Content uploaded by Yung-Ju Chang
Author content
All content in this area was uploaded by Yung-Ju Chang on May 03, 2023
Content may be subject to copyright.
Not Merely Deemed as Distraction: Investigating Smartphone
Users’ Motivations for Notification-Interaction
Xi-Jing, Chang Fang-Hsin, Hsu En-Chi Liang
siliconcrystal.c@nycu.edu.tw vivi160638.cs09@nycu.edu.tw n7enchi.ee08@nycu.edu.tw
National Yang Ming Chiao Tung National Yang Ming Chiao Tung National Yang Ming Chiao Tung
University University University
Hsinchu, Taiwan, Taiwan Hsinchu, Taiwan, Taiwan Hsinchu, Taiwan, Taiwan
Zih-Yun Chiou Ho-Hsuan Chuang Fang-Ching Tseng
dobelle.11@nycu.edu.tw stonechuang.ee07@nycu.edu.tw r11922012@ntu.edu.tw
Institute of Articial Intelligence National Yang Ming Chiao Tung National Taiwan University
Innovation, National Yang Ming University Taipei, Taiwan, Taiwan
Chiao Tung University Hsinchu, Taiwan, Taiwan
Hsinchu, Taiwan, Taiwan
Yu-Hsin Lin
yuhsin.lin@outlook.com
Department of Computer Science,
National Yang Ming Chiao Tung
University
Hsinchu, Taiwan, Taiwan
ABSTRACT
Notications are commonly considered a distraction when they
arrive during a task, and consequently, prior research has consis-
tently sought eective ways of deferring their arrival until task
transitions. However, many smartphone users still interact with
notications during tasks. In our qualitative study combining diary
study and semi-structured interviews, we examined 34 research par-
ticipants’ motivations for interacting with smartphone notications
at dierent times, including during tasks. Our ndings resulted in a
human-notication interaction framework comprised of 12 unique
motivations frequently associated with three activity timings for
interacting with notications, including before-task, during-task,
and after-task. Notably, participants frequently perceived interac-
tion with notications as a tool for improving task performance,
making the most of their time, and promoting personal well-being,
rather than only as a distraction. The before-the-task timing, in
particular, has received little attention in previous research and de-
serves more attention as it was related to specic user motivations
for notication interaction.
∗Corresponding author.
Permission to make digital or hard copies of all or part of this work for personal or
classroom use is granted without fee provided that copies are not made or distributed
for prot or commercial advantage and that copies bear this notice and the full citation
on the rst page. Copyrights for components of this work owned by others than the
author(s) must be honored. Abstracting with credit is permitted. To copy otherwise, or
republish, to post on servers or to redistribute to lists, requires prior specic permission
and/or a fee. Request permissions from permissions@acm.org.
CHI ’23, April 23–28, 2023, Hamburg, Germany
© 2023 Copyright held by the owner/author(s). Publication rights licensed to ACM.
ACM ISBN 978-1-4503-9421-5/23/04.. . $15.00
https://doi.org/10.1145/3544548.3581146
Yung-Ju, Chang∗
armuro@nycu.edu.tw
National Yang Ming Chiao Tung
University
Hsinchu, Taiwan, Taiwan
CCS CONCEPTS
• Human-centered computing → Empirical studies in HCI.
KEYWORDS
Attention Management; Multitasking; Interruptions, Notications;
Opportune Moment
ACM Reference Format:
Xi-Jing, Chang, Fang-Hsin, Hsu, En-Chi Liang, Zih-Yun Chiou, Ho-Hsuan
Chuang, Fang-Ching Tseng, Yu-Hsin Lin, and Yung-Ju, Chang. 2023. Not
Merely Deemed as Distraction: Investigating Smartphone Users’ Motiva-
tions for Notication-Interaction . In Proceedings of the 2023 CHI Conference
on Human Factors in Computing Systems (CHI ’23), April 23–28, 2023, Ham-
burg, Germany. ACM, New York, NY, USA, 17 pages. https://doi.org/10.1145/
3544548.3581146
1 INTRODUCTION
The growing popularity of smartphone use in recent years is not
only due to the increasing availability and range of mobile Internet,
but also to the variety of mobile applications that allow users to
access a wide range of digital services [
10
,
13
]. However, sharp in-
creases in the average number of such applications installed on each
phone have inescapably resulted in people receiving an increasing
number of notications, which has prompted debates over whether
notications have become distraction or even damaging to smart-
phone users [
14
,
15
,
23
,
56
,
108
]. As a result, a strand of research
has attempted to identify moments for deferring notications to, in-
cluding opportune moments [
35
,
89
], breakpoints [
2
,
45
,
51
,
81
,
85
],
and task transitions [
45
,
53
], for the purpose of reducing the distrac-
tion or disturbance they cause. Research has also shown, however,
that smartphone users have a propensity to initiate notication-
checking themselves [
19
], and that this behavior can occur even
while a person is in the middle of a task [
5
,
27
,
33
,
36
,
105
]: a
CHI ’23, April 23–28, 2023, Hamburg, Germany Chang and Wu, et al.
phenomenon that has been widely presumed to harm task per-
formance. Yet, if interacting with notications during a task is
harmful, why, and in what situations, would smartphone users
want to do so? Answering this question will shed light on smart-
phone users’ perceptions of notications’ benets, as well as the
situations in which such benets are more or less evident to them:
factors that have been neglected in the literature on notications
and interruptibility, presumably due to the dominant notion of
notication-as-interruption [
14
,
15
,
23
,
37
,
56
,
108
]. Accordingly,
we ask:
RQ 1: What are smartphone users’ motivations for interacting with
notications when they are performing another, primary task?
RQ 2: What factors trigger or inuence such motivations?
We believe that the answers to these two questions will facil-
itate improvement of notication systems, by shedding light on
the types of notications users want to receive at dierent specic
timings, including in the middle of their primary tasks. To answer
them, we conducted a qualitative study that involved a week-long
diary study and semi-structured interviews with 34 participants,
with both these approaches focused on obtaining in-depth under-
standing of these individuals’ motivations for reading and handling
notications at dierent timings in their daily lives. It makes the
following four substantial contributions to the HCI literature.
•
It identies 12 distinct motivations for interacting with noti-
cations and reveals that participants often associated these
motivations with one of three specic points in time relative
to the primary task - before, during, or after the task.
•
It reveals that notications were often perceived by partic-
ipants as more than merely distractions, but instead, as a
means to improve task performance, time utilization, and
personal well-being.
•
It distinguishes between before-task and after-task transition
timings from the users’ perspective, revealing that partici-
pants more frequently associated a specic set of motivations
in the former vs. the latter. This has important implications
for future research, insofar as more attention should be paid
to the before-the-task transition timing, which has long been
neglected in the literature.
•
It proposes a human-notication interaction framework that
includes, in addition to the aforementioned 12 motivations
grouped by task timings, contextual factors that triggered
and inuenced these motivations, and can be used as a theo-
retical tool for analyzing the various functions and meanings
of the notications smartphone users wish to receive in a
variety of task situations.
2 RELATED WORK
Notications and opportune moments have been of interest to re-
search communities. We rst review HCI research at the opportune
moments. We then discuss the literature related to the phenomenon
of self-interruption and media multitasking. Lastly, we review the
literature on digital well-being.
2.1 Perceptions of and Interaction with
Notications
Notications have long been regarded as interruptions, disturbance,
and/or distraction in the literature [
14
,
15
,
23
,
37
,
56
,
108
], not only
because numerous research has shown that ill-timed notications
could result in such negative impacts on workers’ performance of
the tasks at hand [
2
,
14
,
59
,
62
,
109
] and delay their task completion
time [
15
,
108
], but also because prior research has also shown that
users perceive alerts of notications as disrupting [
19
] and distract-
ing them from the task at hand [
15
,
108
]. An early study showed
that compared to receiving notications without any alerts, receiv-
ing notications with ringtone alerts is likely result in symptoms
of inattention and hyperactivity [59].
