A Wandering Mind Is an
Matthew A. Killingsworth* and Daniel T. Gilbert
events that happened in the past, might happen
in the future, or will never happen at all. Indeed,
“stimulus-independent thought” or “mind wan-
dering” appears to be the brain’s default mode
of operation (1–3). Although this ability is a re-
markable evolutionary achievement that allows
people to learn, reason, and plan, it may have an
emotional cost. Many philosophical and religious
traditions teach that happiness is to be found by
living in the moment, and practitioners are trained
to resist mind wandering and “to be here now.”
These traditions suggest that a wandering mind is
an unhappy mind. Are they right?
Laboratory experiments have revealed a great
deal about the cognitive and neural bases of mind
wandering (3–7), but little about its emotional
consequences in everyday life. The most reliable
method for investigating real-world emotion is ex-
perience sampling, which involves contacting peo-
ple as they engage in their everyday activities and
asking them to report their thoughts, feelings, and
actions at that moment. Unfortunately, collecting
real-time reports from large numbers of people as
they go about their daily lives is so cumbersome
and expensive that experience sampling has rarely
been used to investigate the relationship between
mind wandering and happiness and has always
been limited to very small samples (8, 9).
We solved this problem by developing a Web
application for the iPhone (Apple Incorporated,
Cupertino, California), which we used to create
an unusually large database of real-time reports
of people as they went about their daily activ-
their iPhones at random moments during their
waking hours, presents them with questions,
and records their answers to a database at www.
trackyourhappiness.org. The database currently
contains nearly a quarter of a million samples
from about 5000 people from 83 different coun-
tries who range in age from 18 to 88 and who
collectively represent every one of 86 major oc-
To find out how often people’s minds wander,
what topics they wander to, and how those wan-
derings affect their happiness, we analyzed samples
from 2250 adults (58.8% male, 73.9% residing in
the United States, mean age of 34 years) who were
randomly assigned to answer a happiness question
(“How are you feeling right now?”) answered on a
continuous sliding scale from very bad (0) to very
good (100), an activity question (“What are you
doing right now?”) answered by endorsing one or
nlike other animals, human beings spend
a lot of time thinking about what is not
going on around them, contemplating
more of 22 activities adapted from the day recon-
struction method (10, 11), and a mind-wandering
question (“Are you thinking about something
other than what you’re currently doing?”) answered
with one of four options: no; yes, something pleas-
ant; yes, something neutral; or yes, something un-
pleasant. Our analyses revealed three facts.
First, people’s minds wandered frequently, re-
gardless of what they were doing. Mind wandering
occurred in 46.9% of the samples and in at least
30% of the samples taken during every activity
except making love. The frequency of mind wan-
dering in our real-world sample was considerably
higher than is typically seen in laboratory experi-
ments. Surprisingly, the nature of people’s activ-
ities had only a modest impact on whether their
minds wandered and had almost no impact on the
pleasantness of the topics to which their minds
Second, multilevel regression revealed that peo-
ple were less happy when their minds were wan-
dering than when they were not [slope (b) = –8.79,
P < 0.001], and this was true during all activities,
including the least enjoyable. Although people’s
minds were more likely to wander to pleasant topics
(42.5% of samples) than to unpleasant topics
(26.5% of samples) or neutral topics (31% of sam-
ples), people were no happier when thinking about
pleasant topics than about their current activity (b =
–0.52, not significant) and were considerably un-
happier when thinking about neutral topics (b =
–7.2, P < 0.001) or unpleasant topics (b = –23.9,
P < 0.001) than about their current activity (Fig. 1,
bottom). Although negative moods are known
to cause mind wandering (13), time-lag analyses
strongly suggested that mind wandering in our
sample was generally the cause, and not merely
the consequence, of unhappiness (12).
