Computers in Human Behavior Reports 10 (2023) 100299

Contents lists available at ScienceDirect

Computers in Human Behavior Reports

journal homepage: www.sciencedirect.com/journal/computers-in-human-behavior-reports

Do persuasive designs make smartphones more addictive? - A

mixed-methods study on Chinese university students
Xiaowei Chen a, b, *, Anders Hedman c, Verena Distler a, Vincent Koenig a
a
Human-Computer Interaction Research Group, University of Luxembourg, 11 Porte de Sciences, L-4366, Esch-sur-Alzette, Luxembourg
b
Institute for Advanced Studies, University of Luxembourg, Campus Belval, L-4365 Esch-sur-Alzette, Luxembourg
c
Division of Media Technology and Interaction Design, KTH Royal Institute of Technology, SE-10044, Stockholm, Sweden

A R T I C L E I N F O A B S T R A C T

Keywords: Persuasive designs have become prevalent for smartphones, and an increasing number of users report prob­
Persuasive designs lematic smartphone use behaviours. Persuasive designs in smartphones might be accountable for the develop­
Persuasive technology ment and reinforcement of such problematic use. This paper uses a mixed-methods approach to study the
Problematic smartphone use
relationship between persuasive designs and problematic smartphone use: (1) questionnaires (N=183) to
Smartphone addiction
HCI ethics
investigate the proportion of participants with multiple problematic smartphone use behaviours and smartphone
designs and applications (apps) that they perceived affecting their attitudes and behaviours, and (2) interviews
(N=10) to deepen our understanding of users’ observations and evaluations of persuasive designs. 25% of the
participants self-reported having multiple problematic smartphone use behaviours, with short video, social
networking, game and learning apps perceived as the most attitude- and behaviour-affecting. Interviewees
identified multiple persuasive designs in most of these apps and stated that persuasive designs prolonged their
screen time, reinforced phone-checking habits, and caused distractions. Overall, this study provides evidence to
argue that persuasive designs contribute to problematic smartphone use, potentially making smartphones more
addictive. We end our study by discussing the ethical implications of persuasive designs that became salient in
our study.

1. Introduction (eCorner, 2013). However, studies have shown that the “like” button
negatively affects users’ mental health, resulting in social comparisons
Fogg was one of the first scholars to research computers as persuasive and increased envy and depression (Blease, 2015). On the other hand, in
technologies (Fogg, 1998). In the early days of persuasive technology, the context of the attention economy, persuasive designs (for example,
designers focused on promoting positive attitudes and behaviour recommendation algorithms on video and e-commerce platforms)
changes in the health, economics and education fields (Hamari et al., insatiably seek users’ attention and consume their leisure time (Wil­
2014). Later, persuasive designs became ubiquitous online, with attitude liams, 2018). Companies adopt persuasive designs to prolong users’ time
and behaviour change design methods being increasingly integrated on their digital services and seek to make their products more engaging
into social networking, retail, news and entertainment sites. Billions of and habit-forming than those of their competitors (Eyal, 2014). Experts
users are exposed to such persuasive designs each day by internet giants fear that the precipitous rise in smartphone and social media usage
such as Facebook, TikTok, Amazon, Netflix, Alibaba, YouTube and (Rosenquist et al., 2021) is leading to mental (Lei et al., 2020) and
others. physical harm (Kim et al., 2015), especially among children and ado­
Persuasive designs might negatively affect users’ attitudes and be­ lescents (Lewis, 2017).
haviours (Borgefalk & Leon, 2019). On the one hand, for products Screen time and mental/physical problems associated with smart­
designed to serve their customers better, it is possible that unintended phone use are increasing worldwide (Busch & McCarthy, 2021; Olson,
impacts on users might arise despite firms’ good intentions. One Sandra, Colucci, et al., 2022; Rozgonjuk et al., 2018). Studies have found
prominent case is the introduction of the Facebook “like” button, which that some mobile applications (apps) predominantly occupy users’
was intended to enable users to share affirmation and positivity easily screen time, i.e., lifestyle, social networking (Noë et al., 2019) and

* Corresponding author. Human-Computer Interaction Research Group, University of Luxembourg, 11 Porte de Sciences, L-4366, Esch-sur-Alzette, Luxembourg.
E-mail addresses: xiaowei.chen@uni.lu (X. Chen), ahedman@kth.se (A. Hedman), verena.distler@uni.lu (V. Distler), vincent.koenig@uni.lu (V. Koenig).

https://doi.org/10.1016/j.chbr.2023.100299
Received 13 September 2022; Received in revised form 7 May 2023; Accepted 20 May 2023
Available online 22 May 2023
2451-9588/© 2023 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
X. Chen et al. Computers in Human Behavior Reports 10 (2023) 100299

instant messaging apps (Ding et al., 2016). Furthermore, Noë et al. found
some features of Snapchat that make users prolong their usage, for
example, making friends compete against each other for the top position
(competition) and incentivising users not to break their “Snapstreaks”
with close friends (reward) (Noë et al., 2019). It seems legitimate to
consider Snapchat’s success in light of its integration of these persuasion
strategies.
The majority of studies on applications of persuasive technology are
in the health, wellness, and education domains(Devincenzi et al., 2017;
Orji & Moffatt, 2018); recently, scholars have paid growing attention to
its adverse effects on volunteerism, privacy, transparency, and users’
awareness (Nyström & Stibe, 2020). However, the relationship between
persuasive designs and problematic smartphone use (PSU) is less stud­
ied. This study makes two main contributions.

