The 2026 study found that participants’ distributions of sensed day patterns and transitions were more similar to themselves across observation windows than to other participants. It does not establish routine persistence as a personality trait or explain an individual’s routine. In one long-window analysis, conscientiousness had a small association with change in routine distributions; short-window student analyses found no significant personality predictors.
What does the 2026 study establish?
The 2026 study supports a descriptive finding: in three datasets, participants’ distributions of recurring sensed day patterns and their transitions were more similar to themselves across observation windows than to other participants. It does not establish routine persistence as a personality trait, or explain why any individual follows a particular pattern. The distinction matters: recurrence is a feature of recorded behavior; interpreting that recurrence as a broader tendency is a further claim.
In the June 15, 2026 version 2 preprint of “Quantifying the Persistence of Daily Routines,” researchers analyzed passive sensing from 1,086 participants across more than 153,000 person-days. The data came from three longitudinal studies, with different populations and observation periods. The paper reports that each person’s mix of routine types, as well as the way those types followed one another, was more consistent within that person across split windows than in comparisons with others. This is evidence of relative stability in the measured patterns over weeks or months. It does not mean that a person repeated the same day, or that their routines never changed.
The study’s headline language about person-specific behavioral “fingerprints” should therefore be read within the boundaries of its method. The result concerns clusters built from selected sensed behaviors, not a complete account of daily life, preference, or motive. Nor does a pattern that persists identify its cause: schedules, obligations, available opportunities, cues, and personal tendencies may all contribute. The preprint offers evidence that measured day-pattern signatures can be distinctive and persistent in these samples. It does not show that persistence itself is a stable personality characteristic, and it cannot turn a reader’s routine into a character verdict.
What did the study count as a routine?
The study treated a routine as a recurring pattern across measured features of a day, not a person’s complete account of what they did or why. In the June 15, 2026 v2 preprint, Quantifying the Persistence of Daily Routines, researchers grouped days by passively collected indicators of sleep, movement, and device use. They tracked how often a participant’s days fell into each derived pattern and how one pattern was followed by another. Here, “routine” means a statistical summary of selected signals, not proof that someone prefers or consciously repeats a schedule. A routine type is a data-derived cluster of days with similar measured features. The researchers summarized each day across morning, afternoon, evening, and night, using features such as sleep timing or duration, activity, and screen use. Clustering groups observations that resemble one another on the variables supplied. The resulting types therefore depend on the recorded features and method. They are not personality categories or complete descriptions of every day assigned to them. The paper’s routine signature has two parts. First, it records the proportion of a participant’s days assigned to each routine type, ordered from most to least frequent. This captures the mix of day patterns, but not their sequence. Second, a transition signature summarizes the probabilities of moving from one type of day to another on the next day. It adds adjacent-day order, though it does not describe every longer rhythm, such as a weekly cycle or gradual change. Two people could have a similar mix of day types but different transitions between them. The analysis combined 1,086 participants and more than 153,000 person-days from three longitudinal datasets. Tesserae followed information workers in the United States; MoMo-Mood included patients with depressive episodes and healthy controls in Finland; and GLOBEM followed university students in the United States. The populations and collection periods differed, so the combined total is not a representative sample of everyone. The paper reports that eight types captured the dominant structure in each dataset, a result for these data rather than a universal count of people’s routines. The signals differed across datasets. Sleep was measured with wearable devices in Tesserae and GLOBEM, while MoMo-Mood estimated it from phone lock and unlock status. Movement came from step counts in two datasets and accelerometer activity in MoMo-Mood. Device use was represented by screen-use duration. The paper compares the phone-based sleep estimate with actigraphy during a two-week validation phase, reporting adequate correlation with slight bias. That check does not make all features equivalent across cohorts. The clusters are summaries built from available proxies, not direct observations of a whole day. Important parts of everyday life were outside the measure. The authors excluded location because GPS data were substantially missing, and communication records were too sparse for a useful broader picture. The analysis did not capture full activities, their reasons, unrecorded obligations, or how a participant understood the day. A late start in sensed data cannot distinguish a chosen schedule from a shift requirement, poor sleep, travel, or missing device readings. Repeated screen use also does not show whether it reflects work, leisure, or another activity. The study’s “routine” is narrower than the ordinary-language idea of someone’s usual day. It is a pattern in selected, passively sensed features, summarized into recurring day types and their frequencies or adjacent transitions. This framework describes whether measured patterns recur in the studied data. On its own, it cannot establish a preference, motive, habit, or personality tendency for any person. Those interpretations require other evidence about behavior, setting and constraints, and whether a pattern appears across different contexts. The authors normalized features within each participant, focusing on variation from that person’s own average.
