Having Too Little or Too Much Time Is Linked to Lower Subjective Well-Being

Having Too Little or Too Much Time Is Linked to Lower

Subjective Well-Being

Marissa A. Sharif

, Cassie Mogilner

, and Hal E. Hershfield

Department of Marketing, The Wharton School, University of Pennsylvania

Marketing and Behavioral Decision Making Areas, Anderson School of Management, University of California, Los Angeles

Many people living in modern society feel like they do not have enough time and are constantly searching

for more. But is having limited discretionary time actually detrimental? And can there be downsides of hav-

ing too much discretionary time? In two large-scale data sets spanning 35,375 Americans and two experi-

ments, we explore the relationship between the amount of discretionary time individuals have and their

subjective well-being. We ﬁnd and internally replicate a negative quadratic relationship between discretion-

ary time and subjective well-being. These results show that whereas having too little time is indeed linked to

lower subjective well-being caused by stress, having more time does not continually translate to greater sub-

jective well-being. Having an abundance of discretionary time is sometimes even linked to lower subjective

well-being because of a lacking sense of productivity. In such cases, the negative effect of having too much

discretionary time can be attenuated when people spend this time on productive activities.

Keywords: discretionary time, happiness, life satisfaction, subjective well-being, time

Supplemental materials: https://doi.org/10.1037/pspp0000391.supp

Between their many obligations, people today feel like they do not

have enough time and want more. This “time famine” reaches across

the globe (Hamermesh & Lee, 2007) and is particularly acute among

Americans, who report feeling more time-constrained than ever before

(Carroll, 2008; Roxburgh, 2004). A nationwide poll shows that nearly

half of Americans report they do not have enough time to do what they

want to do (Newport, 2016). Hoping to lessen the strain, many people

search for ways to save time to increase hours spent doing what they

want. For i nstance, there a re more than 35,000 books available on

Amazon aiming to improv e time management (e.g., How to Get 12

Hours Out of an 8-Hour Day), and 50% of Americans spend money to

buy out of chores, such as cooking, shopping, and household mainte-

nance (Whillans et al., 2017). Although prioritizing time over money

and spending money to buy more free time have been empirically

linked to greater happiness (Hershﬁeld et al., 2016; Whillans et al.,

2017, 2016), would actually having more time make people better

off? Pushing this question further, is it possible to have too much dis-

cretionary time? Across two large-scale data sets and two experiments,

we examine the relationship between the amount of discretionary time

people have and their subjective well-being.

The likely harm of having too little time is straightforward. People

who work longer hours and have a greater proportion of their sched-

ules consumed by obligations have less time to do what they want.

Less discretionary time means less time spent on activities that are

linked to greater happiness (e.g., socializing and engaging in active

leisure; Csikszentmihalyi & Hunter, 2003; Kahneman et al., 2004;

Lathia et al., 2017; Mogilner, 2010). In addition, feeling pressed for

time takes its toll. People who report greater feelings of time-stress

exhibit more unhealthy behaviors, such as eating poorly and not

exercising (Banwell et al., 2005; Strazdins et al., 2011). Those who

report feeling time-stressed also report being less happy (Kasser &

Sheldon, 2009; Masuda et al., 2020), more depressed (Roxburgh,

2004), and more emotionally exhausted (Teuchmann et al., 1999).

Although a negative effect of being temporally impoverished

seems likely, what is the effect of having an abundance of time?

Ample time for discretionary activities may have a diminishing

effect on people’s enjoyment of those activities. Both fortunately

and unfortunately, people are prone to hedonic adaptation, making

them grow accustomed not only to life’s pains but also to life’spleas-

ures (Frederick & Loewenstein, 1999; Lyubomirsky et al., 2005;

Wilson & Gilbert, 2008). For instance, despite socializing being

among the most enjoyable ways to spend time, time spent on social

activities has been shown to have diminishing returns for subjective

well-being (Kushlev et al., 2018). Indeed, excessive access to

enjoyed activities leads people to savor them less (Kurtz, 2008;

Quoidbach et al., 2015). Therefore, spending hours upon hours sim-

ply doing what one wants may lose its positive impact on happiness.

Beyond a reduced positive effect, could there ever be a negative

effect? Is it possible to have too much discretionary time? Are there

cases in which having additional discretionary hours is associated

Marissa A. Sharif https://orcid.org/0000-0001-7910-7000

OSF Link for data, syntax, materials, and preregistrations: https://osf.io/

hw85m/.

Correspondence concerning this article should be addressed to Marissa

A. Sharif, Department of Marketing, The Wharton School, University of

Pennsylvania, 3730 Walnut Street, Jon M. Huntsman Hall 751,

Philadelphia, PA 19103, United States, or to Cassie Mogilner, Marketing

and Behavioral Decision Making Areas, Anderson School of Management,

University of California, Los Angeles, 110 Westwood Plaza, Gold Hall B-

515, Los Angeles, CA 90095, United States. Email: masharif@wharton

.upenn.edu or [email protected]

Journal of Personality and Social Psychology:

Personality Processes and Individual Differences

ISSN: 0022-3514 https://doi.org/10.1037/pspp0000391

This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

with lower subjective well-being? Emerging work tangentially sug-

gests so. People dread being idle and have been observed as happier

when busied by a task (Hsee et al., 2010). Recent research suggests

that busyness has become a status symbol, signaling competence,

ambition, and being in high demand (Bellezza et al., 2017). If people

derive satisfaction from being productive with their time (i.e., spend-

ing it on worthwhile activities; Keinan & Kivetz, 2011), the effect of

having an abundance of time in one’s daily life may be more insidi-

ous than mere boredom. With too much discretionary time, people

may infer lack of productivity and purpose (Csikszentmihalyi, 2000)

—thus feeling less happy and less satisﬁed in their lives.

Even though many people in modern society feel time-stressed

and want more time, we hypothesized that actually having more

discretionary time would not increasingly make people better off.

More speciﬁcally, we predicted a negative quadratic relationship

between discretionary time and subjective well-being, such that

beyond a certain amount, more discretionary time would not be

further associated with greater subjective well-being. We further

expected that in some cases, a large amount of discretionary time

may actually be associated with lower subjective well-being,

depending on how this time is spent. In particular, we predicted

that an abundance of time spent on nonproductive discretionary

activities would exhibit a negative effect on subjective well-being.

However, if people spent their discretionary time in productive

ways, we predicted that the negative effect of having too much dis-

cretionary time would be attenuated.

Drawing on prior research (Holbrook & Lehmann, 1981), we

deﬁne discret ionar y time as the amount of time spent on leisur e

activities or other pursuits where the primary function is the use of

time for pleasure or another intrinsicall y worthwhile purpose. There-

fore, discretiona ry time is not simply whatever time remains outside

of paid work hours. For one, not everyone works for pay (e.g., stay-

at-home parents and retirees). And among those who do, the amount

of time one chooses to spend working might be well over what is

obligatory (Snir & Zohar, 2008). Additionally, at least some of one’s

time outside of work is likely to be absorbed by other obligations,

including household chores, going to the dentist, taking a child to the

dentist, or standing in line at the DMV

(Eriksson et al., 2007). Thus,

we deﬁne and operationalize discretionary time as the number of

hours a person spends in a day doing what they want.

To examine the relationship between discretionary time and subjec-

tive well-being, we ﬁrst analyze two large-scale data sets representing

adults from across the United States (Studies 1 and 2). We then sup-

plement these data with two mental simulation experiments to gain

insight into mechanisms driving the observed effect. In particular, in

Study 3, we measure two theoretically motivated mediators: greater

stress from having too little discretionary time and a lacking sense of

productivity from having too much discretionary time. In Study 4, we

then manipulate whether a high (vs. moderate) amount of discretion-

ary time is spent on product ive (vs. nonprod uctive ) activities to test

for an attenuation of the negative effect of having too much time.

Study 1: National Study of the Changing Workforce

Method

We analyzed the data of 13,639 working Americans who partici-

pated in the National study of the Changing Workforce (NSCW)

between 1992 and 2008.

2,3

The NSCW surveyed representative sam-

ples of the nation’s labor force on four occasions: in 1992; in 1997; in

2002; and in 2008 (ages 18–99, M

age

= 42.27; 47.2% male; 78.8%

Caucasian; 58.1% married; 41.7% have children; 36.9% with at least

a bachel or's degree; M

income

= $46,398.47). To qualify to take this sur-

vey, par ticip ants had to be working at a paid jo b or operating an

income-producing business as part of the civilian labor force (see

Galinsky et al., 1993 for more information about this survey).

Among the survey’s many questions, participants reported their

amount of discretionary time (“On average, on days when you’re

working, about how many hours [minutes] do you spend on your own

free-time activities?”; M =1.80hr,SD = 1.82, Median = 1 hr, Min =

0 hr, Max = 20 hr; see Figure S1 for distribution) and their subjective

well-being, which was measured as life satisfaction (“All things con-

sidered, how do you feel about your life these days? Would you say

you feel 1 = very satisﬁed ,2=somewhat satisﬁed,3=somewhat dis-

satisﬁed,or4=very dissatisﬁed?”). For our analysis, we reverse-

coded the life satis factio n measure, such that larger numbers corre-

spond with greater subjective well-being (M =3.24,SD =.71).

Whereas many papers have examined this data set, https://www

.familiesandwork.org/research/workforce-research-national-study-of

-the-changing-workforce, research-to-date has not examined the

relationship between the amount of discretionary time people have

and their life satisfaction.

Results

We conducted an OLS regression predicting life satisfaction

from the linear and quadratic terms of people’s reported amount of

discretionary time. Consistent with our prediction, we found a sig-

niﬁcant negative quadratic relationship between the amount of dis-

cretionary time people have and their subjective well-being (N =

13,639, B = .003, SE = .001, t[13636] = 5.28, p , .001, 95%

CI [.004, .002], R

= .004; Figure 1, Table 1).

We also examined whether the quadratic term explained more

variance in the model than did the signiﬁcant linear term alone (N =

13,639, B =.017,SE =.003,t[13639] = 5.11, p , .001, 95% CI

[.011, .024], R

= .002). Indeed, by adding the quadratic term in the

model, there was a signiﬁcant increase in the variance explained: an

increase of .002 in the R

, F Change (1, 13636) = 27.85, p , .001.