(1) Our findings extend existing knowledge on how users perceive

the influence of persuasive design features.
Fig. 1. The PSD model (Oinas-Kukkonen & Harjumaa, 2009).
(2) We identify potential relationships between persuasive designs
and problematic smartphone use behaviours, and links to
“feedback”, “social support, sharing and comparison”, “reminders”,
addiction-like behaviours.
“alert, reward, points, credits”, “objectives”, and “personalisation”.
Kaptein et al. (2015) presented persuasion profiling as a method of
Our study makes empirical contributions relevant for researchers
personalising messages, which represent different persuasion principles,
and designers in Human-computer Interaction (HCI), as well as psy­
to influence users. Oyibo and Vassileva (2019) studied how participants
chologists, by providing qualitative insights into the longitudinal effects
with different personality traits exhibit different levels of susceptibility
of persuasive designs on technology users. This study is among the first
to social influencing strategies (i.e., social learning, social proof and
to use a mixed-methods approach to investigate the relationship be­
social comparison). All of these studies advance our knowledge of how
tween problematic smartphone use and the design features of smart­
different persuasion strategies could be deployed in designing persua­
phones in China, one of the countries with the most problematic
sive applications.
smartphone use (Busch & McCarthy, 2021; Olson, Sandra, Colucci, et al.,
Ethical concerns regarding persuasive technology: Berdichevsky
2022). This is timely in an era in which many cutting-edge technologies
and Neuenschwander (1999) discussed the potential negative impacts of
are applied to exert influence on users across the world.
persuasive technology on its users and proposed a guideline for de­
The remainder of this paper is organised as follows. In section 2, the
signers. They formulated a golden rule: interaction designers should
definitions, applications and ethical concerns of persuasive technology
never seek to persuade users of “something they themselves would not
and studies about PSU are examined; subsequently, research questions
consent to be persuaded to do" (Berdichevsky & Neuenschwander, 1999).
are formulated. In section 3, the study methods and data analysis process
Fogg (2002) was concerned that ethical issues regarding persuasive
are described in detail. We then present our results in section 4 and
technology could be the same as those for persuasion in general and
discuss the results in section 5. In section 6, we present our conclusions.
recommended that designers perform stakeholder analysis in compli­
cated situations. In addition, Fogg predicted that persuasive technology
2. Related work
might encounter increasing scrutiny from policymakers because of its
potential impact on the public, prompting regulatory efforts to guard
Definitions and applications of persuasive technology: Fogg
against certain coercive tactics and protect specific audiences. Smids
(2002) defined persuasive technology as “interactive computing systems
(2012) found that persuasive technologies often influence users beyond
designed to change people’s attitudes and/or behaviours, without using
the point of self-control. Smids (2012) recommended that persuasive
coercion or deception”(p.15) (PT1). He excluded unethical applications
technology designers perform voluntariness assessments, specifically
from the definition. However, when Kampik et al. (2018) studied the
reviewing whether there is manipulation (“influence users in ways of
persuasive properties of several popular apps, including Duolingo,
which they are not aware and cannot control”) and coercion (“apply direct
Facebook, Slack, and YouTube, they found prevalent use of deception
force or credible threat to control users”) in their persuasive designs and
and coercion and suggested redefining persuasive technology as “any
whether users act intentionally during persuasion. Timmer et al. (2015)
information system that proactively affects human behaviour, in or against
analysed new challenges brought on by the integration of persuasive
the interests of its users” (PT2). Kampik et al. (2018) listed four core in­
technology and ambient intelligence. They emphasized transparency
dicators of persuasive technology, i.e., intentionally persuasive,
about the interests of different stakeholders and proposed that users and
behaviour-affecting, technology-enabled, and proactive (“some extent of
providers should have open discussions and reach agreements on the
autonomy or a high degree of automation”). This study adopts the PT2
goals, methods and interests of persuasive technology to achieve trust­
definition and applies the four core indicators in identifying persuasive
worthiness. Borgefalk and Leon (2019) observed the rise and prolifera­
designs.
tion of digital platforms that use persuasive designs in business
Over the past two decades, multiple frameworks and psychological
operations, proposing interdisciplinary research approaches, combining
theories have been introduced to the persuasive technology design
persuasive technology, governance, and management studies to address
process as persuasion strategies. Oinas-Kukkonen and Harjumaa (2009)
associated ethical challenges.
developed Fogg’s taxonomy of persuasive design principles and pro­
PSU and smartphone addiction: There is a divide in academia on
posed a framework for the design and evaluation of persuasive sys­
the most precise term to describe users’ prolonged screen time and the
tems—the Persuasive System Design (PSD) model (see Fig. 1). The PSD
mental/physical problems associated with smartphone use (Billieux,
model divides persuasion strategies into four categories: primary task
Maurage, et al., 2015). Based on the internet addiction scale, Kwon, Lee,
support, dialogue support, system credibility support, and social sup­
et al. (2013) validated a self-diagnostic scale for smartphone addiction.
port. Orji and Moffatt (2018) analysed 85 articles on persuasive tech­
More than 3300 articles have cited the scale and its short version (Kwon,
nologies for health and wellness. They found that the most frequently
Kim, et al., 2013) as of April 2023. By contrast, some scholars use the
employed strategies in this field were “tracking”, “monitoring”,

2
X. Chen et al. Computers in Human Behavior Reports 10 (2023) 100299

term PSU to study the “craving to use a smartphone in a way that is difficult that 89% of smartphone interactions are initiated by users, not by no­
to control and leads to impaired daily functioning” (Busch & McCarthy, tifications. Importantly, they suggested investigating the routines and
2021). Lanette et al. (2018) noted that the expression “smartphone habitualised behaviours users develop over time in interacting with
addiction” was overused in academia and the media. They questioned smartphones. One possible explanation is that persuasive designs in
the objectiveness of self-reported scales and stressed the situated and smartphone operating systems and apps cultivate and reinforce users’
complex nature of phone use. They suggested looking deep into smart­ habitual behaviours. Studies have identified persuasive designs in pop­
phone functionalities and users’ self-reflective behaviour. Likewise, ular apps such as YouTube, Facebook, Snapchat, Flipkart and AliExpress
Billieux, Schimmenti, et al. (2015) advised caution in overpathologizing (Adib & Orji, 2021; Kampik et al., 2018; Noë et al., 2019), while
everyday behaviour as behavioural addictions. They highlighted the persuasive technology is effective in motivating users to change be­
multi-faceted nature and heterogeneity of daily life disorders. Therefore, haviours (empirical studies) (Cellina et al., 2021; Khalil & Abdallah,
this study does not aim to examine which terms describe current 2013) and in sustaining behaviours (theoretical framework) (Kaptein
smartphone use problems most precisely; it instead seeks to study factors et al., 2010).
that contribute to the problematic situation, and studies using both Teenagers spend prolonged time online with social media and games,
terms (“smartphone addiction” and “PSU”) will be referred to. which might be linked to persuasive designs (Daniel, 2018). Cemiloglu
Almourad et al. (2020) reviewed different definitions of digital et al. (2021) compared theories explaining digital addiction behaviours
addiction in 47 studies, including internet, gaming and smartphone with the PSD model, suggesting that certain PSD strategies, such as
addiction. Several features were identified and classified into categories reduction (simplifying tasks), rewards (rewarding target behaviours),
that provide an image of digital addiction with respect to four aspects, i. social comparison (comparing oneself with others), liking (being visu­
e., device usage, social interaction, psychological states, and clinical ally attractive to users) and personalisation (offering personalised con­
symptoms (see Fig. 2). In addition to providing a holistic view of the tent or services), may trigger and expedite digital addiction in specific
digital addiction research field, they observed a set of shared features contexts. To the best of our knowledge, there is no empirical study on the
across terms, including impulsivity, compulsion, lack of control, nega­ relationship between persuasive technology and PSU; as a result, we
tive emotional outcomes, and impairment in work and study (Almourad formulate the following research questions (RQs) to fill this research
et al., 2020). This integrated approach considers the diversity of gap:
smartphone functionalities, given that smartphones have become syn­
thetic devices for communication, connectivity, and gaming. RQ1. What proportion of study participants self-report having mul­
For the purposes of this article, we will use the term Problematic tiple problematic smartphone use behaviours?
Smartphone Use (PSU) to refer to users overusing smartphones to a RQ2. Which smartphone designs and apps do participants report
degree that causes a perceived negative impact on their productivity, influence their attitude or behaviour?
mental/physical health, and other aspects of life. RQ3. How do participants perceive persuasive designs and their
Scholars have striven to identify the factors that contribute to PSU. influences on smartphone use?
Some researchers investigated users’ susceptibility to PSU. They found
that user characteristics, such as antagonism and negative affect, serve 3. Research methods
as positive predictors for latent PSU behaviours (Marciano et al., 2021).
Further, researchers found that fear of missing out (FOMO)/lower 3.1. Overview
well-being/distress were correlated with PSU (Della Vedova et al., 2022;
Elhai et al., 2016; Horwood & Anglim, 2019). Higher levels of FOMO We used a mixed-methods approach to answer our research ques­
were found to be associated with greater impact of social media on one’s tions. We administered a questionnaire to answer RQ1 and RQ2 (section
daily activities and productivity (Rozgonjuk, Sindermann, Elhai, & 3.2) and interviewed ten study participants to provide additional an­
Montag, 2020). Meanwhile, scholars studied users’ interactions with swers to RQ1 and answer RQ3 (section 3.3).
smartphones to explore why smartphones are disruptive. Frequent
phone-checking habits make up a substantial amount of users’ smart­
3.2. Questionnaire
phone use (Oulasvirta et al., 2012). Heitmayer and Lahlou (2021) found
3.2.1. Participants
Participants were recruited through multiple channels. We published
the link to the questionnaire on the university intranet forum (Beijing
Institute of Graphic Communication), student group chats (Energy and
Sustainability program at Zhejiang University), and Tencent question­
naire service1 (Survey distribution set to college students aged 18 to 26).
248 users completed the survey. With Tencent’s automatic spam screen2
and manual age-grade consistency check (i.e., Chinese students, in
general, start their freshman year between age 17-19 years old; This age-
grade consistency check filtered out participants who deviated more
than two years from the average number), 183 questionnaires were
verified as valid.
The participants of our study are Chinese university students. There
were 90 male and 93 female participants, ranging from 18 to 26 years
old (mean=21.7, SD=1.8). 83% (n=152) of the participants use Android

1
Tencent questionnaire service: a free and professional questionnaire
design and distribution platform operated by Tencent company, the largest
social networking company in China.
2
Tencent automatic spam screen: the Tencent machine learning algorithm
studied respondents’ answering behaviour, number of answers, question types,
Fig. 2. Digital Addiction features (Almourad et al., 2020). and other factors to filter spams.

