methods), Larry V. Hedges and Martyna Citkowicz
The Use of Geophysical Survey in Archaeology (Methods), Timothy J.
Horsley
Network Research Experiments (Methods), Allen L. Linton and Betsy Sinclair
Longitudinal Data Analysis (Methods), Todd D. Little et al.
Data Mining (Methods), Gregg R. Murray and Anthony Scime
Remote Sensing with Satellite Technology (Archaeology), Sarah Parcak
Quasi-Experiments (Methods), Charles S. Reichard
Digital Methods for Web Research (Methods), Richard Rogers
Virtual Worlds as Laboratories (Methods), Travis L. Ross et al.
Modeling Life Course Structure: The Triple Helix (Sociology), Tom Schuller
Content Analysis (Methods), Steven E. Stemler
Person-Centered Analysis (Methods), Alexander von Eye and Wolfgang
Wiedermann
Translational Sociology (Sociology), Elaine Wethington