Third, what people were thinking was a better
predictor of their happiness than was what they
were doing. The nature of people’s activities ex-
plained 4.6% of the within-person variance in hap-
piness and 3.2% of the between-person variance in
happiness, but mind wandering explained 10.8%
of within-person variance in happiness and 17.7%
of between-person variance in happiness. The var-
iance explained by mind wandering was largely
independent of the variance explained by the na-
ture of activities, suggesting that the two were in-
dependent influences on happiness.
In conclusion, a human mind is a wandering
mind, and a wandering mind is an unhappy mind.
The ability to think about what is not happening
is a cognitive achievement that comes at an emo-
References and Notes
1. M. E. Raichle et al., Proc. Natl. Acad. Sci. U.S.A. 98, 676
2. K. Christoff, A. M. Gordon, J. Smallwood, R. Smith,
J. W. Schooler, Proc. Natl. Acad. Sci. U.S.A. 106, 8719
3. R. L. Buckner, J. R. Andrews-Hanna, D. L. Schacter,
Ann. N. Y. Acad. Sci. 1124, 1 (2008).
4. J.Smallwood, J.W. Schooler,Psychol.Bull. 132,946(2006).
5. M. F. Mason et al., Science 315, 393 (2007).
6. J. Smallwood, E. Beach, J. W. Schooler, T. C. Handy,
J. Cogn. Neurosci. 20, 458 (2008).
7. R. L. Buckner, D. C. Carroll, Trends Cogn. Sci. 11, 49 (2007).
8. J. C. McVay, M. J. Kane, T. R. Kwapil, Psychon. Bull. Rev.
16, 857 (2009).
9. M. J. Kane et al., Psychol. Sci. 18, 614 (2007).
10. D. Kahneman, A. B. Krueger, D. A. Schkade, N. Schwarz,
A. A. Stone, Science 306, 1776 (2004).
11. A.B.Krueger,D.A. Schkade,J.PublicEcon.92, 1833(2008).
12. Materials and methods are available as supporting
material on Science Online.
13. J. Smallwood, A. Fitzgerald, L. K. Miles, L. H. Phillips,
Emotion 9, 271 (2009).
14. We thank V. Pitiyanuvath for engineering www.
trackyourhappiness.org and R. Hackman, A. Jenkins,
W. Mendes, A. Oswald, and T. Wilson for helpful comments.
Supporting Online Material
Materials and Methods
18 May 2010; accepted 29 September 2010
Harvard University, Cambridge, MA 02138, USA.
*To whom correspondence should be addressed. E-mail:
Fig. 1. Mean happiness reported during each ac-
tivity (top) and while mind wandering to unpleas-
ant topics, neutral topics, pleasant topics or not
mind wandering (bottom). Dashed line indicates
mean of happiness across all samples. Bubble area
indicates the frequency of occurrence. The largest
bubble (“not mind wandering”) corresponds to
53.1% of the samples, and the smallest bubble
(“praying/worshipping/meditating”) corresponds to
0.1% of the samples.
12 NOVEMBER 2010VOL 330
on November 11, 2010
Supporting Online Material for
A Wandering Mind Is an Unhappy Mind
Matthew A. Killingsworth and Daniel T. Gilbert*
*To whom correspondence should be addressed. E-mail: email@example.com
Published 12 November 2010, Science 330, 932 (2010)
This PDF file includes:
Materials and Methods
Killingsworth & Gilbert / Supporting Online Material / Page 1 of 6
Supporting Online Material
Participants and Procedure
1. We stated “The most reliable method for investigating real-world emotion is experience
sampling.” Experience sampling is generally considered the “gold standard” for investigating
real-time emotion in vivo because it reduces or eliminates many of the potential biases
inherent in other survey methodologies (S1). For example, surveys normally require people
to report on emotional experiences long after they are over (“How did you feel last week?”)
and to integrate experiences over time (“In general, how happy have you been in the last
year?”). Although answers to such questions are good predictors of many consequential
outcomes, they are poor indices of people’s momentary emotional experiences because both
memory and integration are notoriously susceptible to error (S2, S3). In addition, surveys
normally ask people to report on many aspects of their lives (e.g., income, health, etc.), and
answering such questions can distort subsequent reports of emotional experiences (S4).