The predicted negative quadratic relationship held when con-

trolling for gender, age, parental status, marital status, race, the

natural log of respondent income, employment status (i.e.,

self-employed or not), and education

(N = 11,649, B = .003,

SE = .001, t[11638] = 3.72, p , .001, 95% CI [.011, .004],

= .058; Table 1). We did not ﬁnd any consistent interactions

between these covariates and the predicted quadratic relationship

across studies.

Other than work, these activities were among those viewed as least

discretionary (and most obligatory) in the crowdsourcing study we

conducted to determine our measure of discretionary time for Study 2. See

Table S7 in the online supplemental materials for the complete set of

activities listed from most to least discretionary.

Studies 1 and 2 analyzes large datasets collected by a third party with

no identifying information about the participants; thus, IRB approval was

not necessary for these studies.

Data are not publicly available.

Gender, age, parental status, marital status, race, education, income,

and self-employment status are all mean-centered.

SHARIF, MOGILNER, AND HERSHFIELD

This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Outliers

The predicted negative quadratic effect held up to a series of

robustness checks. Namely, the effect persisted when we excluded

individuals who reported discretionary time four or more standard

deviations away from the mean (n = 108 with 9.1þ hours of discre-

tionary time; without covariates: N = 13,531, B = .010, SE =.002,

t[13528] = 6.02, p , .001, 95% CI [.014, .007], R

=.006;

with covariates: N =11,564,B = .010, SE =.002,t[11553] =

3.39, p , .001, 95% CI [.023, .006], R

=.060;Table S1).

The effect also persisted when we excluded individuals who were

identiﬁed as outliers using Cook’s Distance (without covariates: n =

290 with Cook’s distance greater than 4/N, N = 13,349, B = .004,

SE =.001,t[13346] = 5.75, p , .001, 95% CI [.005, .003], R

= .007; with covariates: n = 436 with Cook’s distance greater than

4/N, N =11,213,B = .004, SE =.001,t[11202] = 4.61, p ,

.001, 95% CI [.005, .002], R

=.072;Table S2). These checks

conﬁrm that the observed negative quadratic effect was not driven

by a few participants who reported having very large amounts of

discretionary time.

U-Shape Test

We then used Simonsohn’s (2018) two-lines approach (http://

webstimate.org/twolines/) to test for an inverted U-shape in the

relationship between discretionary time and subjective well-being.

This method conﬁrmed that for low values of discretionary time,

the regression line was positive and statistically signiﬁcant (b =

.090, z = 9.291, p , .001); however, for high values of discretion-

ary time, the regression line was negative but did not reach statisti-

cal signiﬁcance ( b = .005, z = .899, p = .368; break point = 2

hr). Thus, as we predicted, having more discretionary time did not

show a continued positive effect on subjective well-being. How-

ever, in this dataset, we did not observe our predicted signiﬁcant

negative effect among people who had an abundance of discretion-

ary time. In light of our results from Study 2, we later propose

why we likely did not observe the predicted drop in this dataset.

Figure 1

Scatterplot of the Relationship Between Discretionary Time and Subjective Well-Being in Study 1

0 5 10 15 20

Subjective Well-Being

Discretionary Hours

Note. For ease of visualization, a jitter was added to subjective well-being scores and discretionary hours. Fit

line represents the negative curvilinear relationship between discretionary time and subjective well-being.

Table 1

Regression Results of Study 1: The Influence of Discretionary

Time on Subjective Well-Being

Variable (1) (2)

Hours of discretionary time 0.044*** 0.057***

(0.006) (0.007)

Hours of discretionary time squared 0.003*** 0.003***

(0.001) (0.001)

Male 0.025

(0.013)

White 0.012

(0.016)

Age 0.002*

(0.001)

Married 0.293***

(0.014)

Children 0.034*

(0.015)

4-year college 0.034*

(0.014)

Natural log-transformed income 0.047***

(0.007)

Self-employed 0.078***

(0.018)

Constant 3.18*** 3.15***

(.010) (.011)

.004 .058

Note. (1) The coefficients reported above are the unstandardized coeffi-

cients. (2) All predictor variables, except Hours of Discretionary time and

Hours of Discretionary time Squared, are mean-centered.

p , .10. * p , .05. ** p , .01. *** p , .001.

DISCRETIONARY TIME AND SUBJECTIVE WELL-BEING

This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Study 2: American Time Use Survey

The dataset we analyzed in Study 2 advances our investigation

in several important ways. First, Study 1 only included working

Americans. In Study 2, we leveraged data from an even larger and

more representative sample of Americans, including those not in

the workforce.

Second, Study 1 relied on a fairly subjective measure of discre-

tionary time. Participants used their own interpretation of what

constitutes “free-time ac tivities” to report their amount of discre-

tionary time on an average workday. However, an activity that

one person views as discretionary (e.g., going for a run) might be

viewed as a tortuous chore by someone else, or an activity that

someone views a s a tedious, incessant obligation (e.g., cooking)

might be viewed as an enriching hobby by others. Although the

results of Study 1 were informative in capturing individuals ’

views of th eir activities in a typical day, in Study 2 we sought to

replic ate the predicted negative quadratic effect using a more

objective and conservative measure of discretionary time. Based

on the activities that most pe ople (i.e., more than 90%) view as

discretionary, we calculated the amount of time e ach of the tens

of thousands of individuals in the dataset spent on discretionary

activities in a given day. This level of detail in the dataset allowed

us to further explore the role that different types of discretionary

activities play in affecting the relationship between the amount of

discretionary time individuals have and their subjective well-

being.

Method

We analyzed the data of 21,736 Americans who participated in

the American Time Use Survey (ATUS) between 2012 and 2013,

the years in which our key variables were administered (ages

15–85, M

age

= 47.92; 44.5% male; 79.3% Caucasian; 47.7% mar-

ried; 43.5% have children; 33.5% with at least a bachelor's degree;

57.8% employed full time; M

income

= $52,597.74). Data are avail-

able at: https://www.bls.gov/tus/#database. In answering the

ATUS, respondents provide a detailed account of the activities

that ﬁlled their prior 24 hr—indicating the time period and dura-

tion of each activity. We assessed discretionary time by calculat-

ing the amount of time each individual spent on discretionary

activities in a day.

Because there is no standard deﬁnition for which speciﬁc activ-

ities count as discretionary, we used a crowdsourcing platform

(Amazon’s Mechanical Turk) to determine the activities that most

people (i.e., more than 90%) consider to be discretionary time.

Speciﬁcally, we conducted a preregistered study in which we ﬁrst

provided a deﬁnition of discretionary time (“time spent on leisure

activities or other pursuits where the primary function is the use of

time for pleasure or some other intrinsically worthwhile purpose”)

and then asked participants (N = 500; M

age

= 34.59, SD

age

= 10.83;

50.8% male) to indicate for each activity in the American time

Use Survey whether it was representative of discretionary time or

not (0 = “is NOT discretionary time,” 1=“IS discretionary time”).

The categories of activities that at least 90% of the sample consid-

ered discretionary were Relaxing & Leisure (e.g., doing nothing,

watching TV, listening to the radio, playing games); Socializing &

Communicating with Others (e.g., hanging out with family, hang-

ing out with friends); Arts & Entertainment Other than Sports

(e.g., attending a comedy club, attending an art gallery, attending a

movie); Travel Related to Socializing, Relaxing, & Leisure; Perso-

nal Activities (e.g., having sex, making out); Attending Sporting/

Recreational Events (e.g., watching sports); Playing Sports with

Household and Nonhousehold Children (e.g., riding bikes with

child, strolling with child); and Participating in Sports, Exercise,

or Recreation (e.g., biking, playing basketball, ﬁshing, running,

golﬁng, doing yoga, working out). See Table S7 of the online sup-

plemental materials for more details about Study 2, including the

full list of 139 activities and the percentage of participants who

identiﬁed each as discretionary.

Applying this crowdsourced metric for determining which

activities count as discretionary, we calculated the total amount

of time each American time Use Survey respondent spent engag-

ing in discretionary activities over the previous 24 hr (M =5.48

hr, SD =3.70,Median=4.92hr,Min=0hr,Max=23.98hr;see

Figure S2 for the distribution).

Subjective well-being was

assessed with a ladder-style question used in prior research

(Kahneman & Deaton, 2010): “Please imagine a ladder with

steps numbered from 0 at the bottom to 10 at the top. The top of

the ladder represents the best possible li fe for you. If the top step

is 10 and the bottom step is 0, on which step of the ladder do you

feel you per sonally stand at the present time?” (M = 7.11, SD =

2.03). In doing so, unlike previous research examining this data,

https://www.bls.gov/tus/research.htm, we were able to uniquely

examine how amount of time spent on discretionary acti vities

affects subjective well-being.

Results

We conducted an OLS regression predicting life satisfaction

from the linear and quadratic terms of our calculated amount of

individuals’ discretionary t ime. Replicat ing the r esults from

Study 1, we found a signiﬁcant negative quadratic relationship

between the amount of discretionary time people have and their

subjective well-being (N = 21,736, B = .004, SE = .001,

t[21733] = 4.88, p , .001, 95% CI [.005, .002], R

= .003;

Figure 2 and Table 2). This effect held when controlling for gen-

der, age, parental status, marital status, race, education, natural

log-transformed respondent income, and employment status, all

of which are mean-centered ( N = 20,275, B = .003, SE = .001,

t[20264] = 4.23, p , .001, 95% CI [.005, .002], R

= .032;

Table 2).

Inclusion of the predicted negative quadratic term explained

more variance in the model than did the linear term alone, which

contrary to Study 1 was signiﬁcant and negative (N = 21,736, B =

.026, SE = .004, t[21734] = 6.89, p , .001, 95% CI [.033,

.018], R

= .002). By adding the quadratic term in the model,

there was a signiﬁcant increase in the variance explained: an

increase of .001 in the R

, F Change (1, 21733) = 23.77, p , .001.

It is apparent in Figure S2 that a small portion of respondents had a

very large number of discretionary hours in their day. For instance, 5.9% of

the sample had more than 12 hours of discretionary time. Although it may

seem difﬁcult to have this large amount of discretionary time in a day, it is

not unreasonable. The ATUS asks individuals about a randomly selected

day, which might be one in which the person had an unusually large

number of discretionary hours (with little sleep and/or few obligations)

SHARIF, MOGILNER, AND HERSHFIELD

This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Robustness Checks

The predicted negative quadratic effect held up to a series of

robustness checks.

