03 · What You Need to Know
How Wearable Monitoring Can Influence the Behavior Being Observed
What Is Measurement Reactivity?
Measurement reactivity occurs when the process of observing or measuring a phenomenon influences that phenomenon.
In wearable research, participants may behave differently because they are wearing a device, know that their activity is recorded, or receive information about their performance.
For example, a participant who normally walks relatively little may become more active after receiving an activity monitor, even when the research team provides no exercise intervention.
Such changes may be intentional or arise through increased awareness of behavior.
However, a change observed after device distribution does not independently establish reactivity. Seasonal variation, study recruitment, other interventions, and ordinary behavioral fluctuations may also explain the pattern.
Distinguish Passive Measurement, Self-Monitoring, and Behavioral Intervention
Passive measurement
Recording information with limited active participant input, without necessarily providing behavioral feedback.
Self-monitoring
Participants observe or review information about their own behavior or physiological state.
Behavioral intervention
A planned procedure intended to influence behavior, potentially involving feedback, goals, reminders, or reinforcement.
These categories may overlap.
A wearable device can passively record movement while simultaneously displaying daily step counts. The recording may be passive from the researcher's perspective, but the feedback may encourage participants to monitor or change their behavior.
Researchers should describe the complete participant experience rather than classify a device as passive solely because data capture is automated.
What Is the Hawthorne Effect?
The term Hawthorne effect is commonly used to describe changes in behavior associated with awareness of being observed or studied.
However, the historical evidence and conceptual interpretation of the Hawthorne effect are contested. The term is sometimes applied broadly to several different mechanisms, including attention, expectations, and observation-related changes.
For wearable studies, measurement reactivity is often a more precise description because it identifies the methodological concern without assuming a particular psychological mechanism.
Researchers should avoid attributing every behavioral change during monitoring to a Hawthorne effect without suitable evidence.
How Can Wearable Devices Influence Participants?
| Potential mechanism |
Illustrative participant response |
Possible research consequence |
| Awareness of observation |
Participant becomes more conscious of activity |
Behavior differs from usual routine |
| Visible feedback |
Participant checks daily step counts |
Activity may increase or become goal-directed |
| Notifications and goals |
Device prompts movement or rest |
Monitoring becomes an intervention component |
| Perceived researcher expectations |
Participant attempts to appear more active |
Possible socially influenced behavior |
| Physical discomfort |
Participant avoids certain activities |
Behavior and wear adherence may change |
| Privacy concerns |
Participant changes routines or removes device |
Selective observation and missing data |
| Habituation |
Participant becomes accustomed to wearing device |
Initial behavioral effects may diminish |
These mechanisms are plausible pathways rather than inevitable consequences. Their relevance depends on the device, participant population, research context, and monitoring protocol.
Can Participants Change Behavior Without Receiving Feedback?
Yes. Awareness that behavior is being recorded may influence participants even when the device has no display or notifications.
For example, a person wearing a research accelerometer may become more attentive to daily movement because participation has made physical activity salient.
However, removing feedback does not establish that monitoring has no behavioral effect.
Researchers should distinguish the effect of receiving information from the effect of knowing that observation is occurring.
These influences may require different study designs to investigate.
How Does Feedback Affect Activity Monitoring?
Many consumer wearables display step counts, activity goals, exercise summaries, or achievement notifications.
Such features may encourage behavior change, particularly when participants use the information to set goals or evaluate progress.
Research on pedometer-based interventions has found that self-monitoring approaches can be associated with increased physical activity, although effects depend on intervention design and population.
These findings do not establish that every wearable produces the same effect or that passive recording without feedback necessarily changes activity.
Researchers should determine whether feedback is part of the intended intervention or a potentially unwanted influence on observation.
Can Monitoring Cause Participants to Become Less Active?
Potentially. Reactivity need not involve improvement in a socially desirable direction.
A device may cause discomfort, interfere with particular activities, or increase concern about physiological measurements.
Participants may also avoid activities because they believe the device could be damaged or because monitoring feels intrusive.
In other circumstances, monitoring may have little measurable effect.
Researchers should avoid assuming that reactivity always increases activity or improves outcomes.
Can Monitoring Affect Sleep?
Sleep monitoring may influence bedtime routines, attention to sleep quality, or concern about device-generated scores.
For example, participants may change their sleeping arrangements because of discomfort or become preoccupied with achieving a favorable sleep score.
The term orthosomnia has been used in the clinical literature to describe problematic preoccupation with obtaining ideal sleep-tracker measurements.
However, this does not mean that ordinary sleep tracking commonly causes such problems.
Researchers should consider whether the device changes sleep-related behavior or experience, particularly when studying subjective sleep quality.
Does Reactivity Necessarily Threaten Research Validity?
Not always. Its importance depends on the research question.
If the study aims to describe usual behavior without monitoring, observation-induced changes may threaten ecological validity.
However, if the research investigates the effect of providing activity trackers, behavior change caused by the device may be part of the intended intervention effect.
Similarly, when all participants are monitored under equivalent conditions, some monitoring effects may be shared across groups.
Even then, common monitoring does not guarantee that comparative estimates are unaffected. Participants may respond differently, and monitoring may interact with the intervention.