However, in addition to the negative impacts that arise due to the
notications’ presence or arrival, being unaware of their presence
or arrival is also likely to cause negative emotions. Pielot et al [
92
]
has shown that disabling notications is likely to provoke users’
anxiety and worry about missing notications. Probably because of
this reason, Chang et al. [
19
] showed that smartphone users who
adopted a silent ringer mode would proactively check their message
notications when their phones, of which the overall attentiveness
to notications is not less than that when the users’ phones deliver
sensible notication alerts. Smartphone users’ high attentiveness
to notications has also been reported in [
89
], which shows that
they are attentive to their phones on average every ve minutes.
Prior research has provided some explanations for modern smart-
phone users’ high attentiveness to and habitual use of their phones.
For example, it is suggested that some individuals who have greater
needs for being synchronized with their social contacts and the
world, are likely to nd it more challenging to turn o notications,
especially for those who tend to have the feeling of fear of missing
out (FOMO) [
37
,
41
] and greater need to belong (NtB) [
9
]. For ex-
ample, Liao et al. [
64
] shows that for individuals who scored high
NtB and FOMO, disabling notications on the phone might make
them more worried about missing important or urgent information.
For them, notications oers a practical function to them [
118
];
and as a result, turning notications o may conversely distract
them more. As a consequence, they may self-interrupt themselves
and check their phone often to resolve their such feelings [
29
,
64
].
Probably due to this reason, research has shown that these individ-
uals are more likely to display problematic and addictive mobile
phone usage [61, 97, 106] .
On the other hand, despite the large body of research on the neg-
ative impacts of notications, a few others argued that notications
could be benecial in certain contexts and being highly receptive
regardless of the task at hand is important to their work. For exam-
ple, Kim et al. [
86
] showed that information workers whose works
heavily rely on communication appreciated notications that are
relevant and benecial to their task [
16
,
58
,
86
]. Mazmanian et
al. [
70
] also argued that professional service providers (e.g. invest-
ment bankers, corporate lawyers, management consultants) desire
to be highly available for notications because their availability is
an "integral aspect of the product" they should oer. Likewise, Chou
et al. [
25
] also suggested that certain occupations such as sales or
representatives deem it important to be available and responsive
Motivations for Notification-Interaction CHI ’23, April 23–28, 2023, Hamburg, Germany
for communication notications because it is part of their job and
duty.
In addition to users’ experiences with and perceptions of noti-
cations, a large body of research investigates how smartphone
users attend to arriving notications and manage existing ones in
a notication drawer. Overall, although smartphone users have
been shown to be highly attentive to their phones in prior re-
search [
19
,
89
], as mentioned earlier, other studies suggested that
smartphone users’ attendance to notications is selective. For ex-
ample, prior research consistently shows that smartphone users
favor notications related to communication more than the other
types [
11
,
66
,
76
,
90
,
100
,
115
]; they attend to urgent, important,
useful, and attractive ones quicker than otherwise [
66
,
76
,
93
,
100
,
115
]; and they also attend to notications from specic senders
quicker [
60
]. Often, the decisions of whether to attend to the certain
notications have been made when they sensed the notications,
which is according to their speculation about the source and "about-
ness" of the notications [
18
]. For notications that are already
in their notication drawers, researchers have also investigated
the users’ dismissing [
115
], ’snoozing’ [
6
,
85
,
114
], and deferral
behavior [
19
,
59
]. For example, Turner et al [
111
] found their par-
ticipants generally deal with their notications from top to bottom,
but this strategy can be aected by how many notications have
resided in the drawer. Arguing that users may not always want
to follow the top-bottom order to deal with notications, Lin et
al [
65
] shows that their participants sometimes desired to place
unimportant notications on the top because they served specic
functions to them, such as a reminder or a portal; based on their
results, they argue that nowadays notications have served dier-
ent roles and it is important to recognize their various roles when
designing notication systems.
The current paper resonates with the tendency of self-initiated
phone checking behavior, and adds to the literature that there are
four motivations that may be prominent to the user for doing so if
it happens during a primary task. In addition, it resonates with the
multiple functions of notications to users and further adds that
people’s utilization for these functions of notications is inuenced
by their situation, such as relative to a primary task, before it,
during, or after it.
2.2 Interruptions and Receptivity
Interruptions have received denitions in dierent domains, and
in the context of a person attempting to perform a primary task
in HCI, Mcfarlane [
71
,
74
] dened human interruption as “the
process of coordinating abrupt changes in people’s activities,” which
was believed to be a general and interdisciplinary denition. It
was assumed that four elements are involved, including “(P) the
people involved in the interruption; (T) the task(s) the person is
attempting; (In) the interruption itself; and (C) the working context
or environment” [
71
, p. 4]. Similarly, in the context of multitasking
in HCI, interruptions are referred to as unanticipated requests for
switching between dierent tasks during multitasking [
71
,
96
].
Given the notion that the presence of an interruption is established
when there is a primary task being interrupted, interruptions have
long been regarded as harmful. This is expected, as research has
consistently shown that interruptions may be costly to people’s
primary tasks and decrease their task performance [
8
,
42
,
47
,
72
].
For example, interruptions may draw people’s attention and distract
them from the primary task [
4
,
40
]; and it is not always easy to
resume the original task [
43
,
50
,
62
,
69
]. In addition to cognitive
and performance costs, it also can lead to emotional ones, such as
increased stress [67] and pressure to respond quickly [7, 26, 67].
As mentioned earlier, notications have long been considered as
interruptions, especially when they arrive during a primary task. To
reduce the interruptions and perceived disturbance caused by them,
a large body of research has sought to study users’ high-receptive
moments for delivering interruptions/notications, including fac-
tors that aect their receptivity to notications, and the use of
machine learning techniques to detect such moments for delivering
notications. In the former line of research, a number of studies
have identied factors that aect users’ receptivity to notications.
Factors that have been identied include but are not limited to: the
characteristics of the interrupted task/situation [
24
,
34
,
52
,
76
], the
characteristics of the interruption [
28
,
35
,
44
,
52
,
60
,
79
,
100
,
116
],
the level of engagement the interrupted people have in the inter-
rupted task [
38
,
88
], and other personal context [
78
], social context
[85], and environmental factors [32, 76].
In the second line of research, many researchers have developed
machine-learning models to predict moments at which users would
be receptive to notications (e.g. [
21
,
75
,
83
,
84
,
87
,
89
,
95
,
107
,
112
]). Several similar concepts that describe users’ such moments
have been developed and become targets of prediction, including
interruptible moments [
39
,
46
,
49
,
50
,
57
,
113
], breakpoints [
2
,
80
–
82
,
85
], transitions [
98
], opportune moments [
17
,
21
,
22
,
31
,
89
,
95
], bored moment [
91
] and so on. By delivering notications at
such detected moments, Park et al. [
85
], for example, showed a
54.1% decrease in perceptions that notications were interruptive.
Likewise, Okoshi et al. [
81
] showed 71.8% greater reduction of users’
perception of workload when sending notications during detected
breakpoints.