3
X. Chen et al. Computers in Human Behavior Reports 10 (2023) 100299

smartphones, while 17% (n=31) use iPhones (see Fig. 3 for the histo­ a) No. b) Very rarely. c) Rarely. d) Occasionally. e) Frequently.
grams by age, gender, and operating system). The most common study Q11. Does the smartphone negatively affect your studies or profes­
programs in the survey sample were engineering (n=27%), economics & sional life?
management (n=25%), computer science (n=23%), and e-commerce & a) No. b) Very rarely. c) Rarely. d) Occasionally. e) Frequently.
marketing (n=14%).
Perception of persuasive applications Q12-Q13 investigate par­
3.2.2. Procedure ticipants’ perceptions of persuasive applications on their smartphones.
We pre-tested the questionnaire with four university students. The Specifically, the participants contribute to identifying intentionally
first two students were from a Swedish university and complete the pilot persuasive and behaviour-affecting features, and we examine whether
questionnaire in English. They provided valuable feedback for con­ the mentioned apps/OS features are technology-enabled and proactive.
structing the smartphone usage and open-ended questions. Then, the The questionnaire ends with Q14, about whether the participants would
pretests were held with two Chinese university students using a Chinese be willing to participate in an interview:
translation of the questionnaire. All pretest sessions were conducted via
Zoom. We collected the data using the Chinese version of the Q12. Are there any apps that have changed your attitude or behav­
questionnaire. iour? (If yes, please elaborate briefly.)
The questionnaire was published on April 3, 2021, and data collec­ Q13. Are there any functions, apps, or designs of your smartphone
tion remained open until April 25, 2021. During this period, no lock­ that led you to develop new habits? (If yes, please elaborate briefly.)
down measures were implemented in the cities in which our participants Q14. Would you like to participate in a 30-minute interview about
resided. However participants needed to use smartphones to track their your smartphone usage habits? (If yes, please leave your contact
contacts and install e-learning apps. Questionnaires could be answered details.)
through web pages, QQ, and WeChat. On average, users took 3 minutes
to complete the survey. Each valid survey participant received ¥3 (US 3.3. Interview
$0.5) as compensation.
3.3.1. Participants
3.2.3. Materials Ten interviewees were sampled on April 6, 2021, from the survey
The questionnaire included multiple-choice questions and open- participants who had submitted contact information; 172 valid answers
ended questions in three sections: had been collected by then. Participant gender and screen time were
Demographics Q1-Q5: Participants’ age, gender, smartphone taken into account to match the survey sample distribution. We sampled
operating system (OS), study program, and year in university (i.e., five females and five males. The sampled interviewees reported
Bachelor [freshman, sophomore, junior, senior] or graduate program spending on average 6.0 h/d on their smartphones, while the survey
[master’s student, PhD student]). participants reported 5.6 h/d. Interviewees came from various study
Smartphone usage Q6-Q7: We asked the participants to report their programs, including energy and sustainability, computer science, media
average daily screen time and gaming time. Q8-Q11: We adapted and civil engineering (see Table 1).
questions from published studies and chose loss of control, perceived
negativeness and overuse as key PSU indicators (Huang et al., 2021), 3.3.2. Procedure
including: We designed and pre-tested interview questions using the same
procedure as described above for the questionnaire. The in-depth in­
Q8. Do you feel that the use of smartphones takes up too much time? terviews started with a self-evaluation of their problematic smartphone
a) Yes. b) No. c) Hard to tell. d) Occasionally. use. Then, we asked interviewees questions related to smartphone usage
Q9. Have you tried to control your smartphone usage time? and screen time. Next, the interviewees were asked to identify persua­
a) Yes, I reduced my usage time. sive designs that they could observe in their smartphone usage; if the
b) Yes, but I failed to reduce my usage time. interviewees weren’t familiar with the concept of persuasive technology
c) No, I do not intend to reduce my usage time. before the interview, the definitions, applications and examples of
d) No, but I plan to reduce my usage time in the future. persuasive technology were explained to them. The interviews ended
Q10. Do you inadvertently use your smartphone for longer times with a self-evaluation of how the persuasive designs observed by in­
than you planned? terviewees affect their screen time and smartphone usage habits.
The interviews lasted between 18 and 45 minutes and were con­
ducted remotely via WeChat voice call. The interviews were recorded
with permission. Each interviewee was compensated ¥50 (US$7.3) for
participation.

3.3.3. Materials
A Chinese translation of Fig. 2 was presented to the interviewees, and
we provided the definitions of the DSM- 5 Factors and ADHD as foot­
notes to the interviewees. We referred to the recommendation and red
dots notification designs of WeChat as examples of persuasive technol­
ogy in question 5, because all participants used WeChat and were
familiar with these features. We prepared the following seven questions
for the interviews:

1. Would you mind going through the Digital Addiction Features

(Fig. 2) and telling me which features match your experience?
2. Please indicate the occasions when you have to use your smartphone
daily.

Fig. 3. Age, gender, operating system count.

4
X. Chen et al. Computers in Human Behavior Reports 10 (2023) 100299

Table 1
Summary of interviewee information.
Participants Ages Gender OS Study Program Grade Screen Time (h/d)

P1 19 Female Android Media Freshman 6.0

P2 22 Male Android Civil Engineering Senior 7.0
P3 22 Male Android Computer Science Master Student 4.0
P4 22 Female iOS Mathematics Senior 6.0
P5 22 Female Android Language Senior 7.0
P6 23 Female Android Medicine Junior 8.0
P7 23 Male iOS Computer Science Senior 6.0
P8 23 Male Android Energy and Sustainability PhD Student 5.0
P9 24 Male Android Energy and Sustainability Master Student 5.0
P10 24 Female iOS Media Master Student 6.0