Researchers should identify the estimand before deciding whether reactivity is a bias, a mechanism of interest, or an acceptable feature of the study context.
How Can Reactivity Affect Baseline Measurement?
Baseline monitoring is often intended to establish participants' usual behavior before an intervention.
However, receiving a wearable may itself change behavior during the baseline period.
For example, participants may become unusually active during the first few days because the monitoring process draws attention to their activity.
If behavior subsequently returns toward its previous pattern, the baseline may not represent an ordinary unmonitored state.
Researchers should examine whether early monitoring differs from later observations and consider competing explanations for such changes.
Can a Familiarization Period Reduce Reactivity?
A familiarization or run-in period allows participants to become accustomed to the device before observations are used for the primary analysis.
This may reduce some effects associated with novelty or discomfort.
However, there is no universal duration that guarantees the disappearance of reactivity.
Some participants may adapt quickly, while others may continue responding to monitoring throughout the study.
Researchers should justify familiarization periods using relevant evidence and feasibility considerations rather than assume that excluding the first few days automatically eliminates observation effects.
Should Researchers Hide Activity Feedback?
Disabling feedback may be appropriate when the objective is to observe behavior without introducing activity goals or performance information.
For example, a research-grade accelerometer without a visible step counter may reduce opportunities for deliberate self-monitoring.
However, participants may still know they are being observed.
Researchers should also consider whether disabling feedback is technically possible, ethically appropriate, and consistent with participant information.
Concealing the existence or purpose of monitoring is a different ethical issue from disabling optional device feedback.
Researchers should not use undisclosed monitoring without appropriate ethical and legal authorization.
Can Researchers Measure Reactivity Directly?
Potentially, through a design that compares behavior under different monitoring conditions.
For example, researchers might randomly assign participants to feedback-enabled and feedback-disabled devices while using comparable measurement procedures.
Such a design could help estimate the effect of feedback, assuming appropriate randomization, implementation, and outcome measurement.
However, it would not necessarily isolate the effect of wearing a device because both groups are monitored.
Estimating the effect of awareness of monitoring may require an independent measurement method or another defensible design.
Researchers should distinguish which component of reactivity their comparison can actually identify.
Why Is Comparing the First and Last Monitoring Days Insufficient?
Behavioral differences across a monitoring period may suggest adaptation, but they do not independently demonstrate reactivity.
Activity may vary because of weekdays, weather, assessment schedules, illness, seasonal conditions, or other events.
For example, students monitored during an examination week may become less active for reasons unrelated to device awareness.
Researchers should consider temporal confounding and whether the study design supports attribution of observed changes to monitoring.
Descriptive comparisons can identify patterns worth investigating without necessarily establishing their cause.
How Can Reactivity Affect Group Comparisons?
Monitoring may influence groups differently.
For example, participants who already use fitness trackers may respond differently from those receiving a wearable for the first time.
Similarly, participants with high health motivation may respond more strongly to monitoring than those who find the device burdensome.
If these differences are associated with treatment assignment or other study characteristics, estimated group effects may be distorted.
Researchers should consider prior device experience, feedback exposure, adherence, and relevant participant characteristics when evaluating potential differential reactivity.
How Does Reactivity Interact With Missing Data?
Participants may remove devices during activities they prefer not to monitor or when wearing them becomes inconvenient.
This can create two related problems: the monitoring may influence behavior, and the resulting observations may omit particular activities.
For example, a participant might remove a wearable during vigorous exercise because it feels uncomfortable.
The dataset would then underrepresent exercise, while the discomfort may also influence the participant's willingness to exercise.
Researchers should evaluate both behavioral reactivity and the missing-data mechanism rather than treating them as entirely separate issues.
Can Monitoring Reactivity Be Eliminated Completely?
Not necessarily. Researchers may reduce some influences through device selection, standardized instructions, limited feedback, and appropriate familiarization.
However, participants' awareness of observation may remain.
Attempts to eliminate reactivity should also be proportionate to the research question and ethical requirements.
The methodological objective is not to guarantee that observation has no influence. It is to understand whether the influence could materially affect the intended inference and to design accordingly.
How Should Reactivity Be Reported?
Researchers should describe whether participants could view measurements, receive notifications, or interact with monitoring features.
Reports should identify relevant familiarization periods, monitoring instructions, and procedures used to standardize device exposure.
When reactivity is investigated empirically, the design and limitations of that investigation should be explained.
Where direct evidence is unavailable, potential reactivity should be presented as a plausible limitation rather than an established explanation for the findings.
When Might Passive Monitoring Be Unsuitable?
Passive monitoring may be unsuitable when the device substantially changes the behavior of interest and the research question requires observation of an unmonitored state.
For example, a conspicuous or uncomfortable wearable may be inappropriate for studying ordinary behavior during activities that it disrupts.
In such circumstances, researchers may consider less intrusive instruments, alternative observation methods, or a revised question.
However, the choice should be supported by evidence about measurement quality and participant experience rather than the assumption that another method is automatically free from reactivity.
Watch Out
Do not assume that automated recording is behaviorally neutral. A device may influence participants through awareness, feedback, discomfort, or perceived expectations. Conversely, do not attribute every change during monitoring to reactivity without considering alternative explanations.