On the other hand, despite the negative inuences of notica-
tions and/or interruptions, research regarding the multitasking
practices and interruptions in organizations has also shown that
they may provide some benets in some situations. For example,
Hudson et al. [
48
] showed that although interruptions may delay
the task at hand, they allows people to rstly deal with problems
that may have become overwhelming later. In addition, interrup-
tions themselves may carry useful information that assists other
tasks/activities. Carroll et al. [
16
], likewise, showed that for indi-
viduals who collaborate in an activity with shared goals, resources,
deadlines, to-do items, social roles, and work practices, interrup-
tions that carry information about updates and changes can help
these individuals react to those changes and keep track of and
integrate them in their own work. Other studies investigate self-
interruptions, causes, and their benets. For example, Dabbish et
al.[
29
] found their study participants sometimes self-interrupted
themselves in order to switch to work for which they are account-
able. In another work, Jin and Dabbish [
55
] identied seven cat-
egories of self-interruptions on the computers that occurred to
their study participants, including: adjustment, break, inquiry, rec-
ollection, routine, trigger, and wait. For example, it was found that
taking breaks could help the participants refresh their minds so
that they could concentration when they returned to the primary
CHI ’23, April 23–28, 2023, Hamburg, Germany Chang and Wu, et al.
task. Inquiries were also found to positively impact the participants’
primary tasks because these help them nd the information they
needed. Taking small breaks during work has also been suggested
as positive to increase focus and creativity at work [1].
3 METHODOLOGY
3.1 Recruitment and Participants
Our initial pool of eleven participants was recruited via word of
mouth and advertisements posted on social-networking sites in
our country. Each advertisement was linked to a screening ques-
tionnaire that asked respondents to self-report their demographic
information, work schedules, lifestyle, and multitasking propensity,
along with some questions relating to notications reading behavior,
including the devices they typically used for reading/handling noti-
cations and their familiarity of technology. Multitasking propensity
was measured using the polychronicity scale developed by Poposki
and Oswald (2010) [
94
]. The notication-related questions included
what kind of devices they had that could be used for interacting
with notications, and their familiarity with technology products.
Then, guided by a sampling strategy referred to as theoreti-
cal sampling [
20
], in which research participants are recruited,
interviews are conducted, and interview data are analyzed concur-
rently [
101
], nine participants were recruited in the second wave of
recruitment. In this wave, the focus of sampling was on diversifying
the sample’s demographic backgrounds, lifestyles, busyness, and
multitasking propensity. For instance, most participants recruited
in our initial wave were on-site oce workers. When we found
that some of them who worked at home on some days interacted
with notications at quite dierent frequencies than when in the
oce, and also exhibited dierent multitasking behavior, we de-
cided to recruit more participants with exible schedules in the
next wave. Aside from these behavioral aspects, we also sought
more gender and occupational diversity. Therefore, in the last two
waves of recruitment, in which we recruited 10 and four partici-
pants, respectively, we recruited participants who had demographic
backgrounds, occupations, and/or work styles dierent from those
in the previous waves. In practice, this meant that more of the later
recruits had exible work schedules and were relatively senior.
We continued recruiting as long as we were obtaining new ex-
periences, ideas, and meanings from each newly interviewed par-
ticipant, and stopped sampling when we deemed that theoretical
saturation had been achieved [
20
], i.e., that no new concepts and
meanings had arisen in the latest round of interviews. The nal
sample of 34 participants (see Table 1) included 10 students and
24 non-students, and 20 females and 14 males, aged between 22
and 64 (M = 28, SD = 10.9). Half self-reported having exible work
schedules, and the other half, xed schedules. Just over half (n=18)
had high polychronicity scores (i.e., above 43.20 on Poposki and Os-
wald’s scale), and the remainder, lower levels of polychronicity. All
participants received US$23 for their participation, and the study
was approved by our university’s Institutional Review Board.
3.2 Study Design
Each participant completed a week-long digital diary [
12
] and two
semi-structured interviews: one before the diary-writing phase, and
the other, after it. The pre-diary interviews were aimed at gaining
a broad understanding of the participants’ lifestyles, multitasking
practices, and notication management. In the post-diary interview,
in contrast, the researcher walked the participants through their di-
ary entries, seeking more context about each one, and explanations
for their choices and actions. The main themes that both interviews
were intended to capture were the interviewees’ intent behind and
practices around attending to notications while they were about to
start, had been engaged in, and had nished their main task at hand.
Due to the COVID-19 pandemic, all the semi-structured interviews
took place remotely via Google Meet, Microsoft Teams, or Zoom (at
the interviewee’s option), and all participants consented to these
interviews being recorded.
3.3 Study Procedure
Each participant was allowed to select the day of their pre-diary
interview, which on average lasted 15 minutes; and the week-long
diary period started the following day. During that period, the
participants were to record all occasions on which they multitasked
by handling smartphone notications plus some other task at hand
– including scenarios in which they were about to start, were in
the middle of, and had recently nished that task – as well as
how they acted upon those notications they saw. A link to the
digital diary was sent out once, later on the same day that the rst
interview was held. The diary template was structured as a top half
and a bottom half. The top half listed some questions about the
participant’s general habits and self-imposed rules for interacting
with notications, including which apps they deemed important,
and the modalities of their notication settings. The bottom half, on
the other hand, asked participants to record their relevant daily-life
events, separated into the morning, lunchtime, afternoon, and night.
However, the participants could modify the template based on their
own habits. For example, some participants who normally woke
up in the afternoon could skip the template’s morning section. We
also told our participants that they had the option (if they had time
to do so, and their workplaces allowed it) of uploading photos to
help us understand their multitasking environment. We checked
daily whether each participant had written in their diary, and sent
them reminders if they had forgotten to do so.
When each person’s diary period ended, the research team con-
tacted them to schedule their second interview. All second inter-
views took place within one week following the end of the diary-
writing phase, to ensure that the participants’ memories remained
relatively fresh [12].
In the post-diary interviews, as well as probing the diary entries
themselves as described above, the researcher asked about the in-
terviewees’ other experiences of and attitudes towards dealing with
notications outside and within a main task; and followed this up
with questions about why they had made particular choices about
notications (i.e., ignoring, glancing at, attending to, or responding
to them) in particular task situations. On average, the post-diary
interviews lasted 62 minutes.
3.4 Data Analysis
All diary entries, photographs taken during the diary period(Fig. 1),
and transcripts were analyzed in parallel with the data-collection
process using Atlas.ti, an online collaborative qualitative-analysis
Motivations for Notification-Interaction CHI ’23, April 23–28, 2023, Hamburg, Germany
ID Age Occupation LifeStyle Preference
P01 27 F Designer Flexible -
Freelance Poly-chronic
P02 33 M Engineer Structured Poly-chronic
P03 31 F Customer Service Structured Mono-chronic
P04 37 F Museum Sta Structured Mono-chronic
P05 26 M Banker Structured Poly-chronic
P06 41 M Travel Agent Flexible -
Slash(Travel Agent/Tutor) Poly-chronic
P07 36 M Engineer Structured Mono-chronic
P08 32 F Administrative Assistant Structured Poly-chronic
P09 25 F Designer Flexible -
Freelance Poly-chronic
P10 29 M UX Researcher Structured Mono-chronic
P11 24 M Student Flexible Mono-chronic
P12 28 F Engineer Structured -
Work-from-home Mono-chronic
P13 26 F Product Manager Structured Poly-chronic
P14 22 F Administrative Assistant Structured Mono-chronic
P15 28 F Product Designer Structured -
Work-from-home Mono-chronic
P16 25 F Product Manager Structured Mono-chronic
P17 38 F Product Manager Flexible -
Work-from-home Mono-chronic
P18 46 M Product Manager Structured Poly-chronic
P19 22 F Student Flexible Poly-chronic
P20 25 M Student Flexible Mono-chronic
P21 24 F Student Flexible Poly-chronic
P22 24 M Product Manager Flexible -
Slash(Student / Product Manager) Poly-chronic
P23 24 F Purchasing Specialist Flexible -
Work-from-home Poly-chronic
P24 23 F Student Flexible Poly-chronic
P25 64 M Artist Flexible Mono-chronic
P26 35 F Political Worker Flexible -
Slash (Student / Political worker) Poly-chronic
P27 38 M Engineer Structured Poly-chronic
P28 22 M Student Flexible Mono-chronic
P29 26 M Student Flexible Mono-chronic
P30 27 M Student Flexible Poly-chronic
P31 64 F Retired Flexible Poly-chronic
P32 45 F Administrative Assistant Structured Mono-chronic
P33 46 F Administrative Assistant Structured Mono-chronic
P34 44 F Library faculty Structured Mono-chronic
Table 1: Demographics of the participants.