3. Please evaluate the hours needed for these essential occasions. What 4. Results
are the factors that caused you to spend more time on your
smartphone? 4.1. Questionnaire
4. Have you heard about Persuasive Technology before? (If yes, can you
elaborate a bit.) 4.1.1. Problematic smartphone usage (RQ1)
5. Discuss Persuasive Technology definitions by Fogg (2002) and In this section, we describe what proportion of users self-reported
Kampik et al. (2018) and applications. having multiple PSU behaviours.
6. Can you recognise some persuasive applications/features/designs on The participants reported spending on average 5.6 h/d using their
your smartphone? smartphones. 15% (n=28) of them spent less than 4 h/d using their
7. Would you mind evaluating the impact of the above-mentioned smartphone, while 85% (n=155) spent 4 h or more. On average, female
persuasive applications on your smartphone use? participants used their phones 5.9 h/d, while male participants used
theirs for 5.4 h. We would like to point out that the screen times in this
3.4. Data analysis study are self-reported, with the exception of Table 3 and interviewees’
transcripts, which were documented based on screenshots. We will
The questionnaires were collected in Chinese, and then the raw data discuss this further in the study limitations.
were translated from Chinese to English. For screen time and gaming 67% (n=122) of participants indicated that they spent too much time
time, means were computed by gender and operating system. Quanti­ on their smartphones. 83% (n=152) of participants had tried to control
tative analyses were performed using Microsoft Excel and SPSS. For their smartphone usage time; among them, 58 (mean=5.3 h) partici­
open-ended questions, sentiment analysis was performed using Excel pants were able to reduce their screen time, while 94 (mean=5.8 h)
Azure Machine Learning. The results were also manually checked for participants failed to do so. 122 (67%) participants (frequently and
errors. Data were visualised using Python Seaborn Library (Waskom, occasionally) used their smartphones for longer times than planned,
2021). while 81 (44%) participants (frequently and occasionally) thought
We used thematic analysis (Clarke et al., 2015) to analyse the smartphones negatively affected their studies or professional life.
qualitative data collected from the interviews. The audio recordings To sum up, 25% (n=46) of participants reported multiple PSU be­
were transcribed to text via the iFlytek3 automatic transcription service, haviours at the same time, including smartphone overuse (Q8: a or d and
then manually checked and extracted by author 1. The interviewee Q10: d or e), had failed to control or planned to control their phone use
transcripts were translated into English by author 1. Authors 1 and 2 in future(Q9: b or d), and felt that they had been negatively affected by
discussed together how to choose matched codes in the coding process phones (Q11: d or e). They reported spending 6.3 h daily on their
and search for the relevant themes. The themes were reviewed and phones.
discussed between the authors in the process of structuring and
composing the results section. 4.1.2. Reported persuasive features (RQ2)
In this section, we describe the perceived persuasive applications
3.5. Privacy and ethical considerations reported by survey participants.
145 (79%) participants answered the open-ended question: Are there
The study was conducted at a Swedish University which does not, any apps that have changed your attitude or behaviour? Among participants
under normal circumstances, require ethical review board approval for who filled in answers, 38 participants only mentioned app names, with
HCI studies. Participants who were younger than 18 years old were no specification of how these apps influenced them. Consequently, 107
automatically deleted from the survey results to protect the privacy of valid answers were analysed by Azure to identify sentiments. The most
minors. The quantitative and qualitative data collected in the study were mentioned apps were TikTok4, WeChat (social networking), Honor of
analysed anonymously, and participants’ identity was protected. The Kings (a popular game), Kuaishou (short video platform), Little Red
collected data were stored in an encrypted hard drive. We informed all Book (social networking), Weibo (Chinese Twitter) and Taobao (e-
study participants of the purpose of the study and their right to with­ commerce) (see Table 2, the most mentioned apps that changed users’
draw at any time. We obtained verbal consent from all interviewees attitude or behaviour with sentiment analysis, count, and coded com­
prior to the interviews. ments). These are all popular apps among young Chinese. Surprisingly,
TikTok, WeChat, Honor of Kings, and Taobao were also the most
frequently mentioned as having negative influences on participants.
“Time” was the most frequently mentioned keyword, which was
recorded in 23 answers. Positive sentiments were associated with

3
iFlytek: a Chinese information technology company specialising in voice
recognition and communication technology. We employed automatic tran­
4
scription, and the audio and transcripts were deleted from the researchers’ TikTok: TikTok in this study refers to Douyin which is the Chinese version
iFlytek account after downloading them. of TikTok international version.

5
X. Chen et al. Computers in Human Behavior Reports 10 (2023) 100299

Table 2
The most mentioned apps that have changed participants’ attitudes or behaviours.
Apps Sentiment Analysis Count Comment

Pos. Neu. Neg. N.

D.

TikTok 11 2 10 5 28 Learn skills/Broad horizon/Inspirational/Kill boredom/New Knowledge/Lose control/Can’t stop/Too much

time/You can find everything on it
WeChat 5 2 8 4 19 Help me with social/Inspiring tweets/Convenient to make payment/Without notice, half of my screen time was on
social media/FOMO/Constantly check my phone/Distraction
Honor of Kings 3 7 10 Angry/stay up late/Decreased self-control and study time
Kuaishou short video 4 3 2 9 Life tips/Diverse/You can watch everything/Vulgar/Don’t want to put down your phone/not so optimistic
Little Red Book 7 2 9 Learn new knowledge/Different people and diverse world/Help me with dress code
Learning apps (2); 6 2 8 Learn a lot in extracurricular/TED videos/Help me with my studies/Learn knowledge and ideas about life and
XueXiaoyi (1); science
Xuexi Tong (1);
Zhihu (4)
Forest (4); 6 1 7 Focus/Reduce time playing on my phone/Concentrate more on studying/Schedule my study time
Toma Todo (3)
Weibo 4 1 2 7 Diverse world/Meet new people/Check phone too frequently
Taobao 2 3 1 6 Addiction/No longer enjoy shopping malls/Spent too much money on it

Notes: Pos. for Positive; Neu. for Neutral; Neg. for Negative; N.D. for No Detail.
‘(n)’ in Apps means count of the specific app.

Countdown (a timer app with schedule features), Forest (time man­ users was both positive and negative; social networking, time manage­
agement), Douban (an online community of book, music and movie ment, and digital payment apps were the most mentioned apps leading
lovers), e-learning, Screen Time (iOS and Android digital health func­ to new habits; additionally, some of these apps lead users to adopt daily
tions), and Toma Todo (a timer app with screen locker function). In check-in habits.
contrast, negative sentiments were linked to Honor of Kings (decreased
self-control), TikTok (spent too much time on TikTok, “cannot stop”),
WeChat (half of the screen time goes by without noticing it). 4.2. Interviews
More than ten users positively commented on a set of lifestyle, hobby
and learning apps: Keep (fitness), Mint (healthy diet), National Karaoke 4.2.1. Self-evaluation of smartphone use (RQ1)
(hobby), Kuwo Music, Duolingo (language learning), and Fluently In this section, we document interviewees’ self-evaluations of
Speaking (English learning). Words such as “fun”, “inspiring”, “helpful”, smartphone use and their reflections.
“time well-spent”, and “productive” were found in these positive After going through the features of digital addiction (Fig. 2), in­
comments. terviewees indicated which experiences matched their own. The most
125 (68%) participants answered the question: Are there any func­ frequently mentioned were using smartphones “over 4 h per day”,
tions, apps, or designs on your smartphone that led you to develop new “habitual checking/unconsciously unlocking phone”, “checking specific
habits? Among participants with filled-in answers, 26 participants did content on smartphones”, “time distortion/forget about time”, and
not elaborate on their answers. Azure marked 4 of the answers as “prolonged usage”.
negative, 25 as neutral, and 70 as positive. 31 (17%) participants P2, P8 and P9 self-reported that they felt addicted to their phones.
mentioned the functions of their smartphone operating system with They reported spending 5.7 h/d on average on their phones. They
positive sentiments, such as: “AI assistant is so smart, I get used to operating mentioned the following symptoms: Time distortion (P2, P8), mild
my phone using voice”, “the Digital Health function gives me a clear idea depression (P9), low productivity in studies (P8, P9), blurring eyesight
about how much time I spend on my phone”, “I use phone memos to write lab and sore fingers (P2), and habits that could cause physical harm (P8 and
notes, it is so convenient”, and “Turning on NFC by double-clicking makes the P9 share a common habit of scrolling through their phones while
payment process easier, saving commute time”. It was observed that these walking). For participants who did not consider themselves addicted,
participants were satisfied with the utilisation and application of fore­ some symptoms were nevertheless reported: bad sleep quality associated
front technologies, and they accepted and appreciated the convenience with phone usage (P10), sometimes feeling anxious when scrolling
brought by smartphones. Many identified useful operating system through one’s phone (P1), cannot stop scrolling (P4, P6) and habitually
functions that had nothing to do with persuasive technology. unlocking one’s phone (all participants except P7).
The apps most mentioned as leading to new habits were WeChat (11 P8 and P10 expressed being worried about spending over 2 h daily on
times, about changing ways of socialising and making payments, WeChat to socialise with peers because of FOMO. P2, P3 and P5
refreshing the app to check updates, walking 10000 steps daily, etc.); expressed that their performance in their studies/internship had been
Toma Todo (6 times, about concentrating on learning); Alipay (4 times, less productive recently due to excessive use of smartphones:
about digital payment and feeding pets on virtual farms); Baidu (3 times, “I know that I spend too much time on my smartphone, it negatively af­
about the benefits of maps and search engine). All these comments were fects me. I cannot focus on studying and often drift away. Tried a few
either positive or neutral, except one participant mentioned that times to reduce screen time; however, I never succeeded.” (P2)
“WeChat has negative influences on my sleep time”. Survey participants
mentioned a set of apps that led them to adopt daily check-in habits: QQ “Playing with my smartphone causes me to delay the hand-in of assign­
(instant messaging), Taobao (e-commerce), Banking (finance), Alipay ments. When stress is high, it is more difficult to put aside my phone. This
(finance), Forest (time management), vocabulary apps and Xuexi Tong leads to a cycle of inefficiency and self-indulgence.” (P3)
(e-learning). “I was troubled by the notifications. I fear that I will miss something
Overall, short video, social networking, gaming, e-learning and time important if I do not read them. Some make me emotionally disturbed,
management apps were the most persuasive apps reported by partici­ which affect my study and productivity.” (P5)
pants. Habit changes brought about by built-in features of smartphones
were almost entirely described as positive, while the impact of apps on In order to gain a deeper understanding of the roles of smartphones
in interviewees’ daily lives, the interviewer asked in which situations