Outliers

The effect held when we excluded individuals who reported dis-

cretionary time four or more standard deviations away from the

mean (n = 13 with 20.28þ hours of discretionary time; without

covariates: N = 21,723, B = .004, SE = .001, t[21720] = 4.42,

p , .001, 95% CI [.005, .002], R

= .003; with covariates: N =

20,262, B = .003, SE = .001, t[20251] = 3.76, p , .001, 95%

CI [.005, .002], R

= .031; Table S3). It also held when we

excluded outliers using Cook’s Distance (n = 1,088, without cova-

riates: N = 20,648, B = .003, SE = .001, t[20645] = 3.21, p =

.001, 95% CI [.005, .001], R

= .001; n = 858, with covariates:

N = 19,417, B = .002, SE = .001, t[19406] = 3.01, p = .003,

95% CI [.004, .001], R

= .036; Table S4). These results sug-

gest that the predicted effect is not reliant on the inclusion of out-

liers who had an extremely large amount of discretionary time

that day.

Discretionary Time Calculations

Further, the effect held when we loosened the restriction of

what constitutes a discretionary activity from those activities iden-

tiﬁed by at least 90% of participants as discretionary to those that

at least 75% of participants identiﬁed as discretionary (M = 6.08

hr, SD = 3.76, Median = 5.62 hr, Min = 0 hr, Max = 23.98 hr;

without covariates: N = 21,736, B = .004, SE = .001, t[21733] =

5.56, p , .001, 95% CI [.006, .003], R

= .003; with covari-

ates: N = 20,275, B = .004, SE = .001, t[20264] = 4.57, p ,

.001, 95% CI [.005, .002], R

= .031; Table S5).

The effect also held when we used an alternate measure of dis-

cretionary time. For this, we calculated the total amount of time

respondents spent doing any activity that fell within the American

time Use Survey’spredeﬁned category of “Socializing, Relaxing,

and Leisure” (e.g., socializing and communicating with others,

Table 2

Regression Results of Study 2: The Influence of Discretionary

Time on Subjective Well-Being

Variable (1) (2)

Hours of discretionary time 0.027* .035

(0.011) (0.012)

Hours of discretionary time squared 0.004*** 0.003***

(0.001) (0.001)

Male 0.189***

(0.029)

White 0.051

(0.035)

Age 0.006***

(0.001)

Married 0.577***

(0.031)

0.076*

Children (0.036)

0.117***

(0.032)

4-year college 0.030***

(0.008)

Natural log-transformed earnings 0.279**

Employed (0.087)

Constant 7.134*** 7.070***

(0.034) (0.037)

.003 .032

Note. (1) The coefficients reported above are the unstandardized coeffi-

cients. (2) All predictor variables, except Hours of Discretionary time and

Hours of Discretionary time Squared, are mean-centered.

p , .10. * p , .05. ** p , .01. *** p , .001.

Figure 2

Scatterplot of the Relationship Between Discretionary Time and Subjective Well-Being in Study 2

0 5 10 15 20

Subjective Well-Being

Discretionary Hours

Note. For ease of visualization, a jitter was added to subjective well-being scores. Fit line represents the nega-

tive curvilinear relationship between discretionary time and subjective well-being.

DISCRETIONARY TIME AND SUBJECTIVE WELL-BEING

This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

attending or hosting social events, etc.). Using this alternative mea-

sure, we again observed the predicted negative quadratic relation-

ship (M =5.04hr,SD = 3.57, Median = 4.42 hr, Min = 0 hr, Max =

23.98 hr; without covariates: N =21,736,B = .004, SE =.001,t

[21733] = 4.64, p , .001, 95% CI [.006, .002], R

=.005;

with covariates: N = 20,275, B = .003, SE =.001,t[20264] =

3.91, p , .001, 95% CI [.005, .002], R

=.033;Table S6).

Working Versus Not

Whereas Study 1 only included working Americans, Study 2

included individuals who were working (n = 12,558) as well as

those who were not working (n = 9,178), including those who

were employed but absent from work that day (n = 543), those

who were unemployed (n = 1,159), and those not in the labor force

(n = 7,476). Even though those who were working had less discre-

tionary time on average (M = 4.39 hr, SD = 3.20) than those who

were not working (M = 6.96 hr, SD = 3.82), we found the pre-

dicted negative quadratic relationship between individuals’

amount of discretionary time and subjective well-being both

among workers (N = 12,558, B = .003, SE = .001, t[12555] =

2.23, p = .025, 95% CI [.005, .000], R

= .001) and non-

workers (N = 9,178, B = .004, SE = .001, t[9175] = 3.43, p =

.001, 95% CI [.007, .002], R

= .005). As additional evidence

that the effect was not driven by people who were involuntarily

unemployed and thus whose abundance of discretionary time was

potentially unwelcomed and uncoordinated with their social net-

work (Young & Lim, 2014), we still observed the signiﬁcant nega-

tive quadratic effect when we excluded only individuals who

self-identiﬁed as being “unemployed” (excluded n = 1,159;

Discretionary time of Unemployed Participants

= 6.40 hr, SD = 3.77; N =

20,576, B = .004, SE = .001, t[20574] = 5.03, p , .001, 95%

CI [.006, .003], R

= .003).

We also found that the effect was not driven by retirees. As evi-

dence, the predicted negative quadratic relationship persisted when

we excluded individuals who were above the standard retirement

age of 66 (excluded n =3,709;M

Discretionary time of individuals over 66

7.50 hr, SD =3.65;N =18,027,B = .004, SE =.001,t[18024] =

4.61, p , .001, 95% CI [.006, .002], R

=.007).Furthermore,

whether or not an individual was above the standard age of retire-

ment or not did not moderate the negative quadratic term: the inter-

action effect was not signiﬁcant (B =.000,SE =.002,t[20261] =

.198, p =.843,95%CI[.005, .004]).

Weekday Versus Weekend

Further, even though people on average have more discretionary

time on the weekends (n = 11,039; M =6.16hr,SD = 3.71) than on

the weekdays (n = 10,697; M =4.77hr,SD = 3.56), the negative

quadratic effect persisted among people reporting their discretionary

activities either on a weekend (N = 11,039, B = .004, SE = .001, t

[11036] = 3.98, p , .001, 95% CI [.006, .002], R

= .003) or

on a weekday (N = 10,697, B = .004, SE = .001, t[10694] = 3.21,

p = .001, 95% CI [.006, .001], R

= .004). This analysis suggests

that people can experience too little time and too much discretionary

time irrespective of whether it’s during the week or on the weekend.

U-Shape Test

We again used Simonsohn’s (2018) two-lines procedure to test

for an inverted U-shape relationship between discretionary time

and subjective well-being. Here, the regression line for low values

of discretionary time was again positive and statistically signiﬁ-

cant (b = .068, z = 3.105, p = .002). However, unlike in Study 1,

the regression line for the high values of discretionary time was

negative and statistically signiﬁcant, as we had initially predicted

(b = .041, z = 6.657, p , .001; break point = 3.42 hr). That is,

in this dataset, we again found evidence for having too little time;

at the low end of the continuum for discretionary time, having

more time was associated with greater subjective well-being.

However, we also found evidence for having too much time; at the

high end of the continuum for discretionary time, having more

time was associated with lower subjective well-being.

For a clearer view into the results, we supplemented this analy-

sis with a histogram of respondents’ amount of discretionary time

binned per 30 minutes. Based on Figure 3, it appears that subjec-

tive well-being increases as discretionary time increases between

zero and two hours, peaks between two and ﬁve hours, and

decreases above ﬁve hours. These results provide a rough approxi-

mation that having less than two hours of discretionary time is

“too little,” whereas having more than ﬁve hours of discretionary

time is “too much.”

This identiﬁcation of what counts as too little or too much dis-

cretionary time offers a potential clue into why we did not observe

the “too much” effect in Study 1. Although both studies showed a

signiﬁcant negative quadratic effect, the U-shape test in Study 1

showed that the positive relationship between discretionary time

and life satisfaction leveled off after the break point, whereas the

U-shape test in Study 2 showed that the relationship became sig-

niﬁcantly negative after a point. If having more than ﬁve hours of

discretionary time counts as too much, we see that while 48.1%

of participants in Study 2 qualiﬁed as having too much, only 3.5%

of participants in Study 1 did. See Figure S1 and Figure S2 in the

online supplemental materials for the distribution of discretionary

time in each study. The very small proportion of individuals in

Study 1 with too much discretionary time may not have provided

enough power to detect a signiﬁcant negative effect on subjective

well-being from having an overabundance of discretionary time.

Yet another difference between the two studies was the measure

for discretionary time. Whereas Study 1 asked participants to

report their amount of discretionary time on average, in Study 2

we calculated the amount of time people actually spent on discre-

tionary activities on a given day. With its larger dataset that repre-

sents individuals both in and out of the workforce, and its more

precise and objective measure of discretionary time, we suggest

that the pattern of results found in Study 2 is more conclusive.

Types of Discretionary Time

With the robustness of the predicted negative quadratic effect,

along with the signiﬁcant test conﬁrming the inverted U-shape, the

results of Study 2 suggest that not only is it possible to have too lit-

tle discretionary time, it is also possible to have too much. We

found that having an abundance of discretionary time is associated

with lower subjective well-being. We then wondered whether the

way people spent their discretionary time might affect this relation-

ship. That is, if people were to spend their discretionary time in

more worthwhile ways, we may see the too much time effect atte-

nuated. For instance, prior research has identiﬁed socializing and

active leisure as particularly positive uses of time (Csikszentmihalyi

SHARIF, MOGILNER, AND HERSHFIELD

This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

& Hunter, 2003; Kahneman et al., 2004; Mogilner et al., 2018).

Research has also shown that being productive with one’s time pro-

duces greater subjective well-being (Etkin & Mogilner, 2016;

Keinan & Kivetz, 2011). To explore whether the type of discretion-

ary time moderates the relationship between amount of discretion-

ary time and subjective well-being, we reexamined the data from

the ATUS. For this analysis, we calculated the amount of time par-

ticipants spent in these more worthwhile ways (Kuykendall et al.,

2018): on social (vs. solo) discretionary activities, active (vs. pas-

sive) discretionary activities, and productive (vs. nonproductive)

discretionary activities.