Figure 1: Sample photographs provided by four dierent study participants in their diary entries that showed their multitasking
practices of reading notications during their work. All these images were approved by the participants who oered them.
CHI ’23, April 23–28, 2023, Hamburg, Germany Chang and Wu, et al.
software. Specically, we adopted the grounded approach suggested
by Charmaz [
20
], in which, as mentioned earlier, sampling, data
collection, and data analysis take place in parallel. Our data analysis
began with open coding, to explore important concepts in the data
gathered from the initial wave of participants. Generation of the
initial codebook was guided by our interview protocol. Later, the
research team regularly met and discussed the emerging codes
and categories that the team felt they needed to focus on during
focused coding, as well as the new concepts (or the lack thereof, i.e.,
theoretical saturation) obtained from the latest group of participants.
Codes and categories were regularly compared against new data
from new participants; and when new codes emerged or existing
codes were revised, they were compared against existing transcripts
for validation.
This process resulted in the following four main theoretical
coding categories: 1) motivations behind actions upon receiving no-
tications; 2) personal, work, and lifestyle characteristics (including
polychronicity); 3) in-the-moment context (personal, situational,
relational, activity); and 4) timing (i.e., before, during or after the
primary task). It should be noted here that our coding of timing
–, i.e., whether an instance of notication-reading was pre-task,
post-task, or during-task –,depended on how the participant re-
garded that instance when we talked about it in his/her interview.
We observed that the participants’ self-reported timings reected a
clear relationship with the relative demands, in terms of time and
cognitive resources, of the just-completed task and the upcoming
task. That is, when they perceived that they were reading noti-
cations before a demanding upcoming task, they described that
moment as a pre-task timing; but when the more demanding of
the two tasks was already nished, they described the moment as
post-task. In the following sections, we present our study’s key
ndings.
4 KEY FINDINGS
This section is structured as follows. The rst subsection presents
the framework we employed to analyze and showcase the motiva-
tions that participants frequently associated with specic activity
timings for interacting with notications. Then, the second subsec-
tion describes the participants’ motivations for attending to and
dealing with notications in these timings in greater detail.
4.1 Human-Notications Interaction: Timings,
User Motivations, and Factors
The conceptual framework we developed for analyzing and depict-
ing participants’ motivations for interacting with their notications,
and two levels of factors that triggered or inuenced their decisions
about when to interact with notications, is shown in Figure 2.
4.1.1 Personal, Work, and Lifestyle Characteristics.
We found that participants’ decisions about when to interact with
notications were strongly aected by their personal, work, and
lifestyle characteristics. For instance, we found that participants
who described themselves as being in relatively exible and relaxing
daily-life routines – e.g., who worked remotely, were freelancers,
or had retired – tended to be more exible about when they read
notications. For participants who tended to follow a structured
daily rhythm and discipline, on the other hand, catching up with
notications was deemed part of this structure, and the timings for
doing so were more or less embedded in it. These participants were
mostly oce workers with xed work schedules. Specic features
of their work also led participants to prefer reading notications at
specic moments. For example, some participants’ jobs involved in-
tensive interpersonal communication, and thus they had to engage
in frequent notication-checking to accomplish their work. Some
others, whose jobs were creative, felt that checking notications
periodically allowed them to gain inspiration. And others men-
tioned that their companies’ cultures, social norms, and regulations
aected when they felt it was appropriate to check their phones.
Personality traits, habits, ability for sustained concentration, and
level of polychronicity also impacted notication-checking behav-
ior. For example, some participants said they easily experienced fear
of missing out (FOMO), which made it hard for them to delay check-
ing notications until opportune moments. Also, in terms of poly-
chronicity, whereas some participants claimed that concentrating
on one thing at a time was best for their productivity, others argued
that their productivity was highest when they quickly switched
among multiple tasks, or let them proceed simultaneously. Notably,
our participants were quite diverse in these dimensions, and not all
of them mentioned all of these dimensions during their interviews.
Nevertheless, these characteristics oer important explanations of
why participants diered in their preferences for certain timings
over others, despite their situations being similar.
4.1.2 In-the-moment Context.
Contextual factors also inuenced participants’ decisions about
whether to read notications at a given moment. Our results sug-
gested that these occurred in four categories, i.e., personal, situa-
tional, relational and activity factors. Personal factors include, but
are not limited to, one’s attentional, emotional, and physical states.
For example, some participants mentioned that feeling bored made
them want to read notications in that moment. Situational factors
included social surroundings and locations: e.g., some participants
mentioned if they had a meal with their friends, they would prefer
not to use their phone, as doing so might be deemed disrespect-
ful. Relational factors involved their relationships and patterns of
interaction with particular people, typically the senders of the in-
stant messages and emails that had resulted in notications being
delivered. For example, some participants thought that it was im-
portant to address their superiors’ messages in a timely manner,
regardless of their timing. Finally, activity factors described the
participants’ primary tasks in terms of their familiarity; continuity;
divisibility; cost of resumption; requirement for waiting; demands
on cognitive, physical, and emotional resources; anticipation of sub-
sequent events’ resource demands; and social consciousness. For
example, some participants mentioned that cooking occupied both
their hands and was dicult to pause, and thus it was perceived
as not a good time at which to act on notications. Attending a
meeting that they were not closely involved in, on the other hand,
was deemed undemanding of their attention, and thus they felt able
to read notications while waiting for it to begin.
In the following subsections, we rst present our participants’
reasons for deferring their interaction with their notications until
after they have completed tasks, i.e., interacting with notications
after tasks, as these reasons seem the most straightforward and
Motivations for Notification-Interaction CHI ’23, April 23–28, 2023, Hamburg, Germany
Figure 2: Analytical framework for notication-management strategies: The visual grouping of the 12 motivations into three
activity timings, namely before-task, during-task, and after-task, represents the most commonly identied and prominent
associations that participants made between the motivations and the activity timings.
resonant with the ndings of prior research that sought to reduce
interruptions to tasks. Then, we present the participants’ various
motivations for interacting with notications during and before
tasks. Again, it is important to note that these timings were not
mutually exclusive. That is, the assignment of a moment to a before-
task, after-task, or during-task timing depended on the participants’
perceptions, and it was possible, and indeed not unusual, for a
moment between two tasks to be deemed both before-task and
after-task, or for notication-reading to also be associated with
multiple motivations. Additionally, it is important to note that our
framework does not rigidly dene the relationship between activ-
ity timings and motivations. That is, it does not necessarily imply
that these are the only activity timings during which a motivation
can occur. The framework mainly indicates that participants pre-
dominantly associated certain motivations with specic activity
timings. It is possible that in actual practice, a user may experience
a motivation at a dierent activity timing than what was observed
in our study.