How persistent were the recorded patterns?
The study found that each participant’s routine patterns were more similar to themselves across two observation windows than to patterns from other participants. This held for two different summaries: the share of a person’s days assigned to each routine type, and the probabilities of moving from one type of day to another on the next day. In Quantifying the Persistence of Daily Routines, the authors call these routine and transition signatures. “Persistent” therefore means comparatively consistent across the study’s two windows; it does not mean that a person repeated one identical schedule every day.
The comparison matters. The researchers split each participant’s observation period into two halves, calculated both signatures separately for each half, then compared the distance between that person’s two signatures with distances involving other people. A smaller distance means the two summaries are more alike. Within-person similarity was greater than the between-person reference for both signatures in all three datasets. This is evidence that the measured distributions and transitions carried information about whose days were being summarized. It is not evidence that every activity was stable, that participants followed a fixed timetable, or that the pattern had a particular motive.
The day types themselves were data-derived clusters, not categories chosen in advance to represent familiar lifestyles. The paper reports that roughly eight clusters captured the dominant structure in each dataset. Across participants, the two most frequent types together accounted for more than half of a participant’s days on average. That descriptive result helps explain how persistence can coexist with variation: a person’s record could contain several recurring kinds of day, with some appearing more often than others. The allocation signature summarizes those proportions but does not preserve the order of days. The transition signature adds a different detail by summarizing adjacent-day moves, yet it too compresses a complex schedule into a limited set of probabilities.
The observation windows also differed. For the GLOBEM student dataset, the comparison used two 30-day halves. For Tesserae and MoMo-Mood, it used two 135-day halves. Those are not interchangeable tests of persistence: a month can capture shorter-term recurrence, while a longer window allows more opportunities for a range of day patterns to appear. This supports the bounded claim across sampled datasets, but it does not establish how long a signature would remain similar outside those windows or under a major schedule change.
The useful interpretation is that a person may have a relatively distinctive mix of multiple day patterns and a recurring way of moving among them, rather than one unchanging daily routine. The study’s passive measures describe sensed sleep, movement, and device-use patterns; they do not explain why a pattern recurs. A regular work start, for example, could be consistent because a schedule constrains it, while other days vary. That is an illustration of the distinction, not a participant case from the paper. Persistence identifies a repeatable structure in the observations. It leaves the causes, personal meaning, and connection to personality open for further evidence.
What does the personality result actually say?
The personality result is narrow: in the current version of “Quantifying the Persistence of Daily Routines,” conscientiousness had a small association with less stable distributions of sensed day patterns in one pooled, long-window analysis. In the separate short-window analysis of university students, no measured personality trait significantly predicted either routine distribution distance or transition distance. To see why, keep the paper’s two outcomes separate. Routine-signature distance compares the proportions of a participant’s days assigned to each data-derived routine type across the first and second halves of observation. Transition-signature distance compares how that participant moved between types on adjacent days across those halves. A greater distance means the two halves were less similar on that measure. It is not a score for liking novelty, resisting habits, or choosing change. For the long-window analysis, the researchers pooled Tesserae and MoMo-Mood, with 135 days in each half. Their model related within-person distance to age, gender, and Big Five traits. Conscientiousness was positively associated with routine-signature distance: the reported coefficient was 0.015, with a 95% confidence interval from 0.004 to 0.026 and p = .006. The coefficient is small, and it is an association in observational data, not evidence that conscientiousness caused the distributions to change. The result did not extend to the other persistence measure. In the same long-window model, conscientiousness was not a significant predictor of transition distance: the coefficient was 0.006, its confidence interval ranged from −0.002 to 0.015, and p = .155. Someone might keep a similar share of work-like and non-work-like days over time while changing the order in which those day types follow one another, or the reverse. A result for the mix of days cannot be restated as a result about transitions. Nor did the personality association appear in the student analysis. GLOBEM participants were observed in 30-day windows, and the authors analyzed repeated windows with a mixed-effects model. They report that no demographic or personality variables significantly predicted either routine-signature or transition-signature distance there. This is not proof that personality never relates to routines. It means the evidence in this paper varies by outcome and observation design: the small conscientiousness association appears in one pooled long-window model and is absent from the short-window models. Measurement also limits how broadly to read the comparison. Personality was assessed at baseline, using the BFI-10 in Tesserae and GLOBEM and the NEO-60 in MoMo-Mood; these are not identical instruments. The study therefore relates brief or differing trait measures to later summaries of passively sensed sleep, movement, and device use. It does not repeatedly ask how a person experienced a day, why they followed a schedule, or whether a recurring pattern felt chosen. The defensible interpretation is consequently modest. The v2 paper reports one small, model-dependent link between conscientiousness and change in the distribution of sensed day types, alongside null personality findings for transitions in that analysis and for both signatures in the short-window student models. It does not establish a stable personality explanation for routine persistence. A person’s consistent pattern could coexist with many different trait tendencies, and the aggregate coefficient cannot classify an individual. Do not infer that a person with a stable routine is more conscientious, or that conscientious people generally have less persistent routines. The study measured similarity in recorded behavior across windows; motives and personal meaning remain outside what its result can establish.