To determine which uses of discretionary time most people con-

sider to be social (vs. solo), active (vs. passive), and productive

(vs. nonproductive), we again conducted a crowdsourcing study

on Amazon Mechanical Turk among a separate set of participants

(N = 901; ages 18–74, M

age

= 37.55; 60.7% male). In this preregis-

tered study, we provided the same deﬁnition of discretionary time

and presented participants with the discretionary activities identi-

ﬁed in our previous crowdsourcing study (see Table S8–S10). To

reduce the chance of respondent fatigue, each participant was pre-

sented with a random subset of 31 activities from the full list of 94

discretionary activities. For each discretionary activity, partici-

pants indicated whether it was social versus solo (“A social activ-

ity is one that would be more likely to involve engaging with other

people (vs. alone)”;0=solo activity,1=social activity); whether

it was active versus passive (“An active activity is one that is

physically or mentally engaging. The opposite of an active activity

is a passive activity”;0=passive activity,1=active activity) and

whether it was productive versus nonproductive (“By productive,

we mean that you consider this use of discretionary time to not be

“wasted.” This use of time might feel useful, accomplished, fulﬁll-

ing, helpful, purposeful, and/or worthwhile”;0=not productive,

1=productive).

As preregistered, we coded discretionary activities to be social,

active, or productive if 90% or more of the participants rated the

activity as such. Otherwise, the activity was coded as solo, pas-

sive, or nonproductive. See Tables S8–S10 in the online supple-

mental materials for a complete list of the activities and the

percentage of participants who categorized them as social, active,

and productive. Notably, this crowdsourcing study revealed a large

degree of overlap between the discretionary activities that were

viewed as active and productive, with all but one of the productive

activities (i.e., Hobbies) also identiﬁed as active. These results,

along with our rationale that productive discretionary time is likely

to be worthwhile by being physically or mentally engaging, led us

to combine these two categories into discretionary activities that

are productive (vs. nonproductive).

Based on this classiﬁcation process, the discretionary activities

counted as social were Socializing and Communication with

Others, Playing Sports with Household Children, Playing Volley-

ball, Playing Football, Playing Racquet Sports, and Playing Bil-

liards. The discretionary activities we counted as productive were

Working Out, Running, Hobbies, Bowling, Participating in Water

Sports, Playing Volleyball, Playing Rugby, Participating in Eques-

trian Sports, Playing Baseball, Weightlifting/Strength Training,

Biking, Playing Sports with Household or Non-Household

Figure 3

Histogram of the Relationship Between Discretionary Time and Subjective Well-Being in Study 2

6.4

6.5

6.6

6.7

6.8

6.9

7.1

7.2

7.3

7.4

30 min

1 hr

1.5 hr

2 hr

2.5 hr

3 hr

3.5 hr

4 hr

4.5 hr

5 hr

5.5 hr

6 hr

6.5 hr

7 hr

7.5 hr

8 hr

8.5 hr

9 hr

9.5 hr

10 hr

10.5 hr

11 hr

11.5 hr

12 hr

12.5 hr

13 hr

13.5 hr

14 hr

14.5 hr

Subjecitve Well-Being

Discretionary Time

Note. Discretionary time is split into bins of 30 minutes. Error bars represent standard error of the mean.

The pattern of the results is consistent when the active and productive

categorizations are kept separate. See the online supplemental materials for

the separated results.

DISCRETIONARY TIME AND SUBJECTIVE WELL-BEING

This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Children, Doing Aerobics, Participating in Martial Arts, Personal/

Private Activities (e.g., making out, having sex), Rollerblading,

Wrestling, Fencing, Playing Hockey, Playing Soccer, Playing Rac-

quet Sports, and Golﬁng. Using these classiﬁcations, we calculated

the amount of time each participant in the ATUS spent on discre-

tionary activities that were social (vs. solo) and productive (vs.

nonproductive) that day.

For each type of discretionary time, we examined whether there

was a signiﬁcant negative quadratic effect by conducting OLS

regressions predicting subjective well-being from the linear and

quadratic term of each type of discretionary time. Following this,

we examined if there was a signiﬁcant inverted U-shape for each

type of discretionary time.

Social Versus Solo Discretionary Time

For solo discretionary time, we found a signiﬁcant negative

quadratic effect (N = 21,736, B = .003, SE = .001, t[21733] =

4.04, p , .001, 95% CI [.005, .002], b = .079, R

= .004;

with covariates: N = 20,275, B = .003, SE = .001, t[20264] =

3.87, p , .001, 95% CI [.005, .002], b = .078, R

= .032).

Simonsohn’s (2018) two-lines approach conﬁrmed an inverted

U-shape. For low values of solo discretionary time, the relation-

ship with subjective well-being was positive and statistically sig-

niﬁcant (b = .05, z = 1.91, p = .056); however, for high values of

solo discretionary time, the regression line was negative and statis-

tically signiﬁcant (b = .05, z = 7.59, p , .001; break point =

2.53 hr).

However, for social discretionary time, we did not ﬁnd a signiﬁ-

cant negative quadratic effect (N = 21,736, B = .003, SE = .003, t

[21733] = 1.11, p = .269, 95% CI [.008, .002], b = .015, R

.001; with covariates: N = 20,275, B = .000, SE = .003, t[20264] =

.012, p = .991, 95% CI [.005, .005], b = .000, R

= .031). This

nonsigniﬁcant quadratic effect was thus signiﬁcantly smaller than

for solo discretionary time activities, comparing the standardized

coefﬁcients of the quadratics,

F(1, 21731) = 8.26. We instead only

found a signiﬁcant positive linear effect of social discretionary time

on life satisfaction (N = 21,736, B = .033, SE = .009, t[21734] =

3.63, p , .001, 95% CI [.015, .052], R

= .001; with covariates:

N = 20,275, B = .041, SE = .009, t[20265] = 4.36, p , .001, 95% CI

[.023, .059], R

= .030).

Together these results suggest that when people spend their dis-

cretionary time socially, more is better. We only observed the too

much time effect when that discretionary time did not offer the

value of social connection.

Productive Versus Nonproductive Discretionary time

For nonproductive discretionary time, we found a signiﬁcant

negative quadratic effect (N = 21,736, B = .004, SE = .001,

t[21733] = 4.67, p , .001, 95% CI [.005, .002], b = .095,

= .004; with covariates: N = 20,275, B = .003, SE = .001,

t[20264] = 3.91, p , .001, 95% CI [.005, .002], b = .082,

= .032). Furthermore, we found evidence for the inverted

U-shape. For low values of nonproductive discretionary time, the

regression line was positive and statistically signiﬁcant (b = .059,

z = 2.43, p = .015); and for high values of nonproductive discre-

tionary time, the regression line was negative and statistically sig-

niﬁcant (b = .044, z = 7.49, p , .001; break point = 3 hr).

For productive discretionary time, we also found a signiﬁcant

negative quadratic effect (N = 21,736, B = .016, SE = .007,

t[21733] = 2.36, p = .018, 95% CI [.030, .003], b = .030,

= .002; with covariates: N = 20,275, B = .016, SE = .007,

t[20264] = 2.27, p = .023, 95% CI [.030, .002], b = .029,

= .033). However, the negative quadratic effect was signiﬁ-

cantly smaller for productive (vs. nonproductive) uses of discre-

tionary time, comparing the standardized coefﬁcients of the

quadratics, F(1, 21731) = 6.16 (p = .013).

Furthermore, we did

not ﬁnd evidence for the inverted U-shape. For low values of pro-

ductive discretionary time, the regression line was positive and

statistically signiﬁcant (b = 4.09, z = 2.41, p = .016); however, for

high values of productive discretionary time, the regression line

did not reach statistical signiﬁcance (b = .032, z = 1.35, p = .176;

break point = .22 hr).

Altogether, this exploration into types of discretionary time sug-

gests that how people spend their time does affect the relationship

between the amount of discretionary time people have and their

subjective well-being. In particular, the manner in which people

spend their discretionary time appears to largely determine

whether we observe the negative effect of having too much discre-

tionary time. Although an abundance of discretionary time spent

on solo and nonproductive activities did produce a negative effect

on subjective well-being, discretionary time spent on activities

that were social or productive did not. With an already vast litera-

ture asserting the emotional beneﬁts of social connection (Diener

& Seligman, 2002; Gable & Bromberg, 2018), we were particu-

larly interested in the moderating role of discretionary activities

that were productive (vs. nonproductive), revealing that only when

people spent large amounts of discretionary time nonproductively

did they report lower subjective well-being. In light of prior work

showing that people enjoy increased satisfaction from feeling pro-

ductive and busy (Etkin & Mogilner, 2016; Hsee et al., 2010), this

ﬁnding suggests that a lacking sense of productivity may be one

underlying mechanism driving the negative effect of having too

much time. We further examine this ﬁnding in the subsequent two

studies.

Study 3: Discretionary Time Experiment

Even though our analyses of the large nationally representative

data sets in Studies 1 and 2 controlled for likely covariates (e.g.,

To compute this statistic, we z-scored each of the following terms: amount

of discretionary time people spent on solo activities, amount of discretionary

time people spent on social activities, the amount of discretionary time people

had on solo activities squared, the amount of discretionary time people had on

social activities squared, and life satisfaction. We then conducted a linear

regression predicting life satisfaction from this set of predictor variables. The

statistic examines whether the quadratic terms for social vs. solo discretionary

time are statistically different from each other in this regression model.

To compute this statistic, we z-scored each of the following terms:

amount of discretionary time people spent on productive activities, amount

of discretionary time people spent on nonproductive activities, the amount

of discretionary time people spent on productive activities squared, the

amount of discretionary time people spent on nonproductive activities

squared, and life satisfaction. We then conducted a linear regression

predicting life satisfaction from this set of predictor variables. The statistic

examines whether the quadratic terms for productive vs. nonproductive

discretionary time are statistically different from each other in this

regression model.

SHARIF, MOGILNER, AND HERSHFIELD

This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

employment status, marital and parental status, income, etc.), it

remains plausible that inherent differences between people with

low, moderate, and high amounts of discretionary time account for

their differing levels of subjective well-being. For instance,

because we did not have data on mental health status, it is possible

that individuals with depressive symptoms and who are dissatis-

ﬁed with life may also be unmotivated to fulﬁll their obligations,

which would leave them with large amounts of discretionary time.

To address such alternate explanations, in Study 3, we randomly

assigned participants to the experience of having a low, moderate,

or high amount of discretionary time.