4.2 Participants’ Main Motivations for
Interacting with Notications after the
Primary Task
4.2.1 Ensuring the Progress and ality of the Primary Task.
The most common reason our participants mentioned for wanting
to interact with their notications after completing a task was
to protect their performance of the primary task, which can be
further broken down into task progress and task quality [
58
]. This
motivation resonates strongly with prior research that sought to
identify task transitions and opportune moments for delivering
notications [
45
]. Our participants often mentioned this motivation
in combination with an assessment that the task was continuous
and could not be paused (i.e., that they would need to start over if
interrupted), and/or that they themselves needed to concentrate on
it to complete it eciently. As P14 mentioned, “I was working on
this problem set and if someone interrupted me, I would forget what I
was supposed to do next.” There were also cases where the primary
task was perceived as divisible, but the participants preferred not
to interrupt it because it would delay their progress; and they were
especially likely to avoid interrupting tasks when they perceived
that resuming them would take extra eort. For example, P10 noted
in his diary, “I was at work in the afternoon and messages kept
coming in. I’d have delayed my work if I had distracted myself with the
messages. So I didn’t read them until I nished my work.” Participants
also sometimes perceived that a certain level of task quality could
only be ensured and/or achieved when they were fully focused
on the task, as P18 commented about working on high-quality
presentation slides: “I need to be very focused and can’t do anything
else. [...] This is because this level of focus can make me think more
deeply when I work on certain issues.”
4.2.2 Ensuring Enough Resources and Time for Handling Notifica-
tions.
A second reason participants read notications after completing
their primary tasks was to enable them to allocate more of their
cognitive resources and time to dealing with them. They mentioned
CHI ’23, April 23–28, 2023, Hamburg, Germany Chang and Wu, et al.
this reason in connection with a perception that particular notica-
tions would demand relatively large amounts of time and/or eort
to read, think about, and deal with. P11, for instance, noted that
he would not look at email during a task, “even if it’s just trivial or
in one sentence.” The same participant reported not reading news
notications, because “I might encounter some vocabulary that I
don’t know and need to look it up, and this would take me lots of time.
So I’d wait until I had more time for that.” Some notications did not
demand an articulate response, but nevertheless took up time. P02
told us: “Once I start to watch YouTube videos, it’d just take me ten,
twenty minutes. So I’d just glance at what’s there [in the notication
drawer] but watch them during my break or after work.”
4.2.3 Leaving Things for Passing the Time/Avoiding Boredom Later.
A third reason participants gave for reading notications after a
primary task related neither to that task nor to the notications;
instead, they purposely left notications unread until their free/idle
time, specically to ll it. P16 explained:“I postponed reading the
ads sent from my company to some other time [...] Like when I was
waiting for the elevator, I would read newsletters, or emails about
training sent by our company.” In her diary, P08 noted about her
after-work commute that “reviewing notications one by one after I
get on the bus is something I must do every day, until I get out of the
vehicle and walk home.” Later, in her second interview, she reected
on this diary entry as follows: “Those are the things that you do
every day, something that would not take you much time. And you
can’t do anything else on the bus. It’s crowded and so you can’t read
books, unless you have a seat. So it’s phone time. Compared to looking
at Facebook, or Instagram feeds, doing this is more meaningful. It’s
making better use of my time.” P18 likewise commented on deferring
notications to commute time. “This is a period when you can’t do
anything but use your phone. What else can you do, watch people?
So I’d feel that at least you can [. . . ] review your emails, Facebook
posts, messages, and so on.” To sum up, this motivation reected
participants’ widespread aspiration to use their free time wisely.
4.2.4 Leveraging Break-time as Personal Time for Reinvigoration.
Finally, several participants mentioned that they would aggregate
all notications they considered personal (as opposed to business-
related) and postpone dealing with them until their “private time”,
especially after a period of hard work, often as a means of rein-
vigorating themselves. P16, for example, noted in her diary that
after a meeting, she went to a break room on her oce building’s
21st oor, her “ritual space for dealing with private business” and
for “relaxation and private time” such as scrolling Facebook. In her
subsequent interview, when reecting on this diary entry, she said:
“I don’t use my phone during my work. I do it in that break room.”
Several participants said they often needed to work overtime,
which resulted in their personal time being sacriced. Taking small
breaks and dealing with personal business between primary tasks
made them feel that they had earned some “lost” personal time
back. Several mentioned that these interruptions to work were
important to them sustaining their well-being and subsequent work
performance. Some even praised its positive impact on their work
performance, not only because small breaks were refreshing in
themselves, but because during them, they found incentives among
their notications to work more eciently.
4.3 Participants’ Main Motivations for
Interacting with Notications During the
Primary Task
We will now address the primary question of this study: What are
smartphone users’ main motivations for interacting with noti-
cations during a primary task? As the following subsections will
show, many of our participants often did not regard notications
as distractions, but rather, as fullling their momentary functional
and psychological needs, which again, included improving their
utilization of time and personal well-being. As a result, as opposed
to passive feelings of being disturbed or distracted, participants
proactively sought to satisfy these needs by using notications as
an instrument.
4.3.1 Establishing Accountability, Responsibility, and a Proactive
Aitude.
One of the most common motivations participants mentioned for
reading notications during their primary tasks was to establish a
sense of accountability, responsibility and/or proactiveness. As also
discussed in prior research, this was often driven by a need to fulll
obligations due to their work, duty, or position [
54
]. Our own par-
ticipants mentioned that frequently checking for notications was
necessary because some of their duties involved other stakeholders,
and failing to act on certain notications promptly would result
in negative outcomes, such as delaying other people’s decisions or
harming their performance. Therefore, they intermittently attended
to their phone to check if there were any incoming duties or tasks,
accompanied by worry that they might have missed something
urgent. As P04 mentioned in her diary, “In my Yoga class, I would
wear my Mi Smart Band, to sense its vibration and ensure I would
not miss important notications. If it could not receive notications,
it made me anxious. The class is supposed to be relaxing. So doing
this is not good. But I’ve had bad experiences of missing important
notications during the class.”
Sometimes, however, the motivation to check notications dur-
ing a primary task was not rooted in the avoidance of negative out-
comes, but rather, in a desire to feel like a responsible, spontaneous,
or proactive person. That is, we observed that the act of proactively
checking and responding to notications was particularly meaning-
ful to these participants, not only because it helped them become
aware of which tasks might require their attention sooner, but also
because the act itself helped them fulll their self-expectations. P04
explained, “It happens a lot. [...] When I was focusing on something
and my family or colleagues suddenly messaged me to do something,
I’d just stop my work and work on theirs. [...] I feel it’s like a sense
of responsibility. In another example, P13 explained, "If the message
was a question, I presume that it would be more urgent to respond, so
I tended to check if there was any question. Or, if the message came
from my boss, I would assume it to be more important too, which
is also typically true. If you didn’t respond, he might wonder why
you were gone for so long, because it is, after all, working hours."
In other words, for these participants, it was important to look
at arriving notications because they could involve duties – not
necessarily because such duties required urgent execution in and
of themselves, but because being attentive to their duties was vital
to their self-images and value systems.
Motivations for Notification-Interaction CHI ’23, April 23–28, 2023, Hamburg, Germany
4.3.2 Sustaining Connectedness with One’s Social Network and the
Wider World.
The second motivation participants often mentioned for this be-
havior was related to FOMO: a nding that resonates with many
earlier investigations of the reasons behind high notication atten-
dance [
64
], phubbing [
3
], addictive smartphone use [
61
,
97
,
106
],
and distraction [
40
,
85
]. As in many of those prior studies, a com-
mon theme of FOMO in our study was related to social interactions
and a sense of belonging. For example, P24 commented on her de-
sire to take part in her friends’ chat group, despite its overwhelming
number of notications. “They were friends I’m close to and want
to know more about their updates or gossip, like small secrets about
their professors. There are a lot of conversations going on [...]. It’s a
feeling of convergence, a feeling like my friends are around me.”