Is a recurring routine the same as a habit?
No. A recurring routine and a habit overlap in ordinary speech, but they describe different units of behavior in these studies. The 2026 routine paper groups whole days according to sensed patterns across sleep, movement, and device use. Habit research usually follows a chosen action repeated in a particular context, asking whether it becomes easier to initiate or more automatic. A person can have recurring kinds of days without any single action being automatic; one action can become habitual even while the rest of the day varies.
Repetition alone does not reveal why something happens. The same route each morning may follow a fixed work shift; making tea after lunch may follow a cue such as finishing the meal. Both recur, but for different possible reasons. The sensed day-pattern study does not observe such motives or cues, so it cannot distinguish them.
In “Context Stability in Habit Building Increases Automaticity and Goal Attainment,” researchers followed 95 university students building a chosen study habit over six weeks, collecting 2,482 repetitions. Participants defined an action, such as summarizing lecture notes, and a completion goal. One group was instructed to repeat in a stable context; another varied place and time. Context covered physical surroundings and a point in the day. After repetitions, participants reported context stability, automaticity, and goal completion. In this setting, greater context stability predicted more reported automaticity and goal attainment.
The paper also analyzed app records for 308 habits from 218 users. There, context was measured rather than assigned, and stability again predicted automaticity and goal attainment. The student study manipulated instructions; the app analysis examined naturally varying reports. Together, they suggest stable cues can matter while a specific action is being built. They do not show that assigning a stable context guarantees habit formation, or explain whole-day routine distributions. The students were building study behaviors, not a representative sample of daily routines.
A different approach appears in “Creature of Habit: A self-report measure of habitual routines and automatic tendencies in everyday life.” Its 27-item Creature of Habit Scale asks people about habitual responding through two aspects: routine behavior and automatic responses. This is a questionnaire about how respondents characterize their everyday tendencies. It is not a sensor-based classification of their days, nor a record of whether a chosen action followed a cue on each occasion. The measure therefore broadens the view of habit beyond a single study task, while still relying on self-report and addressing a different question from passive sensing.
These methods should not be collapsed into one claim that a person has a “persistent routine.” Day-pattern analysis asks which kinds of sensed days recur and how they follow one another. Habit-building research asks whether a selected action in a stable context becomes more automatic and meets its repetition goal. The questionnaire asks how people report routine behavior and automatic responding. Each captures a different level: day, action, or reported tendency. Evidence at one level does not settle another.
For one repeated action, notice what tends to happen before it, whether its time or place is stable, and whether beginning feels deliberate. These observations may distinguish a planned action from one that feels cue-linked, though a short log is not a formal test. For whole-day patterns, look across behaviors and ask what schedule, access, or obligations they share. A recurring sensed day is descriptive; deciding whether habit or personality helps explain it requires different observations.
Sources: Context Stability in Habit Building Increases Automaticity and Goal Attainment; Creature of Habit: A self-report measure of habitual routines and automatic tendencies in everyday life
What can routine data say about a personality tendency?
A recurring pattern can be consistent with a personality tendency, but the 2026 daily-routine study does not identify which tendency explains any participant’s data. It compares sensed patterns across people and time windows; it does not establish why one person follows a particular schedule. To connect routine with personality, observations need to include behavior and some account of its setting. A group-level association is not an individual explanation.
This is where experience sampling asks a different question. Rather than infer a person’s inner reasons from a trace of movement or device activity, it collects repeated reports in daily settings. A report might capture what someone was doing, how they felt, or how they understood the situation at that moment. In two studies summarized in “Unveiling an exclusive link: Predicting behavior with personality, situation perception, and affect in a preregistered experience sampling study,” personality traits, affect, and perceived situations each contributed uniquely to explaining self-reported behavioral states. The researchers also found overlap between situation perceptions and affect, while the two were not interchangeable. The result supports considering these parts together when studying everyday expression; it does not say which one caused a particular action, or provide a rule for reading a stranger’s routine.