2. Participants volunteered for the study by signing up at trackyourhappiness.org. Our only
advertisement consisted of a link from our laboratory’s website to trackyourhappiness.org,
though the project did receive significant national press coverage. At initial signup,
participants completed an informed consent form in which they certified that they were at
least 18 years old. Participants then answered several questions about themselves, one of
which asked them to select their birth year from a drop-down menu. Twenty-seven of the
2,250 participants in our sample selected a birth year indicating they were less than 18 years
old. Because these participants had already certified that they were at least 18 years old, and
because selecting a birth year required a more complicated response than did certifying their
age, we considered these to be response errors and included these participants in the data set.
Killingsworth & Gilbert / Supporting Online Material / Page 2 of 6
However, we omitted these participants’ ages when computing the mean age of the
3. Next, participants were asked to indicate the times at which they typically woke up and went
to sleep, and how many times during the day they wished to receive a sample request (default
= 3, minimum = 1). A computer algorithm then divided each participant’s day into a number
of intervals equal to the number of samples to be requested, and a random time was chosen
within each interval. New random times were generated each day, and the times were
independently randomized for each participant. At each of these times, participants received
a notification on their iPhone, asking them to respond to a variety of questions about their
feelings, thoughts, behavior, and environment. Samples were collected on all days of the
4. Participants received requests for samples until they chose to discontinue participation. If 50
samples had been collected, sampling stopped for 6 months or until the participant requested
that it be restarted.
5. In each sample, different questions had different probabilities of being asked. The happiness
question and the activity question were asked in all samples, but the mind-wandering
question was asked in a randomly selected subset of samples. Only those samples that were
randomly assigned to include the mind-wandering question are reported here. The happiness
question was always asked before the mind-wandering question, and when the mind-
wandering question was asked it was always asked before the activity question. Other
questions were also asked but are not relevant to the present report. Participants contributed
an average of 7.9 samples (SD = 5.8, range = 1 to 39) to the present report.
Killingsworth & Gilbert / Supporting Online Material / Page 3 of 6
6. We computed “compliance rate” by dividing the number of samples received by the number
of samples requested during a participant’s “active period,” which we defined as the interval
between a participant’s first and last response. For example, if a participant completed 40
samples at the time of analysis but the 40th sample corresponded to the 50th request, then
that participant’s compliance rate would be 80%. Our mean compliance rate was 83% and
our median compliance rate was 93%. Calculation of compliance rates was based on all
samples, and not only those in which the mind-wandering question was asked.
7. The list of activities was adapted from the Day Reconstruction Method (S1, S5). Participants
appeared able to use this scheme to categorize their activities, as in only 6.7% of all samples
did participants indicate “other” as their primary activity.
8. Due to the nested structure of the data, analyses of sample-level data (with the exception of
within-person variance calculations as detailed below) were performed using multilevel
regression with samples nested within persons. These analyses were performed in R using
the function lmer from the lme4 package. Between-person analyses did not contain nested
data and were performed using OLS regression.
9. We stated that “the nature of people’s activities had only a modest impact on whether their
minds wandered.” Evidence for this statement includes the facts that (a) there was a
consistently high rate of mind-wandering across all activities except for making love, and (b)
the nature of people’s activities explained only 3.5% of the between-person variance in mind-
10. We stated that “the nature of people’s activities… had almost no impact on the pleasantness
of the topics their minds wandered to.” Evidence for this statement includes the facts that (a)
Killingsworth & Gilbert / Supporting Online Material / Page 4 of 6
although mind-wandering to an unpleasant topic was associated with less happiness (p <
.001), multilevel logistic regression revealed that the probability of mind-wandering to an
unpleasant topic was unrelated to a person’s activity (p > 0.25), and (b) person-level
regression revealed that differences in people’s activities explained less than 1% of the
between-person variance in the rate of mind-wandering to an unpleasant topic (Adj R2 =
0.0085, p < .05) or a neutral topic (Adj R2 = 0.0088, p < .01), and less than 2% of the
between-person variance in the rate of mind-wandering to a pleasant topic (Adj R2 = 0.016, p
11. We stated: “time-lag analyses strongly suggested that mind-wandering in our sample was
generally the cause—and not merely the consequence—of unhappiness.” We used multi-
level regression to determine whether there was a relationship between happiness in given
sample (T) and mind-wandering in the previous sample (T-1) and/or the next sample (T+1).