However, in addition to social connectedness, several partici-
pants’ worries on these lines involved informational connectedness.
That is, they were anxious about becoming “outdated” by missing
breaking news, live updates, and time-limited oers, which felt
they needed to catch up with as soon as possible. Many of them
reported being particularly alert when they received a burst of noti-
cations in a short period of time, on the grounds that there might
be something important happening. When this occurred, to resolve
their worries and curiosity, they would proactively check these
notications. “I was at work, and a lot of messages suddenly popped
up from my friend’s chat group. It makes me curious about what
happened that made them send so many [messages ...]. The more mes-
sages, the more I’m curious, and want to check it out quickly” (P13).
Although these participants were aware that their performance
on their primary tasks would be negatively aected as a result of
their self-interruption for checking these notications, they were
more worried about becoming personally un-synchronized from
the world. This pattern was especially evident when the cost of
missing updates was perceived as greater than the cost of being
distracted from the primary task. For example, P20 mentioned his
close attention to his investment in cryptocurrency, which led him
to frequently check related news and the latest prices so that he
could react quickly to market changes: “If there’s some news about
the issue of certain items, and I didn’t notice it in time, it’d have a
large impact on me. [...] If it has problems, I’d also like to know what
problems there are. [...] I’d try my best to keep myself posted on the
latest news. It’s never a bad thing.”
4.3.3 Pursuing the Feeling of Being Productive during the Primary-
task Period.
Third, we observed that some participants switched back and forth
between several tasks-at-hand and notications because they wanted
to pursue a feeling of being productive by pushing the progress of
multiple tasks in parallel. Several participants reported that doing
so made them feel that they were making full use of their time. This
motivation was mentioned especially when participants felt that
their primary task was not attention-demanding, and therefore, that
if they did not divide their attention and devote some of it to other
things during these “attention surplus” [
68
] or micro-moments [
53
],
these participants felt that the time was not wasted and meaning-
fully used., they felt the available time had not been fully leveraged.
P13, for example, told us: “During my entertainment period at night,
I’d play Switch and computer games at the same time. Meanwhile
I’d have my phone next to me so that I can look at it and respond to
messages while waiting for the game to load.” P30, on the other hand,
wrote in his diary that he read notications related to NBA games
while he was watching NBA, a phenomenon often referred to as
media multitasking [
103
] or second screen [
102
,
117
]. “I see that
hour as my NBA time. [...] The most ecient time to look at these noti-
cations from the NBA app was during [. . . ] boring transitions in the
game itself.” P11 commented similarly on utilizing micro-moments
for interacting with notications while watching YouTube videos:
“I could roughly predict which segment was important or not, like
transitions, trivial talk, something you wouldn’t feel you’d missed out
on if you skipped it. Those segments are short, and so you can’t really
use them for doing something else bigger, so I’d use them for doing
these trivial things [reading notications].”
4.3.4 Leveraging Stimulation and Replenishment to Enhance Pro-
ductivity on the Primary Task.
Finally, the fourth motivation participants mentioned was looking
for stimulation and replenishment from notications to boost their
energy and improve their productivity on the primary task when
bored, tired, sleepy, or simply nding the primary task tedious and
mundane. Even while conceding that notications on their phones
were not necessarily entertaining, they perceived them as a space
in which they could nd diverse, and sometimes unexpected and re-
warding, elements that helped them shift gears and replenish their
minds. As a result, though spending time on notications postponed
the completion of their primary tasks, participants perceived this
delay would be made up for, in the form of better productivity. For
example, P09 noted in her diary: “I went to a coee shop and read
my book. I wasn’t very focused on it. I checked the notication center
and read some emails I got. Then I read a newsletter in there, a couple
of paragraphs, and then returned to my book.” In her subsequent
interview, she reected on this diary entry as follows: I felt the book
a little boring [. . . but] was still trying to be productive. So I did not
open YouTube. I read newsletters, just to shift gears. I spent some time
on that, and then went back to my book. Then I felt the book became
more interesting.”
At other times, participants did not feel bored, or nd their pri-
mary tasks tedious, but were struggling, stuck, or annoyed when
encountering negative emotions during their primary task. Staying
in this state bothered them and triggered them to look at noti-
cations in the hope of sharpening their thinking, adjusting their
mood, relieving pressure, gaining inspiration, or thinking outside
the box. As P01 commented: “My work is creative, and when I feel my
ideas exhausted, I will read notications to nd some inspiring ideas
or something interesting.” P02, on the other hand, found reading
arriving notications during a primary task vital to recovering from
struggles. “When you’re exhausted or in a dead end and can’t get out,
and someone texts you and shares something interesting or pleasant
about his life, it makes a dierence to your mood. Sometimes you’re
in an innite loop and just need external forces to pull you out. It’s
like seeing light at the end of the tunnel.” Similarly, P24 reported
in her diary that exchanging messages with a close friend to re-
lease her negative emotions during work helped her calm down
and continue. “While I was eating my breakfast and modifying the
slides for the election campaign, I received messages from Professor B
that the slides needed signicant revisions. So, to calm myself down, I
CHI ’23, April 23–28, 2023, Hamburg, Germany Chang and Wu, et al.
turned to my two close friends to complain about it. Then I returned
to check those slides that were seen as not good enough to see if it was
on me. [. . . ] I was kind of upset about my increased workload recently.
It happened that another senior classmate texted me and complained
about Professor B. Then I complained too, and worked on the slides at
the same time. Doing these two things at the same time made me feel
much better.”
To sum up, the four main motivations imply that notications
were not always deemed by the participants as an unwelcome
distraction in task-completion contexts. Rather, participants some-
times need them to fulll specic needs, two of which were related
to improving their overall productivity and use of time and main-
taining their well-being, in contrast to the widespread idea that
notications better be suppressed or delayed to prevent them from
harming smartphone users’ productivity.
4.4 Participants’ Main Motivations for
Interacting with Notications Before the
Primary Task.
Finally, participants associated four main motivations for interact-
ing with their notications with the before-task timing. We present
them in turn below.
4.4.1 Pushing the Progress of Other Tasks in Parallel with the Pri-
mary Task. Several participants reported that they preferred to
respond to all notications that needed their responses before start-
ing their primary tasks, specically so that other people could make
progress on other tasks. P03, for instance, said, “I understand the
feeling of waiting until the last minute. I don’t want other people to
wait for me as well”. Others reported that their intention was to
improve the overall eciency of all tasks they were involved in,
and that they did not want their engagement with their own tasks
to become a bottleneck. For example, P13 noted in her diary, “Before
I focused on my meal, I replied to all messages from my friends.” Later,
she commented on this diary entry as follows: “I couldn’t reply to
these messages when I was having my meal. So I replied to all of them
before I started eating and then my friends could reply to me while
I was eating, and then I could reply to them after I nished. Life is
short. Doing this can let us chat more.” She also reported replying
to all messages before leaving her seat at work for an unknown
amount of time. “I don’t know how long it’s gonna take. [...] If they
didn’t get my response, they might not be able to do their next step.
Sometimes they just needed my conrmation. And I don’t want them
to freeze there without making any progress.”
4.4.2 Planning the Rest of the Day. Several participants mentioned
that they preferred to be aware of what notications were in their
notication centers before starting any task. Their main motivation
for this was a perceived need to plan the best use of their time for
the rest of the day. Participants who mentioned this motivation also
had a tendency to seek optimization of their use of time for dealing
with every task-at-hand. Perceiving that notications might bring
to light certain tasks that had to be completed by a specic time,
and that their original schedules would need adjustment as a result,
they wanted to be aware of everything that they needed to act
on before they started, such that time-allocation for all remaining
tasks could be optimized and harmonized. P15, for example, said: “I
plan how to optimize my route when I go out. Some routes will take
me more time, so I have to think.” Likewise, P11 mentioned that
he addressed all notications before a task because “I like to keep
my schedule arrangement perfect. I’d check my notications [before
starting my activity] in case anything would change my schedule.”