Study 3 thus employed a between-subjects experimental design.

Because it is infeasible to manipulate the actual number of discre-

tionary and nondiscretionary hours individuals typically have in their

day-to-day lives, we relied on a mental simulation manipulation that

required participants to vividly imagine and describe having a given

number of discretionary hours every day for an extended period of

their lives. This mental simulation approach helped ensure that par-

ticipants considered the nuanced and actual experience of being in

that situation, rather than reporting based on their superﬁcial assump-

tions (Morewedge et al., 2010; O’Brien et al., 2018).

In the previous studies, we had to rely on the measures that

were available in those data sets. In Studies 1 and 2, subjective

well-being was measured using items that assessed satisfaction in

life. However, life satisfaction is just one component of subjective

well-being, and it is one that is more cognitive in nature (Kahne-

man & Deaton, 2010). In these next studies, we used a more com-

prehensive measure of subjective well-being that encompasses

both life satisfaction and positive emotion (Diener et al., 2017;

Lyubomirsky et al., 2005).

Perhaps most important, Study 3 sought to illuminate the under-

lying mechanism for the negative quadratic effect observed in Stud-

ies 1 and 2. In light of previous work showing the detrimental

emotional consequences of feeling time-stressed (Kasser & Shel-

don, 2009; Roxburgh, 2004), we proposed that compared with hav-

ing a moderate amount of discretionary time, having a small

amount of discretionary time would make people feel more stressed

and thus experience lower subjective well-being. On the other end

of the continuum, in light of the moderating role of discretionary

time spent productively (vs. nonproductively) in Study 2, as well as

work showing the beneﬁts of being productive (Etkin & Mogilner,

2016; Hsee et al., 2010; Keinan & Kivetz, 2011), we proposed that

compared with having a moderate amount of discretionary time,

having too much discretionary time would make people feel unpro-

ductive and thus experience lower subjective well-being.

Method

As preregistered, 2,550 American adults were recruited on Ama-

zon’s Mechanical Turk to participate in this study.

Two thousand,

ﬁve hundred sixty-ﬁve participants (M

age

= 37.83; 40.9% male;

58.1% employed; 47.7% married; 50.6% have children; 69.2% with

at least a bachelor's degree; M

HoursWork/Week =

30.26 hours

;

DiscretionaryTime/Day =

3.54

hr) completed the three-cell (amount

of discretionary time: low vs. moderate vs. high) between-subjects

experiment.

After being presented with the deﬁnition of discretionary time,

“time spent on leisure activities or on other pursuits where the pri-

mary function is the use of time for pleasure or some other

intrinsically worthwhile purpose,” participants were led to men-

tally simulate having a given amount of discretionary time every

day for at least 6 months of their lives. Participants were randomly

assigned to have a low (15 minutes per day), moderate (3.5 hr per

day), or high (7 hr per day) amount of discretionary time.

speciﬁed that they would have this amount of time consistently

over a minimum six-month period to ensure that participants in

the high time condition did not imagine a single vacation day or

weekend day that was uncharacteristically relaxed, and that partic-

ipants in the low time condition did not imagine a single workday

that was uncharacteristically busy.

For each condition, participants were prompted to imagine and

vividly describe what it would be like to have the given amount of

discretionary time (e.g., how they would spend this time, what

they would be doing during the other portion of the day, and how

they would be feeling that day) after 1 day, 1 month, 3 months,

and 6 months of having that amount of discretionary time every

day.

Participants then reported their subjective well-being by rating

the extent to which they would experience enjoyment, happiness,

and satisfaction on a 7-point Likert scale, with 1 = would not feel

this at all to 7 = would feel this a lot (a

low

= .96, a

moderate

= .94,

high

= .94). To test our proposed mechanisms, we then asked par-

ticipants to rate on the same 7-point scale their stress (stressful,

rushed, exhausting; a

low

= .91, a

moderate

= .91, a

high

= .88) and

their sense of productivity (productive, purposeful, I accomplished

a lot; a

low

= .90, a

moderate

= .91, a

high

= .93) during this period of

their lives.

Lastly, participants were asked a series of questions capturing

features of their imagined time period: employment status, the

number of hours worked per day, occupation, marital status, pa-

rental status, whether on vacation, whether in-between jobs, and

whether they had experienced any signiﬁcant life changes right

before this period of their life.

All materials, data, syntax, and preregistrations for both Study 3

and 4 can be found at https://osf.io/hw85m/. Both Studies 3 and 4

were conducted with Institutional Review Board approval from

the University of California, Los Angeles under protocol IRB#

16–000935-AM-00006, “Time, Money, and Happiness.”

Results

The results of this experiment were c onsistent with the nega -

tive quadratic effect observed in Studies 1 and 2. We conducted

an OLS regression predicting subjective well-being from two

dummy variables representing conditio ns, with the moderate

discretionary time co ndition serving as the reference group.

Participants r eported lo wer subjec tive well-bei ng from ha ving a

The sample size was determined based on a power analysis of a pilot

of this study.

Two participants reported working more hours than the number of

hours in a week and were excluded from this statistic.

One hundred ﬁfty-four participants reported more than 24 hours of

discretionary time in a day and were excluded from this statistic.

These amounts were determined based on the data from Study 2. The

moderate amount (3.5 hours) was set to be halfway between the identiﬁed

range of optimal discretionary time (between 2 and 5 hours). The high

amount of discretionary time was set to be double this amount of time (7

hours).

DISCRETIONARY TIME AND SUBJECTIVE WELL-BEING

This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

low amount of discretionary time than from having a moderate

amount of discretionary time (M

Low

=5.52, SD = 1.55 vs. M

Moderate

6.00, SD =1.14;B = .480, t[2562] = 7.36, p , .001, 95% CI

[.607, .352], d = .35), and participants reported lower subjective

well-being from having a high amount of discretionary time than

from having a moderate amount of discretio nary time (M

High

=5.88,

SD =1.30vs.M

Moderate

=6.00,SD =1.14;B = .119, t[ 2562] =

1.84, p = .066, 95% CI [.247, .008], d = .10). This effect held

when controlling for characteristics of the imagined time period,

including employment status (employed or not), number of working

hours,

marital status (married or not), parental status (have children

or not), whether they imagined being on a vacation or not, whether

they imagined being between jobs or not, and whether they imagined

there was any signiﬁcant life change right before the imagined time

period or not (Low vs. Moderate Discretionary Time: B = .463, t

[2554] = 7.16, p , .001, 95% CI [.590, .337]; High vs. Moder-

ate Discretionary Time: B = .152, t[2554] = 2.35, p =.02,95%CI

[.279, .025]).

We also found support for our proposed mechanisms. Participants

reported they would feel signiﬁcantly more stress from having a low

amount of discretionary time than from having a moderate amount

Low

= 3.24, SD =1.89vs.M

Moderate

= 2.56, SD =1.62;B = .673,

t[2562] = 8.17, p , .001, 95% CI [.511, .835], d = .38); and these

feelings of stress partially mediated the negative effect of having a

low (vs. moderate) amount of discretionary time on subjective well-

being (B = .26, SE = .04, 95% CI [.339, .193])basedon10,000

bootstrap samples with a reduced direct effect (B = .22, SE = .06,

t[2562] = 3.78, p , .001, 95% CI [.330, .104]). Additionally,

participants reported they would feel signiﬁcantly less productive

having a high amount of discretionary time than having a moderate

amount (M

High

=5.11,SD =1.71vs.M

Moderate

=5.31,SD =1.48;B =

.20, t[2562] = 2.64, p , .01, 95% CI [.355, .052], d =.13),

and this lacking sense of productivity signiﬁca ntly mediated the nega-

tive effect of having a high (vs. moderate) amount of discretionary

time on subjective well-being (B = .09, SE = .04, 95% CI [.167,

.024]) based on 10,000 bootstrap samples with no signiﬁcant direct

effect (B = .03, SE =.05,t[2562] = .464, 95% CI [.131, .081]).

Study 3 revealed that a lacking sense of productivity is one

mechanism explaining the negative effect on subjective well-being

from having a large (vs. moderate) amount of discretionary time.

Along with the moderating role of productive (vs. nonproductive)

discretionary activities in Study 2, these results suggest that the

negative effect of having too much discretionary time would be

attenuated when that discretionary time is spent productively. We

tested this in the next study.

Study 4: Productive Versus Nonproductive

Discretionary Time

In Study 4, we focus on the negative effect of having too much

discretionary time for subjective well-being. In particular, based

on our ﬁndings in Studies 2 and 3, we predicted that if people

spent their discretionary time nonproductively, they would report

lower subjective well-being with high (vs. moderate) amounts of

discretionary time. However, if they instead spent their discretion-

ary time on productive activities, the negative effect of having a

high (vs. moderate) amount of discretionary time would be attenu-

ated. Study 4 thus sought to provide additional evidence for the

exploratory analyses in Study 2 using random assignment in a

between-subjects experimental design.

Method

As preregistered, 5,000 American adults

were recruited on

Amazon’s Mechanical Turk to participate in this study. Five thou-

sand and one participants (ages 18–91, M

age

= 38.30; 50.3%

male

) completed this 2 (amount of discretionary time: moderate

vs. high) 3 2 (type of discretionary time: productive vs. nonpro-

ductive) between-subjects experiment.

Participants were asked to mentally simulate having either 3.5 hr

of discretionary time (moderate discretionary time condition) or 7

hr of discretionary time (high discretionary time condition) for at

least 6 months of their life. They were presented with a revised and

more straightforward deﬁnition of discretionary time: “time spent

on activities that are pleasurable or meaningful to you.”

In addi-

tion, participants were instructed to imagine spending this time on

productive [nonproductive] activities: “you consider this use of dis-

cretionary time to NOT be wasted [to be wasted]. This use of time

would feel [would NOT feel] useful, accomplished, fulﬁlling, help-

ful, purposeful, and/or worthwhile.” We speciﬁed that they would

have this amount of time consistently over at least a six-month pe-

riod, and this is how they would spend it.

For each condition, participants were prompted to imagine and

vividly describe what it would be like to have the given amount of

discretionary time (i.e., how they would spend this time, what they

would be doing during the other portion of their days, and how they

would be feeling each day) throughout this period of their life.