6
X. Chen et al. Computers in Human Behavior Reports 10 (2023) 100299

Table 3
Screen time and App usage of interviewees.
Participants Screen Time Necessary PD Comments Most used apps (h/d)
(h/d) (h/d)

P1 7.8 3.0 − 30% No intention to reduce smartphone usage. Smartphone is a useful tool. Weibo (2.0); WeChat (1.2);
TikTok (1.1); QQ (0.7)
P2 10.0 2.5 − 50% I blocked all notifications and try to reduce the frequency of checking my phone. WeChat (2.4); Videos (1.4);
TikTok (1.4); Gaming (1.1)
P3 4.0 2.0 − 50% I blocked notifications, recommendations, and deleted video apps to reduce my WeChat (1.6); Browser (0.7);
screen time. Music (0.5); Stock app (0.3)
P4 5.7 1.0 − 50% \ Douban (1.6); WeChat (1.2);
YouTube (0.8); TikTok (0.6)
P5 7.0 3.0 − 20% I traveled to another city with friends and do not have much time to play with Gaming (1.7); WeChat (1.5);
smartphone. Meituan (1.4); Weibo (0.4)
P6 9.1 6.0 − 25% Smartphone is my only digital device for assignments, lectures and entertainment. Videos (1.2); Pinduoduo (1.1);
Kuaishou (1.0); WeChat (0.8)
P7 6.1 5.0 − 25% I used iOS Screen Time to control the usage of certain apps. WeChat (2.4); Bilibili (1.5);
QQ Music (0.3); Taobao (0.3)
P8 8.0 5.0 − 65% I used multiple devices to reduce my frequency of checking my phone. All work- WeChat (2.8); Bilibili (1.1);
related tasks have been moved to laptop and iPad. Ciwei reading (0.5)
Xianyu (0.4)
P9 7.0 2.5 − 10% I deleted TikTok. Reading (2.6); WeChat (2.5);
Bilibili (0.8); QQ (0.7)
P10 6.0 5.0 − 40% Smartphone has been integrated into my daily life, makes tasks easier. WeChat (1.4);
Smartphone reduced my time using laptops. Gaming (1.2); TikTok (1.0);
Weibo (0.7)
Mean 7.1 3.5 − 37% – –

Notes: screen times and most used apps in this table were logged from screenshots provided by the interviewees.
PD: Persuasive Designs. This column is the user’s self-evaluation; if they could turn off all persuasive features, their screen time might change n%. ‘\‘: No record. ‘-‘: not
applicable.

interviewees must use their smartphones and with what functions. Ac­ Participants used phrases associated with emotions in their own re­
cording to these functionalities, a list of apps could then be identified flections on overusing smartphones: “bored”, “stress”, “relax”, “fun”,
and divided into six categories: “happy”, “isolated”.
Social networking QQ, WeChat, Weibo, Douban, Little Red Book; In summary, three out of ten interviewees self-reported that they
Shopping Taobao, Pinduoduo, JD, Alipay, Xianyu, Meituan were addicted to their smartphones. Physical harm, depression, FOMO
takeaway; and low productivity were reported to be associated with smartphone
Study/work DingTalk, university apps, Email, NFC commuting overuse. Interviewees reported apps that were essential to them,
card; covering many aspects of their life. Interviewees used emotion-related
Tools Vocabulary apps, maps, Forest, stocks and funds, banks, Toma expressions to reflect on their smartphone overuse.
Todo, Calendar;
Reading Zhihu, WeChat news subscription, Qidian online; 4.2.2. Perception of persuasive designs (RQ3)
Leisure Music apps, games, short video apps, streaming services. In this section, we describe how interviewees perceived persuasive
When the interviewees were asked to evaluate the time needed for designs and which persuasive designs interviewees identified on their
these essential use, this ranged from 1 h to 5 h. The mean time was 3.5 h, smartphones.
which is 58% of their self-reported total screen time (mean=6 h). The Interviewees were divided in terms of knowledge of persuasive
interviewees mentioned some of the factors that caused them to spend technology. On the one hand, two interviewees who were studying
prolonged hours on their smartphones: computer science (P3 and P7) and one interviewee with a media major
(P10) had learnt about persuasive technology in their previous studies.
“I feel bored when commuting, so I play with my smartphone like others;
They defined persuasive technology as “using automatic algorithms and
When I encounter difficulty in writing my bachelor’s thesis and my
notifications to persuade users” (P3), “applying psychological methods in ICT
internship tasks, I check social media and escape from all the stress during
products design” (P7), and “studying users’ preference to cultivate and
the break to relax.” (P5)
reinforce habits” (P10). On the other hand, the other seven interviewees
“My roommate and I compete on the Alipay virtual farm. It is a silly game; had no prior knowledge of persuasive technology before the interviews.
however, it is fun to have a routine game with a friend. Additionally, I To investigate their perceptions of persuasive technology, definitions,
play TikTok videos when I have meals; then, time flies without noticing it.” applications and examples of persuasive technology were discussed with
(P6) the interviewees to ensure that the interviewees understood what
persuasive technologies were and how they worked. After the discus­
“I know that using a smartphone for 6 hours daily is a bit too much. The
sion, the interviewer invited the interviewees to identify persuasive
entertainment provided by the smartphone is very convenient. Since I am
applications and features in the apps they use daily. The persuasive apps,
so happy when playing on the device, and making changes will be painful,
designs and features that the interviewees identified can be classified
why do I need to control my usage? If the purpose of life is to pursue
into the following categories:
happiness, smartphones can indeed fulfil my needs.” (P7)
Social networking Little Red Book integrates purchase links into its
“When I hang out with my friends during the weekends, I have less screen online community, making it easier to place orders from influencers’
time. However, when I spend weekends by myself, I feel isolated if I don’t posts. WeChat has subscriptions; QQ pushes notifications; while Weibo
refresh my social media feeds. Also, commuting between two campuses of and Douban recommend articles and ads based on the users’ viewing
the university takes 3 to 4 hours each week. I play with my phone on history and profiles. WeChat uses tags, such as “N friend(s) favourited”
public transportation.” (P9) and “N friend(s) read this article” to persuade users to click on the
recommended articles (see Appendix Fig. 4). All these social networking