In short, checking notications before their primary tasks made
these participants feel as if they were using their time wisely and
eciently.
4.4.3 Reducing Suspense to Focus on the Upcoming Task. Several
participants mentioned that they would check and address all of
their notications, sometimes clearing them all, so that they could
go on to focus on their primary tasks without any lingering feelings
of suspense. That is, if they did not clear their notications, they
would remain uncertain about whether there was some other task
they should have dealt with, but had not. This motivation was
mentioned mainly in the context of perceiving that their primary
task would take up a relatively long period during which they would
not be able to (or want to) check their phones. P11, for example,
noted in his diary regarding clearing notications before a big date,
“I handled all my notications before I went to the coee shop. That
day was a very special day to me. I hoped I could go in a perfect state
without needing to worry about anything. [...] I was discussing her
project with her, and left my phone alone without even giving it a
look.”
4.4.4 Creating a Smooth Transition into the Upcoming Task. Finally,
some participants mentioned that they sometimes read notica-
tions before an upcoming primary task because they wanted to
adjust their mindset or state of mind for that task. To them, these
notication-checking periods were buer zones that helped them
transition into upcoming tasks more smoothly, as P05 reported.
“You’re transitioning to a state of working from a state of drowsiness.
You need time to make that transition and prepare yourself for it. And
what I usually do is check messages, or something related to my work,
and gradually ow into that state.” Moreover, some felt that doing
this helped improve the subsequent quality of their work, as P26
explained. “I am clever when I am working. But my brain is blank
and operates slowly when I am not working. Like I put it in the diary
that day, I ate my breakfast in my oce and read some notications
to help me enter that working state in the morning.”
To sum up, our participants linked four main motivations for
interacting with notications before they started their primary
tasks, all of which were related to productivity and/or improving
their use of time. In the next section, we will discuss the research
and design implications of the identied motivations.
5 DISCUSSION
Our curiosity about what made smartphone users want to interact
with notications during a primary task led us to analyze their
motivations for reading notications at dierent timings, as well
as the contexts that triggered and inuenced such motivations, via
a diary study and semi-structured interviews. Through the result-
ing human-notication interaction framework, we were able to
understand why, during primary tasks, our participants sometimes
wanted to interact with notications despite their supposed status
as a distraction, and how these reasons diered from those given
Motivations for Notification-Interaction CHI ’23, April 23–28, 2023, Hamburg, Germany
for checking notications before and after primary-task completion.
Yet, despite such dierences, the same core/fundamental needs un-
derlay many of the 12 motivations we identied. Below, we discuss
these ndings and their implications in more detail.
5.1 More than as Distraction - Recognizing
Various Functions of Notications-Reading
During the Primary Task
Notications that arrive during a primary task have long been as-
sumed to constitute interruptions, disturbances, or distractions [
14
,
15
,
23
,
37
,
56
,
108
]. As a result, prior studies of interruptibility
have focused on identifying good moments for delivering noti-
cations, such as at activity transitions [
45
], breakpoints within an
activity [
2
,
53
,
80
–
82
,
85
], negotiated moments [
73
,
99
,
114
], and
detected moments [
31
,
51
]. This view of notications as negative
may be due to the fact that many of these studies were conducted
in organizational or workplace settings where the primary tasks
were work-related and productivity was a main concern. While the
current study did not set out to challenge this widely held view of
notications, it aligns with several other studies that have shown
the benets of notications for productivity [
86
] if they are rele-
vant to the task [
14
,
16
,
58
], and the multiple roles notications can
serve [66].
That is, while all of our participants had experienced situations
in which they preferred to interact with notications before or after
a primary task, as a means of protecting its progress and/or the
quality of its outcome, all had also experienced situations in which
they were eager to interact with notications during it. The reasons
behind such eagerness included not only worry about missing
urgent tasks [
66
,
76
,
93
,
100
,
115
] and FOMO [
37
,
41
], but also a
desire to increase their productivity during the current primary-
task period, whether by making use of idle time during the task,
or by stimulating and replenishing themselves when they were
struggling, stuck, or bored. The former motivation was related to a
need to improve personal productivity and self-ecacy, while the
latter concerned personal well-being, which resonated with recent
studies on the benets of taking small breaks to workers [
1
,
55
].
Therefore, at least in certain circumstances, allowing smartphone
users to easily access notications may enhance their productivity
and/or overall performance rather than reduce them.
It is noteworthy, however, that the varying perspectives on noti-
cations observed in our study compared to previous research could
be attributed to the diverse primary tasks included in our daily life
situations, whereas prior studies primarily focused on organiza-
tional or workplace settings. This distinction in setting may have
led to a less critical view of notications as the primary tasks in
our study were not solely work-related. Additionally, participants
may have become more familiar with checking their phones in non-
work situations, resulting in a potentially less negative outlook on
notications. It is also worth mentioning that even though partici-
pants may see notications as a way to improve their productivity
at times, these notications might still be a source of distraction
during these moments.
Furthermore, aside from productivity, participants were also
motivated to create and maintain an image of accountability, re-
sponsibility, and proactivity towards their values. We observed that
for these participants, the purpose of checking notications dur-
ing a primary task went beyond simply keeping abreast of urgent
tasks. Finally, many participants told us about the various types
of information they valued and were excited about, and how they
subscribed to channels and services that brought them these pieces
of information. For example, several prioritized keeping in sync
with the latest world events over their primary tasks, because they
perceived the former as more consequential. Importantly, these par-
ticipants included many who deemed their primary tasks merely
routine or mundane. As a result, it is not surprising that some of
their notications, however disturbing, distracting, or interruptive,
were more appreciated/welcome than the feeling of accomplishing
the primary task.
One vital implication of these results taken as a whole is that
designers of notication systems should consider the relative im-
portance and value of a given task-at-hand and a given notication.
That is, when a user considers certain notications to be more
valuable or consequential than their task-at-hand, blocking or oth-
erwise regulating their notication use during that task may be
counterproductive, not least because it might result in frequent
self-interruption. In other words, if people opt to interact with their
notications despite being in the middle of performing a primary
task, it is vital to recognize that they might be seeking updates
that are important to them, improving their utilization of time, or
recharging themselves, and therefore that suppressing or deferring
notications in such circumstances may undermine their work pro-
cesses. In place of the current one-size-ts-all solution – deferring
notications until after a task is completed – a crucial next step in
improving notication systems will be to learn smartphone users’
motivations for reading notications during primary tasks, and
address such motivations head-on.
5.2 Reading Notications Before vs. After the
Task: Likely in the Same Transitions, but
with Dierent Motivations
Task transitions are widely considered opportune moments for de-
livering notications [
45
], and there is some recent evidence that
recognizing these transitions and delivering notications during
them reduces disturbance [
2
,
63
]. Importantly, from the viewpoint
of the present study’s structure, before-task and after-task timings
can be one and the same, i.e., in cases where two primary tasks are
executed back to back. Nevertheless, we found that participants’
perceptions of whether they were reading notications before or
after a task, or both, depended on how demanding they perceived
the just-completed and the upcoming tasks to be, relative to each
other. Notably, based on our data about the participants’ percep-
tions, it seems likely that all these motivations can arise in the
same notication-reading situation if a person feels that the just-
completed task and the upcoming task are more or less equally
demanding. Regardless of their perceptions of these two timings’
meanings, however, the participants’ motivations for interacting
with notications during them can dier considerably. For instance,
motivations for interacting with notications after a task were of-
ten associated by the participants with task protection, whether it
was the primary task or handling notications themselves. On the
other hand, motivations for interacting with notications before a
CHI ’23, April 23–28, 2023, Hamburg, Germany Chang and Wu, et al.
task were often associated with planning and time management.