Participants then reported their subjective well-being by rating the

extent to which they would experience enjoyment, happiness, and

satisfaction on a 7-point Likert scale, with 1 = would not feel this at

all to7=would feel this a lot (a

moderate, productive

= .90, a

moderate, not

productive

=.95;a

high, productive

= .91, a

high, not productive

= .96).

Last, participants were asked a series of questions capturing fea-

tures of their imagined time period: employment status, the number

of hours worked per day, occupation, marital status, parental status,

whether on vacation, whether in between jobs, and whether they had

experienced any signiﬁcant life changes right before this period of

their life. They were also asked manipulation check questions about

the amount of discretionary time they imagined having (“During the

period of your life that you previously described, how much discre-

tionary time did you have?” on a 11-point scale with 5=very little

discretionary time and5=a lot of discretionary time) and about

how productively they imagined using this time (“To what extent

would you describe this period of your life as the following?” for

“productive,”“purposeful,” and “I accomplished a lot” on 7-point

Likert scales with 1 = would not feel this at all and 7 = would feel

this a lot; a

moderate, productive

= .88, a

moderate, not productive

=.95;a

high,

productive

=.88,a

high, not productive

=.96).

One participant reported imagining working more than 24 hours a day

and was excluded from this analysis.

This sample size was determined by a power analysis based on Study 3.

Because of an experimenter error, we did not collect information on

additional participant demographics.

This deﬁnition was suggested by one of the reviewers during this

paper’s review process.

SHARIF, MOGILNER, AND HERSHFIELD

This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Results

Eight hundred ﬁfteen participants failed the attention check. Of

those who passed the attention check, 140 participants wrote

meaningless/nonsensical responses when describing how they

would spend their discretionary time. As preregistered, we

excluded these participants, leaving a ﬁnal sample of 4,046 partici-

pants (ages 18–91, M

age

= 38.51; 46.3% male). First, conﬁrming

our manipulation check, participants considered their time to be

more productive in the productive conditions than the nonproduc-

tive conditions (M

Productive

= 6.12, SD = .98 vs. M

NotProductive

=3.31, SD = 1.94; t[4044] = 58.21, p , .001, 95% CI [2.89,

2.71]). Participants also perceived they had more discretionary

time in the high versus moderate amount of discretionary time

conditions (M

High

= 9.67, SD = 1.86 vs. M

Moderate

=8.76, SD =

2.01; t[4044] = 14.90, p , .001, 95% CI [1.03, .789]).

We conducted an OLS regression predicting subjective well-

being from a dummy variable representing the amount of discre-

tionary time (high vs. moderate), a dummy variable representing

the type of discretionary time (productive vs. nonproductive), and

a variable representing their interaction. As predicted, we found a

signiﬁcant 2 (amount of discretionary time: high vs. moderate) x 2

(type of discretionary time: productive vs. nonproductive) interac-

tion (B = .514, SE = .089, t[4042] = 5.76, p , .001, 95% CI [.339,

.689]).

17,18

Simple effect analysis revealed that when people spent

their discretionary time nonproductively, people reported lower

subjective well-being when they had a high (7 hr) versus moderate

(3.5 hr) amount of discretionary time (M

High

= 4.90, SD= 1.84 vs.

Moderate

= 5.30, SD = 1.66; B = .400, SE = .064, t[4042] =

6.27, p , .001, 95% CI [.524, .275], d = .23; Figure 4).

However, when people spent their time productively, they reported

marginally higher subjective well-being when they had a high (7

hr) versus moderate (3.5 hr) amount of discretionary time (M

High

6.20, SD = .98 vs. M

Moderate

= 6.08, SD = 1.02; B = .114, SE =

.063, t[4042] = 1.83, p = .068, 95% CI [.008, .237], d = .11; Fig-

ure 4).

This interaction held when controlling for characteristics of

their imagined time period: employment status, number of work-

ing hours,

marital status, parental status, whether on vacation,

whether in-between jobs, and whether there had been any signiﬁ-

cant life change right before, all of which were mean-centered

(B = .482, SE = .088, t[4031] = 5.47, p , .001, 95% CI [.309,

.655]).

Simple effect analysis revealed that when people spent

their discretionary time nonproductively, they reported lower sub-

jective well-being when they had a high (7 hr) versus moderate

(3.5 hr) amount of discretionary time (B = .378, SE = .064,

t[4031] = 5.94, p , .001, 95% CI [.503, .253]). However,

when people spent their discretionary time productively, they

reported a marginally higher subjective well-being when they had

a high (7 hr) versus moderate (3.5 hr) amount of discretionary

time (B = .104, SE = .062, t[4031] = 1.67, p = .095, 95% CI

[.018, .227]).

As in the exploratory analyses in Study 2, Study 4 further dem-

onstrates that the type of discretionary time moderates the effect of

amount of discretionary time on subjective well-being. In particu-

lar, productive uses of discretionary time attenuate the negative

effect of having too much discretionary time.

General Discussion

In two large-scale nationally representative data sets and two

experiments, we investigated the relationship between the

amount of di scretionary time people have in their daily lives and

their subjective well-being. Leveraging larger sample sizes and

more objectiv e assess men t s of tim e afﬂuence, these results sub-

stantiate the previously observed emotional burden of feeling

time-constrained (Hamermesh & Lee, 2007; Kasser & Sheldon,

2009). We fo und that having a dearth o f discre tionary hours in

one’s day indeed results in greater stress and lower subjective

well-being. Although too little time is bad, we further found that

having more time is not a lways better. In fact, t he resu lts fro m

Studies 2–4 suggest it is possible to have too much time. With

prior literatures having separately examined the negative effects

of being too busy (e.g., Kasser & Sheldon, 2009; Roxburgh,

2004) or not being busy enough (e.g., Bellezza et al., 2017 ;

Hsee et al., 2010; Keinan & Kivetz, 2011), ou r research is

among the ﬁrst to bring these perspectives together. By testing

across the full range of daily discretionary hours, our ﬁndings

clarify the suboptimal emotional impact of having either too lit-

tle or too much time.

Because discretionary time is the amount of time one spends on

activities that are pleasurable or intrinsically rewarding (i.e., activ-

ities one wants to do), how is it possible to have too much? In line

with the Aristotelian urging against excess and toward moderation,

as well as mounting evidence for the possibility of having “too

much of [any] good thing” (Grant & Schwartz, 2011), our ﬁndings

indicate that an overabundance of discretionary hours in one’s

days can too be associated with lower subjective well-being. The

attenuating role of whether people’s discretionary time was spent

productively (Studies 2 and 4), along with the reduced feelings of

productivity reported among those who had been allocated a high

amount of discretionary time (Study 3), point to an explanation.

These results suggest that having too much discretionary time

undermines people’s sense of productivity and purpose, thus leav-

ing them less satisﬁed overall.

Although our investigation centered on the relationship

between amount of discretionary time and subjective well-

being, our additional exploration into how individuals spend

their discretionary time proved revealing. In particular, we

Preregistered as an exploratory analysis, examining just the

satisfaction item as the dependent variable similarly showed a signiﬁcant

interaction (B =.510, SE = .096, t[4042] = 5.30, p , .001, 95% CI [.321,

.698]).

Although qualiﬁed by the signiﬁcant interaction, the results also

showed a signiﬁcant negative main effect of having a high (vs. moderate)

amount of discretionary time on subjective well-being (B = .138, SE =

.045, t[4043] = 3.08, p , .001, 95% CI [.226, .050]), as well as a

signiﬁcant positive main effect of productive (vs. nonproductive) use of

that time on subjective well-being (B = 1.03, SE = .045, t(4043) = 23.08,

p , .001, 95% CI [.946, 1.122]).

Four participants reported imagining working more than 24 hours a

day and were thus excluded from this analysis.

Although qualiﬁed by the signiﬁcant interaction, the results also

showed a signiﬁcant negative main effect of having a high (vs. moderate)

amount of discretionary time on subjective well-being (B = .134, SE =

.045, t[4043] = 2.97, p , .001, 95% CI [.222, .045]), as well as a

signiﬁcant positive main effect of productive (vs. nonproductive) use of

that time on subjective well-being (B = .977, SE = .045, t[4043] = 21.89,

p , .001, 95% CI [.889, 1.06]).

DISCRETIONARY TIME AND SUBJECTIVE WELL-BEING

This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

found that if people spent their discretionary time in worthwhile

ways—on productive or social activities—the negative effect of

having too much time was attenuated. Thus, if the particular

way people spend their discretionary t ime is not accounted for

(as in Study 1 and Study 3), the detrimental effect of having an

overabundance may be weak or nonexistent. These ﬁndings sup-

port self-determin atio n theory’s assertion that well-being

requires fulﬁllment of three fundamental psychological needs:

autonomy, relatedness, and competence (e.g., Sheldo n et al.,

2010; Sheldon & Niemiec, 2006 ). With having discretion in

what activities one does likely contributing to a sense of

autonomy, engaging in social activities likely contributing to a

sense of relatedness, and engaging in productive activities likely

contributing to a sense of competence, our research importantly

highlights that the number of daily hours an individual has

available to spend as they choose, as well as how they allocate

those hours, is critical to well-being.

These ﬁndings also have clear practical implications for indi-

viduals—particularly those who experience t ime poverty. For the

many who feel unhappy from the stress of having too much to do

and not enough time to do it, the answer is not to quit all obliga-

tions. Our ﬁndings suggest that ending up with entire days free

to ﬁll at one’s discretion may leave one similarly unhappy. Fig-

ure 3 from Study 2, for instance, suggests that beyond having

two hours of discretionary time in the day, having more time

does not promise greater happiness. That ﬁgure a lso suggests

that beyond approximately ﬁve hours of discretionary time in the

day, having more is linked to less happiness. Thus, the time poor

should not quit everything and neglect all of their obligations;

instead, they should strive for the reasonably attainable amount

of having a little over two hours to spend how they want during

their days. To be clear though, these amounts are inexact and

based merely on eyeballing a graph, which represents one dataset

and buckets together many types of individuals according to the

amount of time each spent on discretionary activities in one

given day. However, t he overall inverted U-shaped pattern is ro-

bust across people and does offer useful general guidance.

Namely, the signiﬁcant negative quadratic relationship identiﬁed

across multiple large data sets representing the diversity of

American adults, as well as the nonsigniﬁcant interaction effects

comparing segments of the population, preliminarily suggest that

—whether young or old, working or unemployed, male or

female, married or single, with chi ldren or without— most would

beneﬁt from having a moderate amount of discretionary time:

not too little and not too much.