Table S1 presents the results of this analysis. The analysis yielded a significant negative
relationship between mind-wandering at T-1 and happiness at time T (Model 1), but no
significant relationship between mind-wandering at T+1 and happiness at time T (Model 2).
These findings were confirmed when mind-wandering at T-1 and T+1 were entered
simultaneously into the regression (Model 3) and when mind-wandering at T was added as a
control variable (Model 4). In short, across the four models we found a strong negative
relationship between mind-wandering at T-1 and happiness at T, but no relationship between
mind-wandering at T+1 and happiness at T. In other words, a person’s happiness was
strongly related to whether they had been mind-wandering in the previous sample, but was
unrelated to whether they were mind-wandering in the next sample. This is precisely what
one would expect if mind-wandering caused unhappiness, and precisely the opposite of what
Killingsworth & Gilbert / Supporting Online Material / Page 5 of 6
one would expect if mind-wandering and unhappiness were related only because unhappiness
causes mind-wandering. Although this does not preclude the possibility that unhappiness
also caused mind-wandering, such an effect appears to play at most a modest role in the
Table S1. Fixed effects estimates for Happiness at Time T
-1.07 (.331) **
-1.21 (.354) **
-1.34 (.345) *** Mind-Wandering
-0.365 (.331) -0.294 (.357) -0.511 (.348)
-9.04 (.376) ***
Note: ** P < .01; *** P < .001.
12. We stated: “The nature of people’s activities explained 4.6% of the within-person variance in
happiness and 3.2% of the between-person variance in happiness, but mind-wandering
explained 10.8% of within-person variance in happiness and 17.7% of between-person
variance in happiness.” To compute the within-person variation, happiness scores were
centered on each person's mean happiness. Then, OLS regression was used at the sample
level to analyze the amount of within-person variance in happiness explained by mind-
wandering and by activity. To compute the between-person variation, person-level mean
values of happiness, proportion of time spent in each mind-wandering state, and proportion
of time spent in each activity were computed. Then, OLS regression was used at the person
Killingsworth & Gilbert / Supporting Online Material / Page 6 of 6
level to analyze the amount of between-person variance in happiness explained by mind-
wandering and by activity.
13. We stated: “The variance explained by mind-wandering was largely independent of the
variance explained by the nature of activities, suggesting that the two are independent
influences on happiness.” While activity explained 3.2% and mind-wandering explained
17.7% of between-person variance in happiness, together they explained 19.9% of between-
person variance. While activity explained 4.6% and mind-wandering explained 10.8% of
within-person variance in happiness, together they explained 14.8% of between-person
variance. In each case, the variance explained by activity and mind-wandering together
approached the sum of the variances that each factor explained on its own.
S1. D. Kahneman, A. B. Krueger, D. A. Schkade, N. Schwarz, A. A. Stone, Science 306,
D. Kahneman, in Well-being: The foundations of hedonic psychology., D. Kahneman, E.
Diener, N. Schwarz, Eds. (Russell Sage Foundation, New York, 1999), pp. 3-25.
B. L. Fredrickson, D. Kahneman, J. Pers. Soc. Psychol. 65, 45 (1993).
N. Schwarz, F. Strack, in Well-being: The foundations of hedonic psychology., D.
Kahneman, E. Diener, N. Schwarz, Eds. (Russell Sage Foundation, New York, 1999),
A. Krueger, D. Schkade, Journal of Public Economics 92, 1833 (2008).