Repeated reports also make it possible to ask how behavior varies within one person’s life, not only how people differ on average. “Experience Sampling Methods: A Modern Idiographic Approach to Personality Research” describes snapshots in natural settings as a way to study an individual’s patterns over time, including links between situations and behavior. This is a methodological argument, not proof that brief self-tracking reveals someone’s personality. Self-reports depend on what a person notices, remembers, and chooses to report. Frequent prompts can burden participants or change what they attend to. Reports add context that passive sensing may miss, but remain partial records.
Consider an illustrative contrast. A person starts work at roughly the same time each weekday. If a fixed shift sets that hour, repetition may say more about an obligation than a preference for structure. Now imagine that, on workdays and free days, the person also chooses a quiet reading period at a similar point in the day. This is more consistent with a preference that travels across settings, but it still does not prove a personality tendency. Available time, household responsibilities, fatigue, or the reading material could help explain it. This example is a reasoning aid, not a documented participant or research finding.
The distinction is between recurrence and interpretation. The 2026 study’s sensed day patterns can show that a person’s mix of sleep, movement, and device-use patterns recurs relative to other participants. They do not show whether the person chose those patterns, enjoyed them, or could readily change them. Experience sampling can bring reported behavior and perceived context closer together, but its answers remain reports rather than a complete account of causes. Neither kind of evidence turns one repeated action into a trait label.
To make a personality interpretation more plausible, look for a tendency across more than one behavior and setting. Does someone plan ahead only when deadlines are imposed, or also for a personal project? Do they keep a reading period when their schedule shifts, or only when work ends at the same hour? These comparisons help distinguish a broad pattern from a schedule-bound routine. They are not a validated test, and a few observations cannot rule out unrecorded constraints.
A useful next observation is not simply another count of the same routine. Note what happened, where and when, whether timing was required or chosen, and what changed when the context changed. If a pattern persists through meaningful changes in schedule and activity, it becomes more reasonable to consider a broader tendency among possible explanations. If it appears only under one stable obligation, the situation remains a strong alternative. Routine data can invite a personality question; repeated behavior across varied contexts gives it a fairer test.
Sources: Unveiling an exclusive link: Predicting behavior with personality, situation perception, and affect in a preregistered experience sampling study; Experience Sampling Methods: A Modern Idiographic Approach to Personality Research; Quantifying the Persistence of Daily Routines
Which observations would make a personality interpretation more or less plausible?
A personality interpretation gains plausibility when a pattern appears across behaviors and settings, not simply when one schedule repeats. It becomes less persuasive when behavior closely tracks an imposed timetable, reliable cue, access limit, or temporary demand. They are clues, not a checklist that identifies a trait or settles its cause. Separate five things: behavior; setting; whether timing or place was chosen or required; any cue just before it; and what changes when conditions change. “Starts work at eight” describes timing; “starts at eight because the shift begins then” adds a constraint. “Plans the day after breakfast” describes an action and a possible cue. Neither alone establishes a personality tendency. “Quantifying the Persistence of Daily Routines” grouped sensed day patterns and measured their recurrence; it did not identify the motives or preferences behind them. Consider two patterns. Someone may leave home at the same time on weekdays because a required commute fixes the departure. That regularity says little by itself about a preference for structure: the schedule could produce the same pattern across people with different tendencies. Now consider someone who chooses a short planning session before work and also plans before a personal project on days off. Repetition across work and free time makes a broader preference for planning more plausible than the commute alone. Still, deadlines, a learned cue, or practical need may explain both instances. Either routine can coexist with personality influences; neither is a verdict. Repeating a day pattern differs from repeating a cued action. “Context Stability in Habit Building Increases Automaticity and Goal Attainment” followed specified study behaviors over six weeks and examined an app dataset. It reported that greater stability in repetition time and place was associated with more reported automaticity and completion of a chosen repetition goal. This concerns selected actions during habit formation. Note cues, but do not assume one explains a whole-day routine distribution or reveals a trait. An action may be cued, deliberately planned, required by circumstances, or some combination. Repeated reports can add information absent from passive sensing. “Personality Traits, States, and Situations” found in experience-sampling studies that trait measures, affect, and perceived situations each contributed to explaining self-reported behavioral states. It cannot tell a reader what their own routine means, but recording behavior with its situation may sharpen the question: does it recur when the setting changes, and does the person report the same reason or feeling each time? Self-reports depend on what people notice and report, so context helps without making interpretation certain. Keep a practical comparison modest. For one or two weeks, note a specific behavior rather than a label such as “I am disciplined.” Record when and where it happens, whether timing was chosen or imposed, what cue came beforehand, and whether the behavior changed when schedule, access, or demand changed. Compare different contexts, such as a workday and a day off, without treating one unusual day as decisive. The aim is to see whether a tendency appears through multiple behaviors or a condition explains most repetition, not to score a trait. One short log cannot separate these explanations conclusively. A required schedule may hide a preference because it leaves no room to choose; a chosen routine may still reflect convenience or a cue. A personality interpretation gains plausibility through repeated examples across meaningful context changes, especially if behavior remains similar despite different immediate demands. It weakens if the pattern reliably changes when a constraint or cue changes. Even then, personality is one possible contribution among circumstances, habits, resources, and goals. Describe what recurs, name its conditions, and keep the trait explanation provisional until it extends beyond one routine.