7
X. Chen et al. Computers in Human Behavior Reports 10 (2023) 100299

companies incorporated “like” buttons in their apps. commerce apps recommend too many products to users based on their
Shopping Taobao, Pinduoduo, Xianyu, and JD recommend new purchase history and browser history. P3, P7, and P10 identified more
purchases based on users’ searching, browsing and typing history. persuasive designs than other interviewees. Concerns about user privacy
Meituan takeaway suggests menus to users based on their location, and data ownership were also raised by P7 and P10.
profile, and local weather forecast. Pinduoduo uses gambling-like de­ In summary, interviewees identified multiple persuasive designs in
signs5 to deceive users and gamification to attract users to spend more most of their essential daily apps, and they expressed quite positive
time on it (see Appendix Fig. 4). Alipay’s virtual farm uses incentives sentiments towards the functionalities of these apps; however, they
(virtual coins) and competition (rankings) to foster users’ daily check-in complained about the distraction, lack of consent, user privacy and
habits. Additionally, all these apps send coupons to stimulate new blurring data ownership related to persuasive designs. Besides the
purchases. common persuasion strategies, some ethically controversial ones were
Tools Time management apps such as Toma Todo and Forest have also reported by interviewees, that is, using emotional triggers and
persuasive reminders to help users manage their schedules. Keep and gambling-like designs to persuade users.
Mint have persuasive notifications to encourage users to exercise and eat
healthy diets. Vocabulary apps use personalised notifications and goals 4.2.3. Evaluation of persuasive designs (RQ3)
to remind users to check in daily. The Baidu map highlights restaurants In this section, we record how the interviewees evaluate the influ­
and shops that have paid promotion fees to make their locations more ence of persuasive designs on them.
visible. Interviewees indicated that persuasive designs prolonged their
Reading Top Buzz news and Zhihu recommend articles based on screen time, caused distractions, and affected their behaviours; specif­
users’ reading history. Ads are personalised according to users’ unique ically, P5, P7 and P10 mentioned disruptiveness, exploiting human
profiles, making the ads more attractive and relevant to users. Top Buzz, weakness and reinforcing habits:
WeChat subscription channel and Weibo trending news use emotionally
“Without these subtle persuasive designs, I would spend less time checking
triggering titles to lure users. "Smart" algorithms are applied to provide
my phone. When I was reading a paper or working on assignments, the
personalised suggestions.
reminders that a drama I followed had just updated would prompt me to
Leisure Short videos, user-generated content, and streaming plat­
watch the drama first.” (P5)
forms recommend new videos/playlists based on users’ viewing history
(i.e., WeChat video, TikTok, Kuaishou, iQIYI, Youku, Bilibili, YouTube). “All popular apps have persuasive designs integrated. Interaction de­
TikTok and Kuaishou integrate buying links with video content, signers took advantage of human weaknesses and created these apps. It
encouraging users to place orders with only one click. seems that only users are being blamed for their lack of self-control.” (P7)
OS The red dots on app and system icons draw users’ attention and
“Persuasive designs feed users according to their preferences, which could
keep persuading users to click on them (iOS and Android). Xiaomi and
make users stuck in their own habits knowing how it works does not make
Huawei phones show recommendations of readings and app promotions
me immune to the algorithms.” (P10).
after system updates. Participants were annoyed by those solicitations,
which were difficult/impossible to turn off. At the end of the interviews, interviewees were asked to spend 2
Proactive persuasive designs were identified in most essential apps minutes reflecting on their smartphone usage and evaluating the influ­
by interviewees, with the exception of a few tools (stocks & investments, ence of persuasive apps on them. All interviewees indicated that
calendar) and study/work apps (DingTalk, university, email and NFC persuasive applications increased their smartphone usage time. The in­
commute apps). Interviewees used negative expressions to describe their terviewees estimated that if they could turn off all persuasive features on
experience with intrusive persuasions that relied on personal informa­ their smartphones, they might reduce their screen time by 10%–65%,
tion and distracting notifications. On the other hand, interviewees used with a mean value of 37% (see Table 3).
positive expressions to discuss persuasive features of time management All interviewees shared their screen time screenshots with the
apps, Keep, Mint, and vocabulary apps. The interviewees used neutral interviewer during the interviews. The most frequently used apps by
expressions to describe their experience with persuasive designs, but interviewees were social networking apps (WeChat, Weibo, QQ), video
they were annoyed that they could not turn off some persuasive features platforms (Tencent, Youku, Bilibili), short video apps (TikTok,
that were imposed on them without their consent. Kuaishou), shopping apps (Taobao, Pinduoduo, Xianyu, Meituan),
All interviewees expressed positive sentiments towards their phones. reading apps (Ciwei, Qidian) and games. These apps accounted for more
This matches the analysis of the questionnaire’s open-ended answers. than half of the interviewees’ screen time. As we recorded in Section
However, users complained about the red dots on the operating system 4.2.2, nearly all these apps integrate multiple persuasive designs into
icons, which made users click reflexively, and intrusive AI recommen­ their services. We collected comments from the interviewees three
dations (for example, the AI assistant page of Harmony OS), which were weeks after the interviews regarding their participation in the study.
described as “harassing” and “manipulative” by participants. E- Five interviewees had reduced their screen time after the interviews, and
they mentioned blocking notifications, deleting addictive apps, travel­
ling with friends, setting time limits on addictive apps, and using
5
Gambling-like designs: Pinduoduo persuades users to invite their friends alternative devices to reduce the frequency of unlocking their smart­
to download or register with Pinduoduo to endorse them to get “free cash” or phones. For interviewees who did not seem to worry about their screen
“free products”. Pinduoduo’s algorithm calculates how much the invited friends time, their screen times remained nearly the same, and they mentioned
would endorse a user. In many cases, even if the user spends days inviting all his that their smartphone was “useful”, “integrated with everyday life” and
friends to sign up and log in to Pinduoduo, the promised money is never one participant’s “only digital device”.
transferred to the user. This persuasion strategy applies deception and matches Overall, the interviewees evaluated that persuasive designs pro­
the definition of gambling, i.e., “the practice of risking money or other stakes in longed their screen time, reinforced habits and caused distractions. In­
a game or bet” (merriam-webster.com). In these gambling-like designs, users terviewees had different evaluations of the impact of persuasive designs
lost their valuable time and their friends’ personal data; and most users gained
on their screen time, but all reported that persuasive designs increase
nothing at the end of the games. Lawyer Liu Yuhang sued Pinduoduo for
their screen times.
allegedly violating the principle of good faith, using false data, and concealing
rules, which constituted fraud, on March 31, 2021, in Shanghai, China. Even
though Pinduoduo is listed on Nasdaq, there is not much international coverage 5. Discussion
of this lawsuit, which was trending on Chinese social media sites for weeks in
April 2021. In this section, we reflect on the three research questions of this