Recognizing these dierences – and indeed, the existence of the
before-task timing – has at least two important implications.
First, despite the challenges associated with identifying it, the
distinctness of the before-task timing is vitally important, because
it could drive variation in the types of notications users are look-
ing for in a notication center at certain moments (see Table
??
for a list of examples). For instance, if a user’s interaction with
notications during a task-transition period is mainly for the pur-
pose of protecting the task they have just nished, they may not
be looking for specic notications. On the other hand, if they are
mainly motivated by a need for sucient time and resources to
handle specic notications, they are already targeting those noti-
cations in that period. Either of these motivations may arise from
the just-completed task’s demands on the smartphone user’s time
and cognitive resources. However, if that person is mainly seeking
to plan and improve their use of time, they may be looking out for
notications that will aect their future schedules, and/or that they
need to handle before others if multiple tasks are to progress in
parallel. In the latter case, the upcoming task may be perceived as
demanding users’ time and cognitive resources.
Second, prior research has primarily looked at the during-task
and after-task timings, due to its narrow focus on protecting the
progress and outcome quality of the task-at-hand. The present re-
search, in contrast, shows that the perceived nature of upcoming
tasks matters to how people interact with notications before such
tasks commence. Interestingly, some of our participants perceived
that the presence of notications might also aect their upcoming
tasks, either by altering their existing schedule or by creating unwel-
come feelings of suspense. These results suggest that researchers
and practitioners should consider the joint inuence of upcoming
tasks and the presence of certain notications on users’behavior.
5.3 From Behavior to Motivation: Implications
for Human-notication Interaction
Our proposed human-notication interaction framework distin-
guishes among before-task, during-task, and after-task timings and
tentatively suggests that in these dierent timings individuals may
interact with notications for a particular set of motivations than
for others. We deem that this framework may serve as a possible
basis for considering various approaches to meet users’ needs for in-
teracting with notications and may also provide a useful analytical
tool for furthering research in this eld.
Table 2 presents a list of examples of approaches that designers
of future notication systems might wish to consider incorporating
into such systems to address each of the 12 notication-reading
motivations we identied. This list is not intended to be exhaustive,
as there can be numerous dierent ways of satisfying a given moti-
vation; rather, it is intended to provide inspiration. Notably, looking
beyond any such specic approaches, two common needs under-
lie all 12 motivations, and we recommend that practitioners bear
these in mind when designing future notication systems. The rst
such need is for improvement of smartphone users’ productivity
and utilization of time; and the second is for personal well-being,
albeit sometimes also for the purpose of improving overall perfor-
mance. At a high level, then, we recommend that future notication
systems take these two fundamental needs into account.
Among the approaches to satisfying the motivations listed in
Table 2, one critical capability shared by several approaches is
highlighting and/or sorting notications that users want to see rst.
This resonates with Lin et al.’s [
66
] ndings that smartphone users
desire to see certain kinds of notications in particular prominent
locations, and/or that some notications be distinguished from
others in certain contexts.
Another useful capability would be for the system to automati-
cally defer certain notications until certain timings, which could
reduce not only the overall number of notications being presented
at a given time, but also reduce the distraction that would likely
be caused by presenting them at inappropriate times (e.g., before a
task, causing unwelcome feelings of suspense). On the other hand,
all these approaches will call for automated classication of no-
tications’ content and/or roles before the two aforementioned
capabilities can become fully edged.
Finally, it will almost certainly be challenging to know how de-
manding users will think a task is, and thus whether they will
conceive of themselves as being in a pre-task, post-task, or during-
task moment. However, if a system has access to users’ calendar
events, it can infer its users’ intentions/motivations, and further
learn them over time, based on the features of their just-completed
and upcoming tasks and of the notications they interact with. It
might be also worthwhile to verify such intentions/motivations
via prompted questionnaires, a technique that has been employed
not only by researchers [
110
] but in commercial smartphone apps
such as Google Opinion Reward.
1
This would allow a system to
learn patterns of when and how often a user has specic motiva-
tions for interacting with notications and help it sort or highlight
notications based on the motivations it infers.
In terms of its signicance for human-notication interaction
research, our proposed framework extends and augments an exist-
ing body of literature that has hitherto mainly focused on users’
preferences [
66
,
77
,
90
,
93
,
100
], perceptions [
14
,
15
,
23
,
37
,
56
,
108
],
and practices of interacting with notications [
6
,
19
,
59
,
85
,
111
,
114
,
115
]. More specically, our framework sheds new light on the
varying likely motivations that underlie users’ actions involving
notications, and has several important implications. First, it en-
courages researchers to revisit current understandings in the eld of
human-notication interaction and relate them to the motivations
we identied. Taking users’ prioritized and selective attendance to
notications, for example, previous work (e.g., [
18
,
66
]) has con-
sistently reported that users read and act upon notications from
certain apps, from certain senders, or of certain types rst. It would
therefore be worthwhile to examine how users’ selective atten-
dance and choices at dierent timings are associated with the 12
motivations. For instance, users may want to read social-media
notications due to FOMO, or a desire to replenish themselves,
or a need to kill time. The latter two motivations could also be
fullled by other notication types, e.g., those pertaining to news
articles or educational content. In our framework, researchers have
a new tool for explaining the notication-attendance behaviors
and preferences they observed in their studies. The listed moti-
vations can also be used as questionnaire items in studies using
1Google Opinion Reward:
https://en.wikipedia.org/wiki/Google_Opinion_Rewards
Motivations for Notification-Interaction CHI ’23, April 23–28, 2023, Hamburg, Germany
Table 2: Examples of approaches to addressing each of the 12 notication-reading motivations
Associated
Timing Motivation Possible Approaches to Satisfying
the Motivation Potentially Relevant
Notications Types and Content
Before-task
Planning the rest of day Highlight notications that would aect users’
schedules and plans for the rest of the day.
e.g. notications from IM, email, and
work-related apps containing task
or schedule information.
Pushing the progress of other
tasks in parallel with the
primary task
Highlight notications that users need to handle
before others if multiple tasks are to progress
in parallel.
e.g. notications from IM, email, and
collaboration apps containing
time-critical task requests and inquiries.
Reducing suspense to focus on
the upcoming task
Highlight time-critical notications but defer those
among the other notications that would demand
users’ signicant cognitive eort and time until
after they have completed the upcoming task.
(for deferral) e.g. notications from IM,
email, and work-related apps with time
and cognitively demanding but
non-time-critical task requests and inquiries.
Creating a smooth transition
into the upcoming task Highlight notications that provide context for or
are otherwise relevant to the users’ upcoming task. e.g. notications related to the users’
upcoming task.
During-task
Establishing accountability,
responsibility, and a proactive
attitude
Alert users of time-critical tasks that need their
immediate attention and allow them to specify
such tasks by providing rules, such as about senders,
apps, keywords, etc. Devise an area specically for
highlighting these types of notications so that users
can quickly notice their presence.
e.g. notications from IM, email, and
collaboration apps containing
time-critical task requests and inquiries.
Sustaining connectedness with
one’s social network and the
wider world
Keep users posted with specied updates, likely be
grouped, summarized, and presented separately
from other notications. Only send alerts for those
updates that require users’ immediate attention
e.g. notications from IM, social media,
news, nancial, or others checked
frequently for critical real-time updates.
Pursuing the feeling of being
productive during the
primary-task period
Indicate notications