In cases when individuals do ﬁnd themselves in circumstances

with excessive amounts of discretionary time (e.g., upon retire-

ment or having left a job), our results suggest these individuals

would beneﬁt from spending their newfound time with purpose (e.

g., productively or connecting with others). These ﬁndings are

consistent with prior research conducted among retirees (a popula-

tion with ample discretionary time), showing that retirees’ well-

being depends on their level of social interaction (Longino & Kart,

1982) and participation in clubs or volunteer work (Kelly & Ross,

1989; Mishra, 1992; Morrow-Howell, 2010).

Limitations and Directions for Future Research

Although robust, it is worth noting that the observed effect of

discretionary time on subjective well-being is small. However, we

would not expect the quantity of one’s hours spent on discretion-

ary activities in a day to have any larger of an effect on subjective

well-being than we observed. This is because, in addition to other

time-related variables that inﬂuence subjective well-being (e.g.,

how people spend their hours, Kahneman et al., 2004 ; their mind-

set during those hours, Killingsworth & Gilbert, 2010; West et al.,

Figure 4

Means of Subjective Well-Being by Condition in Study 4

3.5

4.5

5.5

6.5

Nonproductive Productive

Subjective Well-Being

Moderate High

Note. Error bars represent standard error of the mean.

SHARIF, MOGILNER, AND HERSHFIELD

This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

2021; and how their obligatory time is ﬁlled, Judge & Shinichiro,

1993); there are a slew of other variables that play into people’s

overall assessment of their satisfaction in life (Lyubomirsky et al.,

2005). Furthermore, despite it being small, we are conﬁdent in its

existence and shape. Our predicted negative quadratic effect repli-

cated across data sets and subsets of the data sets, and it persisted

with and without the inclusion of relevant covariates. The negative

effect of having too much discretionary time is particularly small,

which we found is partly explained by its sensitivity to the particu-

lar ways that time is spent. Albeit small, together these ﬁndings

provide preliminary theoretically and practically important

insights about the role of discretionary time allocation and usage

for well-being.

As an initial exploration into the relationship between time

afﬂuence and subjective well-being, our hope is that this work

spurs future investigations into more precise research questions.

For instance, the current ﬁndings examine how the amount of dis-

cretionary time a person has on a typical day relates to well-being,

but it does not inform the experience of atypical days, like when

on vacation or on a holiday. Furthermore, although we have pro-

vided initial insight into possible mechanisms (i.e., stress for too

little time and lack of productivity for too much time), our identi-

ﬁed effect is likely multiply determined. For example, boredom

may also contribute to reduced well-being from having an over-

abundance of discretionary time. Thus, it is possible that factors

that have been previously identiﬁed as effective in offsetting

hedonic adaptation (e.g., variety across activities; Etkin & Mogil-

ner, 2016; or taking breaks between enjoyable experiences; Quoid-

bach & Dunn, 2013) may be similarly effective in informing

people how to optimally allocate their discretionary time across a

day or week.

Additionally, building off of our proposed mechanism, we

tested a few ways in which our effect might be attenuated (i.e.,

spending discretionary time on worthwhile endeavors, such as

activities that are productive or socially connecting). However,

there are a number of additional ways to characterize the way peo-

ple spend their discretionary time that might also inﬂuence our

observed relationship. We hope future research investigates the

psychological and behavioral nuances within the category of dis-

cretionary activities.

Future work should also experimentally examine the precise point

at which an abundance of discretionary time starts to negatively

affect subjective well-being. The correlational data in Figure 3 sug-

gest that people are happiest having between two and ﬁve ho ur s

of di sc re tion ar y time, and that the s lop e is neg at ive beyond that

point. The results from our experimental studies similarly show

that people with 7 hr of discretionary time experience less sub-

jective well-being than people with 3.5 hr of discretionary t ime.

However, futu re wo rk should experiment ally manipulate at a

more granular level the amount of discretionary time partici-

pants have to more precisely identify the optimal amount of dis-

cretionary time.

Relatedly, the results of Study 2 suggest that people can beneﬁt

from a greater amount of discretionary time before it negatively

affects their subjective well-being as long as the time is spent pro-

ductively or with others. Thus, future work should also identify

optimal amounts of discretionary time depending on how that time

is used, as well as how to optimally allocate given amounts of dis-

cretionary time between relaxation, socializing, and productive

use. Our data cannot speak to, for instance, whether a small

amount of discretionary time is better spent relaxing and unwind-

ing in front of the TV, or whether that limited spare time should

be diligently spent going for a run or knitting a sweater.

It is also possible that the relationship between amount of dis-

cretionary time and subjective well-being varies cross-culturally.

For example, although busyness is associated with higher social

status in the United States ( Bellezza et al., 2017), this may not be

the case in other countries that place greater value on relaxation

(Brislin & Kim, 2003). Without the same compulsion toward pro-

ductivity (Mogilner & Norton, 2019), the negative effect of having

too much discretionary time that we identiﬁed in our American

samples may not extend more broadly.

People often complain about being too busy and express want-

ing more time (Trupia et al., 2021). However, Studies 3 and 4 sug-

gest that when people mentally simulate the experience of having

an abundance of discretionary time, they are able to forecast its

detrimental effects when spent unproductively. Although prior

research has validated the veracity of mental simulation in captur-

ing actual experience (Morewedge et al., 2010; O’Brien et al.,

2018), future research should examine when people are able to

correctly forecast that an abundance of discretionary time would

have a negative impact on their well-being. For example, when

people are simply asked (without being encouraged to mentally

simulate the experience), do they think that having more discre-

tionary time would always make them happier?

Other open questions involve the role of choice and whether

there is an end point to the period of time in which people have an

ample amount of discretionary time during their days. For

instance, would we observe the same pattern of results among

those on a sabbatical (for whom it was a choice, and there is an

endpoint) and for those who decided to retire early (for whom it

was a choice, but there is no endpoint)? Even though the results

from Study 2 persisted for those in and out of the workforce, and

even though the results from Studies 3 and 4 persisted when con-

trolling for whether participants were imagining being employed

or unemployed, it may be the case that if one’s ample amount of

discretionary time is imposed (e.g., being laid off), the relationship

between too much discretionary time and subjective well-being

may be even more pronounced.

Lastly, although our pattern of results persisted across multiple

large-scale correlational data sets and two experiments, future ex-

perimental work is needed to test these ﬁndings outside of a hypo-

thetical paradigm. Our causal evidence for how the amount of

discretionary time people have inﬂuences their subjective well-

being relied on hypothetical mental simulations (Studies 3 and 4),

and

our manipulation for unproductive versus productive uses of

time (Study 4) was rather heavy-handed. Although resources lim-

ited our ability to actually assign very small or very large amounts

of discretionary time in people’s actual daily lives, we hope that

future experimentalists ﬁnd a way to test and conﬁrm these ﬁnd-

ings among participants who actually live through the varied

conditions.

Conclusion

Despite our observed effect being small, the impact of time

afﬂuence (or scarcity) for subjective well-being is important.

These ﬁndings inform the ongoing investigation into the role of

DISCRETIONARY TIME AND SUBJECTIVE WELL-BEING

This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

afﬂuence as a signiﬁcant predictor of subjective well-being—but

extend the focus from money (Kahneman & Deaton, 2010; Kill-

ingsworth, 2021) to include humans’ other principle resource:

time. This work thus contributes to the burgeoning literature on

time and subjective well-being (Mogilner et al., 2018) and adds a

qualiﬁcation to research that has identiﬁed particularly happy

ways to spend time (Kahneman et al., 2004): solely ﬁlling one’s

days with those activities may undermine feelings of purpose and

thus reduce satisfaction in life.

References

Banwell, C., Hinde, S., Dixon, J., & Sibthorpe, B. (2005). Reflections on

expert consensus: A case study of the social trends contributing to obe-

sity. European Journal of Public Health, 15(6), 564–568. https://doi

.org/10.1093/eurpub/cki034

Bellezza, S., Paharia, N., & Keinan, A. (2017). Conspicuous consumption

of time: When busyness and lack of leisure time become a status sym-

bol. The Journal of Consumer Research, 44(1), 118–138.

Brislin, R. W., & Kim, E. S. (2003). Cultural diversity in people’s under-

standing and uses of time. Applied Psychology, 52(3), 363–382. https://

doi.org/10.1111/1464-0597.00140

Carroll, J. (2008). Time pressures, stress common for Americans. Gallup

Poll Social Series. http://www.gallup.com/poll/103456/Time-Pressures

-stress-Common-Americans.aspx

Csikszentmihalyi, M. (2000). The costs and benefits of consuming The

Journal of Consumer Research, 27(2), 267–272. https://doi.org/10.1086/

314324

Csikszentmihalyi, M., & Hunter, J. (2003). Happiness in everyday life:

The uses of experience sampling. Journal of Happiness Studies, 4(2),

185–199. https://doi.org/10.1023/A:1024409732742

Diener, E., Heintzelman, S., Kushlev, K., Tay, L., Wirtz, D., Lutes, L., &

Oishi, S. (2017). Findings all psychologists should know from the new

science on subjective well-being. Canadian Psychology, 58(2), 87–104.

https://doi.org/10.1037/cap0000063

Diener, E., & Seligman, M. E. (2002). Very happy people. Psychological

Science, 13(1), 81–84. https://doi.org/10.1111/1467-9280.00415

Eriksson, L., Rice, J. M., & Goodin, R. E. (2007). Temporal aspects of life

satisfaction. Social Indicators Research, 80(3), 511–533. https://doi.org/

10.1007/s11205-006-0005-z

Etkin, J., & Mogilner, C. (2016). Does variety among activities increase

happiness? The Journal of Consumer Research, 43(2), 210–229. https://

doi.org/10.1093/jcr/ucw021

Frederick, S., & Loewenstein, G. (1999). Hedonic Adaptation. In D.

Kahneman, E. Diener, & N. Schwarz (Eds.), Well-being: The founda-

tions of hedonic psychology (pp. 302–329).

Russell Sage Foundation.

Gable, S. L., & Bromberg, C. (2018). Healthy social bonds: A necessary

condition for well- being. Handbook of well-being. DEF Publishers.

Galinsky, E., Bond, J. T., & Friedman, D. E. (1993). The changing work-

force: Highlights of the national study. DIANE Publishing.