8
X. Chen et al. Computers in Human Behavior Reports 10 (2023) 100299

paper. First, we discuss our results regarding problematic smartphone 5.2. Persuasive designs are making smartphones more addictive
usage (RQ1). Second, we discuss the results concerning perceptions,
identification and evaluation of persuasive designs (RQ2 and RQ3). We found persuasive designs to be more prevalent than we had ex­
Next, we discuss our findings regarding interaction design ethics. We pected. Persuasive designs were identified by interviewees in most daily
also discuss the limitations of the present work and suggest directions for apps, except for a few essential apps (see Section 4.2.2). Reviewing the
future research. apps that changed attitudes or behaviours from the questionnaire results
(Table 2), we found that all but a few learning apps integrated with
5.1. Problematic smartphone use: proportions, screen times and multiple persuasive designs. We can deduce from these findings that
reflections study participants live their everyday life with ambient persuasions.
Persuasive designs are often designed to exploit users for indirect
Throughout our study, similar proportions of participants reported monetisation by manipulating them to spend as much time as possible
multiple PSU behaviours (25% in the survey, 3 out of 10 in the in­ interacting with apps and services. For shopping apps, the most identi­
terviews). Many previous studies have used SAS-SV to measure the PSU fied persuasive designs were personalisation, reduction, suggestion and
of Chinese students, with varying results, ranging from 30% (Chen et al., rewards; for social networking apps, they were recognition, person­
2017) to 72% (Yang et al., 2019). Different sample methods and alisation, social comparison, and reminders; for leisure and reading, the
participant demographics might contribute to the differences in results. most identified were liking, suggestion, tracking, reduction, and moni­
The interviewees’ reported symptoms, such as time distortion, mild toring. These most identified strategies overlapped with Orji and Mof­
depression, low productivity, blurred eyesight, and wrist pains, were fatt’s analysis (Orji & Moffatt, 2018). However, they identified the most
also recorded by other PSU studies (Busch & McCarthy, 2021; Rozgon­ employed designs when analysing persuasive technologies for health
juk, Sindermann, Elhai, Christensen, & Montag, 2020). and wellness. In this study, these designs were found in social
There are multiple reasons why survey participants reported networking, shopping, leisure and reading apps, which seek more clicks
spending 5.6 h/d on average on their phones. Previous studies found and monetisation from users. The persuasive designs in these apps do
that correlation points between FOMO and PSU are much higher for not contribute to user well-being and are misaligned with user interests.
Chinese students compared to Western countries (Elhai, Yang, Fang, Persuasive triggers and reminders play crucial roles in cultivating
et al., 2020). The anxiety of the COVID-19 pandemic has been found to users’ habitual phone-checking behaviours. Both Fogg’s Persuasive
be associated with PSU severity (Elhai, Yang, McKay, & Asmundson, Design behaviour model (Fogg, 2009) and the PSD model emphasize the
2020). Besides these reasons reported in earlier work, we found that the role of triggers/reminders in increasing users’ performance of target
less developed IT infrastructure of Chinese universities (intranet and behaviours. In surveys and interviews, participants reported that a set of
emails), which resulted in the adoption of instant messaging apps (QQ apps fostered their daily check-in habits. This habitual behaviour was
and WeChat) as primary communication tools among students, admin­ triggered by phone vibrations, rings, reminders and flashing, which
istrators and educators, make students spend prolonged hours on their eventually led to users unlocking their phones unconsciously every
phones. 15–30 min, even without such triggers. Habitual checking is one of the
Social network and short video apps were the most mentioned apps symptoms of digital addiction and takes up a considerable amount of
in the questionnaire, occupying much of the interviewees’ screen time. users’ screen time daily (Almourad et al., 2020; Heitmayer & Lahlou,
Users tend to conflate smartphone usage and application usage when 2021).
self-reporting PSU (Rozgonjuk, Sindermann, Elhai, Christensen, & Some most frequently used PSD strategies, such as personalisation,
Montag, 2020). When investigating PSU, researchers should be aware of reduction and rewards, might deprive users of the opportunity to make
this problem and consider gathering information on both general independent decisions in long-term use. Before video platforms and
smartphone usage and application-specific usage to better understand shopping apps introduced algorithmic recommendations, users had
the relationship between the two. Additionally, different social network more time to autonomously explore different topics and products.
applications display varying degrees of addictive potential and impact Algorithmic recommendations have come to increasingly influence
on users (Rozgonjuk, Sindermann, Elhai, Christensen, & Montag, 2020), users’ decision-making, in many cases under the banner of convenience
underscoring the need to analyse the platforms users choose and their for users, but many interviewees in our study indicated that some
potential contribution to problematic behaviours (Rozgonjuk et al., smartphone apps powered by such algorithms “know” users to an un­
2021). comfortable degree. Ten survey participants complained that the videos
Interviewees’ reflections on necessary usage and overuse revealed recommended by TikTok were so addictive that they wasted “too much
that smartphones have become indispensable for study/work, social, time” and “lost control” (see Table 2).
leisure and finance. Societies globally are undergoing large-scale digital Besides popular persuasive design strategies, we found that some
transformations, moving both public service and private business online. companies use manipulative and deceptive strategies to “persuade”
It is almost impossible to live everyday life without a smartphone. users. One example is the WeChat display of FOMO tags on top of rec­
People would face social, study, mobility, and work difficulties without ommended articles to persuade users to spend more time on its services.
apps (see Section 4.2.1). The categories of necessary apps reported by On the one hand, Li et al. (2022) found that FOMO is positively asso­
interviewees can explain why 85% of survey participants spent at least 4 ciated with smartphone addiction. This repeated occurrence of the
h daily on their smartphones. People often cannot refuse to use smart­ FOMO tag might increase users’ level of FOMO, while higher levels of
phones when living in digital societies. In addition, we found that par­ FOMO are associated with a greater impact of social media on one’s
ticipants regard their phones as interactive narrative relays between daily activities and productivity, and particularly messenger and social
them and others (online strangers/peers/friends/family members). network use disorders (Przybylski et al., 2013; Rozgonjuk, Sindermann,
Smartphone users project their mental life through smartphones and Elhai, & Montag, 2020). On the other hand, people who use their
receive the projections of others. These internal and external mental smartphones frequently for social purposes form smartphone habits
projections interact with each other and have impacts on the user’s daily more quickly, which might lead to PSU (Van Deursen et al., 2015).
life. A kind of companionship is formed between people and their Another example is Pinduoduo, which applies gambling-like designs to
phones, as evidenced by the fact that many users use emotional words to exploit users’ time and social contacts. Gambling disorder has been
describe their overuse of smartphones. recognised as a behavioural addiction in DSM-5 (Edition, 2013).
Deceiving users with nontransparent rules, decoy rewards, and inter­
active algorithms can lead to distress and self-blame among users,
accustom them to gambling-like behaviours, and possibly lead them to