Sources: Quantifying the Persistence of Daily Routines; Context Stability in Habit Building Increases Automaticity and Goal Attainment; Unveiling an exclusive link: Predicting behavior with personality, situation perception, and affect in a preregistered experience sampling study
What should I conclude about my own pattern?
Start with a description you could check: “I begin focused work soon after breakfast on most weekdays,” Then ask what happens when the setting changes. Does the pattern appear on workdays and free days, or mainly when a shift, commute, deadline, or household responsibility sets the timing? Is it tied to a chosen cue, or to an obligation that leaves little choice? These details help separate recurrence from its possible explanation. “Quantifying the Persistence of Daily Routines,” the 2026 study, found participants’ distributions of sensed day patterns and transitions were more similar to themselves across observation windows than to other participants. That supports person-specific persistence in the measured data. It does not explain why a particular routine recurs or establish that it expresses a personality trait. Recurrence may be consistent with a tendency, but that interpretation becomes more plausible when similar preferences or behaviors appear across different situations. Even then, the observation is a clue, not a verdict. Choose one pattern and compare it in two genuinely different contexts. Record what you did, whether the timing was chosen or imposed, what cue came first, and what changed when the schedule changed. Keep the conclusion provisional; one contrast cannot settle the cause. The Personality Profile learning library can help you explore broad trait models and the behaviors they describe. If seeing several tendencies together would help, the private Context Profile can serve as a reflection prompt, not a validated measure or explanation of your routine. Collect observations across contexts before deciding what the pattern may mean.
Questions readers ask
Does the 2026 study show that routine persistence is a personality trait?
No. It found that sensed day-pattern distributions and transitions were more similar within participants across observation windows than between participants. That describes measured behavior in the studied datasets; it does not establish a personality trait or explain why an individual’s routine recurs.
What did the study mean by routine persistence?
It compared each participant’s distribution of data-derived day types and their transitions between day types across two observation windows. Greater similarity means those summaries changed less between windows. It does not mean the person followed the same schedule each day.
Does the study show that conscientious people have less stable routines?
Not as a general conclusion. The current v2 reported a small association between conscientiousness and greater change in routine distributions in a pooled long-window analysis. The short-window student analyses found no significant personality predictors for either signature. The results are observational and analysis-dependent.
Sources and notes
- Quantifying the Persistence of Daily Routines
Supports the 2026 study’s dataset, routine and transition signatures, within-person comparisons, and analysis-dependent personality findings.
- Creature of Habit: A self-report measure of habitual routines and automatic tendencies in everyday life
Supports the description of the 27-item Creature of Habit Scale as measuring routine behavior and automatic responses by self-report.
- Context Stability in Habit Building Increases Automaticity and Goal Attainment
Supports the findings on context stability, reported automaticity, and goal completion for specified repeated study habits.
- Unveiling an exclusive link: Predicting behavior with personality, situation perception, and affect in a preregistered experience sampling study
Supports the claim that traits, affect, and perceived situations each contributed to explaining self-reported behavioral states in two experience-sampling studies.
- Experience Sampling Methods: A Modern Idiographic Approach to Personality Research
Supports the methodological role and limits of repeated everyday reports for studying within-person behavior and situation patterns.
Apply it to your own pattern
Compare your tendencies with the conditions around a recurring pattern
From this guide: The study cannot show whether an individual’s routine reflects a broader tendency, a stable cue, or a schedule constraint.
A recurring work pattern can reflect both personal tendencies and the conditions around a role. The study cannot show how those influences combine for you. The private Context Profile can help you reflect on several tendencies together; use it as a prompt for comparing them with your work history and the conditions each role required, not as a career recommendation.