Grant, A. M., & Schwartz, B. (2011). Too much of a good thing: The chal-

lenge and opportunity of the inverted U. Perspectives on Psychological

Science, 6(1), 61–76. https://doi.org/10.1177/1745691610393523

Hamermesh, D. S., & Lee, J. (2007). Stressed out on four continents: Time

crunch or yuppie kvetch? The Review of Economics and Statistics,

89(2), 374–383. https://doi.org/10.1162/rest.89.2.374

Hershfield, H. E., Mogilner, C., & Barnea, U. (2016). People who choose

time over money are happier. Social Psychological & Personality Sci-

ence, 7(7), 697–706. https://doi.org/10.1177/1948550616649239

Holbrook, M. B., & Lehmann, D. R. (1981). Allocating discretionary time:

Complementarity among activities. The Journal of Consumer Research,

7(4), 395–406. https://doi.org/10.1086/208830

Hsee, C. K., Yang, A. X., & Wang, L. (2010). Idleness aversion and the

need for justifiable busyness. Psychological Science, 21(7), 926–930.

https://doi.org/10.1177/0956797610374738

Judge, T. A., & Shinichiro, W. (1993). Another look at the job satisfac-

tion-life satisfaction relationship. Journal of Applied Psychology, 78(6),

939–948. https://doi.org/10.1037/0021-9010.78.6.939

Kahneman, D., & Deaton, A. (2010). High income improves evaluation of

life but not emotional well-being. Proceedings of the National Academy

of Sciences of the United States of America, 107(38), 16489–16493.

https://doi.org/10.1073/pnas.1011492107

Kahneman, D., Krueger, A. B., Schkade, D. A., Schwarz, N., & Stone,

A. A. (2004). A survey method for characterizing daily life experience:

The day reconstruction method. Science, 306(5702), 1776–1780. https://

doi.org/10.1126/science.1103572

Kasser, T., & Sheldon, K. M. (2009). Time affluence as a path toward per-

sonal happiness and ethical business practice: Empirical evidence from

four studies. Journal of Business Ethics, 84(S2), 243–255. https://doi

.org/10.1007/s10551-008-9696-1

Keinan, A., & Kivetz, R. (2011). Productivity orientation and the con-

sumption of collectable experiences. The

Journal of Consumer

Research, 37(6), 935–950. https://doi.org/10.1086/657163

Kelly, J. R., & Ross, J. E. (1989). Later-life leisure: Beginning a new

agenda. Leisure Sciences, 11(1), 47–59. https://doi.org/10.1080/

01490408909512204

Killingsworth, M. A. (2021). Experienced well-being rises with income,

even above $75,000 per year. Proceedings of the National Academy of

Sciences of the United States of America, 118(4), e2016976118. https://

doi.org/10.1073/pnas.2016976118

Killingsworth, M. A., & Gilbert, D. T. (2010). A wandering mind is an

unhappy mind. Science, 330(6006), 932. https://doi.org/10.1126/science

.1192439

Kurtz, J. L. (2008). Looking to the future to appreciate the present: The

benefits of perceived temporal scarcity. Psychological Science, 19(12),

1238–1241. https://doi.org/10.1111/j.1467-9280.2008.02231.x

Kushlev, K., Heintzelman, S. J., Oishi, S., & Diener, E. (2018). The declin-

ing marginal utility of social time for subjective well-being. Journal of

Research in Personality, 74, 124–140. https://doi.org/10.1016/j.jrp.2018

.04.004

Kuykendall, L., Boemerman, L., & Zhu, Z. (2018). The importance of lei-

sure for subjective well-being. In E. Diener, S. Oishi, & L. Tay (Eds.),

Handbook of well-being. DEF Publishers.

Lathia, N., Sandstrom, G. M., Mascolo, C., & Rentfrow, P. J. (2017). Hap-

pier people live more active lives: Using smartphones to link happiness

and physical activity. PLoS ONE, 12(1), e0160589. https://doi.org/10

.1371/journal.pone.0160589

Longino, C. F., Jr., & Kart, C. S. (1982). Explicating activity theory: A for-

mal replication. Journal of Gerontology, 37(6), 713–722. https://doi.org/

10.1093/geronj/37.6.713

Lyubomirsky, S., Sheldon, K. M., & Schkade, D. (2005). Pursuing happi-

ness: The architecture of sustainable change. Review of General Psy-

chology, 9(2), 111–131. https://doi.org/10.1037/1089-2680.9.2.111

Masuda, Y. J., Williams, J. R., & Tallis, H. (2020). Does life satisfaction

vary with time and income? Investigating the relationship among free

time, income, and life satisfaction. Journal of Happiness Studies.

Advance online publication. https://doi.org/10.1007/s10902-020-00307-8

Mishra, S. (1992). Leisure activities and life satisfaction in old age: A case

study of retired government employees living in urban areas. Activities, Ad-

aptation and Aging, 16(4

), 7–26. https://doi.o rg/10.1300 /J016v16 n04_02

Mogilner, C. (2010). The pursuit of happiness: Time, money, and social

connection. Psychological Science, 21(9), 1348–1354. https://doi.org/10

.1177/0956797610380696

Mogilner, C., & Norton, M. (2019). Preferences for experienced versus

remembered happiness. The Journal of Positive Psychology, 14(2),

244–251. https://doi.org/10.1080/17439760.2018.1460688

SHARIF, MOGILNER, AND HERSHFIELD

This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Mogilner, C., Whillans, A., & Norton, M. I. (2018). Time, money, and sub-

jective well- being. Handbook of Well-Being. Noba Scholar Handbook

series: Subjective well-being. DEF Publishers.

Morewedge, C. K., Huh, Y. E., & Vosgerau, J. (2010). Thought for food:

Imagined consumption reduces actual consumption. Science, 330(6010),

1530–1533. https://doi.org/10.1126/science.1195701

Morrow-Howell, N. (2010). Volunteering in later life: Research frontiers.

The Journals of Gerontology: Series B, 65(4), 461–469. https://doi.org/

10.1093/geronb/gbq024

Newport, F. (2016). The Gallup poll: Public opinion 2015. Rowman & Lit-

tlefield Publishers.

O’Brien, E., Kristal, A. C., Ellsworth, P. C., & Schwarz, N. (2018). (Mis)

imagining the good life and the bad life: Envy and pity as a function of

the focusing illusion. Journal of Experimental Social Psychology, 75,

41–53. https://doi.org/10.1016/j.jesp.2017.10.002

Quoidbach, J., & Dunn, E. W. (2013). Give it up: A strategy for combating

hedonic adaptation. Social Psychological & Personality Science, 4(5),

563–568. https://doi.org/10.1177/1948550612473489

Quoidbach, J., Dunn, E. W., Hansenne, M., & Bustin, G. (2015). The price

of abundance: How a wealth of experiences impoverishes savoring. Per-

sonality and Social Psychology Bulletin, 41(3), 393–404. https://doi.org/

10.1177/0146167214566189

Roxburgh, S. (2004). ‘There just aren’t enough hours in the day’:Themental

health consequences of time pressure. Journal of Health and Social Behav-

ior, 45(2), 115–131. https://doi.org/10.1177/002214650 40450020 1

Sheldon, K. M., Cummins, R., & Kamble, S. (2010). Life balance and

well-being: Testing a novel conceptual and measurement approach.

Journal of Personality, 78(4), 1093–1134. https://doi.org/10.1111/j

.1467-6494.2010.00644.x

Sheldon, K. M., & Niemiec, C. P. (2006). It’s not just the amount that

counts: Balanced need satisfaction also affects well-being. Journal of

Personality and Social Psychology, 91(2), 331–341. https://doi.org/10

.1037/0022-3514.91.2.331

Simonsohn, U. (2018). Two lines: A valid alternative to the invalid testing

U-shaped relationships with quadratic regressions. Advances in Meth-

ods and Practices in Psychological Science, 1(4), 538–555. https://doi

.org/10.1177/2515245918805755

Snir, R., & Zohar, D. (2008). Workaholism as discretionary time invest-

ment at work: An experience-sampling study. Applied Psychology,

57(1), 109–127. https://doi.org/10.1111/j.1464-0597.2006.00270.x

Strazdins, L., Griffin, A. L., Broom, D. H., Banwell, C., Korda, R., Dixon,

J., Paolucci, F., & Glover, J. (2011). Time scarcity: Another health in-

equality? Environment & Planning A, 43(3), 545–559. https://doi.org/10

.1068/a4360

Teuchmann, K., Totterdell, P., & Parker, S. K. (1999). Rushed, unhappy,

and drained: An experience sampling study of relations between time

pressure, perceived control, mood, and emotional exhaustion in a group

of accountants. Journal of Occupational Health Psychology, 4(1),

37–54. https://doi.org/10.1037/1076-8998.4.1.37

Trupia, M., Mogilner, C., & Engeler, I. (2021). Stressed or self-important?

Intentions versus perceptions when communicating busyness. Working

paper.

West, C., Mogilner, C., & DeVoe, S. (2021). Happiness from treating the

weekend like a vacation. Social Psychological & Personality Science,

12(3), 346–356. https://doi.org/10.1177/1948550620916080

Whillans, A. V., Dunn, E. W., Smeets, P., Bekkers, R., & Norton, M. I.

(2017). Buying time promotes happiness. Proceedings of the National

Academy of Sciences of the United States of America, 114(32),

8523–8527. https://doi.org/10.1073/pnas.1706541114

Whillans, A. V., Weidman , A. C., & Dunn, E. W. (2016) . Valuing time

over money is ass ociated with g reater h appines s. Social Psychologi-

cal & Personality Science, 7(3), 213–222. https ://doi.org/ 10.1177/

1948550615623842

Wilson, T. D., & Gilbert, D. T. (2008). Explaining away: A model of affec-

tive adaptation. Perspectives on Psychological Science, 3(5), 370–386.

https://doi.org/10.1111/j.1745-6924.2008.00085.x

Young, C., & Lim, C. (2014). Time as a network good: Evidence from

unemployment and the standard workweek. Sociological Science, 1,

10–27. https://doi.org/10.15195/v1.a2

Received November 16, 2019

Revision received May 4, 2021

Accepted May 11, 2021

DISCRETIONARY TIME AND SUBJECTIVE WELL-BEING

This document is copyrighted by the American Psychological Association or one of its allied publishers.

This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.