9
X. Chen et al. Computers in Human Behavior Reports 10 (2023) 100299

become addicted to Pinduoduo. participants we reached were relatively small in number. The survey
Based on the above empirical evidence, we believe that persuasive collected 183 valid results from Chinese university students; it is a small
designs influence the occurrence of PSU. When analysing PSU behav­ sample compared with the target group population size of 32.86 million
iours, it is crucial to consider persuasive designs. Interaction designers (Textor, 2022). More questionnaires with improved sampling methods
relying on such technologies need to consider the long-term impact of need to be distributed to study the group in more detail. Second, the
their products in terms of time spent, habit cultivation, over­ self-reported screen time of the participants in this study differed from
simplification of decision-making, behavioural addiction and human- their actual screen time, as evidenced by the actual screen times
computer relationships. (mean=7.1, source: screenshots of their phones) of ten interviewees,
which are longer than their self-reported screen times (mean=6.0,
5.3. Ethical implications for interaction designers source: questionnaires). Future studies need to examine the validity of
self-reported measures of smartphone use and develop improved tools
Persuasion in interactive computing systems is becoming increas­ for quantifying media use (Parry et al., 2021). Screen time tracking apps
ingly intelligent, subtle and influential. As Fogg pointed out, with could improve the accuracy of capturing usage, but would entail po­
interactive technologies, users receive information and respond imme­ tential invasions of user privacy. We only used three indicators, i.e., loss
diately. Such interactive behaviour is different from traditional media of control, perceived negativeness and overuse, to identify potential
(for example, TV and newspapers), which do not sustain interactive PSU, Therefore, our study’s findings should be interpreted with caution
looping behaviours. Persuasive technology has amplified such interac­ and further research is needed to validate the items used in our ques­
tive looping behaviours radically (Fogg, 2009). Persuasive technologies tionnaire. Third, our follow-up study to assess whether knowledge of
track and produce personalised reminders and suggestions on smart­ persuasive technologies would change users’ screen time is limited by
phones autonomously, instantly and persistently, and they are getting participants having varying levels of knowledge on how to manage their
better at it every day. screen time, since we did not standardise the information provided to
We put more emphasis on investigating the relationship between them if they asked questions at the end of the interviews. We are also
persuasive designs and PSU in our study; however, we documented that unable to make statements about the reasons why participants changed
nearly all participants also expressed positive sentiments towards the their screen time. Contextual factors might have played a role, for
built-in features of smartphones, learning, hobby and time management instance, sickness, family visits or holidays. Further research is needed
apps. Participants believed these apps improved their quality of life and to identify effective interventions for reducing PSU and improving dig­
well-being, which were the initial goals of persuasive technology pio­ ital well-being (Loid et al., 2020; Olson, Sandra, Chmoulevitch, et al.,
neers. According to the user screen time data we collected, users spend 2022). Fourth, we observed that the interviewees who had heard about
far less time using these persuasive technologies designed for user well- persuasive technology before the interviews identified more persuasive
being, while most of their time is occupied by persuasive designs, such as designs than others. Even after the discussion of the definitions and
those for monetisation, that compromise their well-being. examples, interviewees might need more time to understand and
There are no governmental or industry regulations on persuasive observe persuasive designs on their smartphones.
technologies to block deception and manipulation. We have observed We recommend that future research investigate the impact of
some apps applying persuasive designs to exploit users: first, the abusive persuasive designs longitudinally. Currently, there is relatively little
applications of tailoring and suggestions; for example, some trending research on the abusive application of persuasive technology in
articles on Zhihu and Weibo were actually paid promotions targeting commonly used apps and operating systems that consume much of
specific groups of users. Second, the overuse of reminders to seek users’ young adults’ screen time; there might also be other negative effects
attention and exploit users’ time, for example, the broad adoption of red besides the ones we discussed.
dots and notifications on app icons. Third, some algorithms take
advantage of users’ weaknesses; for example, the addictive algorithm of 6. Conclusion
TikTok troubled many study participants. Despite being troubled by
these persuasive features, users cannot turn off these persuasive We found that persuasive designs were perceived to prolong the
functions. screen time of the participants in our study and contributed to PSU.
There are studies on how to design persuasive technologies ethically, Participants reported that short video, social networking, gaming and
and approaches such as stakeholder analysis (Fogg, 2002), moral prin­ learning apps most affected their attitude and behaviour, and these apps
ciples (Berdichevsky & Neuenschwander, 1999), voluntariness assess­ were found to employ multiple persuasive designs and occupy much of
ment (Smids, 2012) and interdisciplinary research methods (Borgefalk participants’ screen time. The most frequently identified persuasive
& Leon, 2019) have been proposed by academia. Seven out of ten in­ design strategies were reminders, personalisation, reduction, rewards,
terviewees in our study had not heard of persuasive technology before recommendations and emotional motivators. These could have negative
the interviews, which might point toward a general lack of awareness of long-term impacts on users by prolonging their screen time, reinforcing
persuasive technology. Persuasive designs are often the default setting in phone-checking habits and oversimplifying decision-making. Some
operating systems and app installations, as far as we could observed in ethically controversial strategies (persuasion without consent, abusively
this study. We argue that one urgent ethical challenge interaction de­ applying emotional triggers and using gambling-like designs) have also
signers face is that most users are persuaded without their consent. Our been documented in our study. To answer the question we set out to
users found themselves being persuaded not with explicit arguments or answer – do persuasive designs make smartphones more addictive – we
reasons, but through the constant harvesting of their private data and find indications that persuasive designs might contribute to problematic
exploitation of their emotions. We see no other way to describe such and addiction-related behaviours. We recommend HCI researchers and
means but manipulative - in other words, as the kind of means Fogg designers, as well as psychologists, examine the long-term impact of
warned against in his inceptional and seminal work on persuasive persuasive designs and other similar designs on their users.
technology. Constant exposure to persuasive designs might lead to an
exhaustion of self-control, which might be addressed in future work.
Declaration of competing interest
5.4. Limitations and future work
The authors declare that they have no known competing financial
Our study has some limitations, and open questions for future work interests or personal relationships that could have appeared to influence
remain. First, we used the convenience sampling method, and the the work reported in this paper.

10
X. Chen et al. Computers in Human Behavior Reports 10 (2023) 100299

Data availability Acknowledgements

Data will be made available on request. Author 1 acknowledges financial support of the Institute for
Advanced Studies of the University of Luxembourg through an Young
Academic Grant (2021).

A. Survey results

The survey results (English translation) can be downloaded at the following link: http://doi.org/10.5281/zenodo.4934731.

B. Examples of persuasive designs

Fig. 4. Screenshots of persuasive designs

References Busch, P. A., & McCarthy, S. (2021). Antecedents and consequences of problematic
smartphone use: A systematic literature review of an emerging research area.
Computers in Human Behavior, 114, Article 106414.
eCorner, S. (2013). Justin rosenstein: No dislike button on facebook. Youtube. URL:
Cellina, F., Marzetti, G. V., & Gui, M. (2021). Self-selection and attrition biases in app-
https://www.youtube.com/watch?v=11WbGqALF_I.
based persuasive technologies for mobility behavior change: Evidence from a swiss
Adib, A., & Orji, R. (2021). A systematic review of persuasive strategies in mobile e-
case study. Computers in Human Behavior, 125, Article 106970.
commerce applications and their implementations. In International conference on
Cemiloglu, D., Naiseh, M., Catania, M., Oinas-Kukkonen, H., & Ali, R. (2021). The fine
persuasive technology (pp. 217–230). Springer.
line between persuasion and digital addiction. In International conference on
Almourad, M. B., McAlaney, J., Skinner, T., Pleya, M., & Ali, R. (2020). Defining digital
persuasive technology (pp. 289–307). Springer.
addiction: Key features from the literature. Psihologija, 53, 237–253.
Chen, B., Liu, F., Ding, S., Ying, X., Wang, L., & Wen, Y. (2017). Gender differences in
Berdichevsky, D., & Neuenschwander, E. (1999). Toward an ethics of persuasive
factors associated with smartphone addiction: A cross-sectional study among
technology. Communications of the ACM, 42, 51–58.
medical college students. BMC Psychiatry, 17, 1–9.
Billieux, J., Maurage, P., Lopez-Fernandez, O., Kuss, D. J., & Griffiths, M. D. (2015). Can
Clarke, V., Braun, V., & Hayfield, N. (2015). Thematic analysis. Qualitative psychology: A
disordered mobile phone use be considered a behavioral addiction? An update on
practical guide to research methods, 222 p. 248).
current evidence and a comprehensive model for future research. Current Addiction
Daniel, J. (2018). Our letter to the apa. URL: https://screentimenetwork.org/apa.
Reports, 2, 156–162.
Della Vedova, A. M., Covolo, L., Muscatelli, M., Loscalzo, Y., Giannini, M., & Gelatti, U.
Billieux, J., Schimmenti, A., Khazaal, Y., Maurage, P., & Heeren, A. (2015b). Are we
(2022). Psychological distress and problematic smartphone use: Two faces of the
overpathologizing everyday life? A tenable blueprint for behavioral addiction
same coin? Findings from a survey on young Italian adults. Computers in Human
research. Journal of behavioral addictions, 4, 119–123.
Behavior, 132, Article 107243.
Blease, C. (2015). Too many ‘friends,’too few ‘likes’? Evolutionary psychology and
Devincenzi, S., Kwecko, V., de Toledo, F. P., Mota, F. P., Casarin, J., & da Costa
‘facebook depression’. Review of General Psychology, 19, 1–13.
Botelho, S. S. (2017). Persuasive technology: Applications in education. In 2017 IEEE
Borgefalk, G., & Leon, N.d. (2019). The ethics of persuasive technologies in pervasive
frontiers in education conference (FIE) (pp. 1–7). IEEE.
industry platforms: The need for a robust management and governance framework.
Ding, X., Xu, J., Chen, G., & Xu, C. (2016). Beyond smartphone overuse: Identifying
In International conference on persuasive technology (pp. 156–167). Springer.
addictive mobile apps. In Proceedings of the 2016 CHI conference extended abstracts on
human factors in computing systems (pp. 2821–2828).

11
X. Chen et al. Computers in Human Behavior Reports 10 (2023) 100299

Edition, F. (2013). Diagnostic and statistical manual of mental disorders, 21 pp. 591–643). Marciano, L., Schulz, P. J., & Camerini, A. L. (2021). How smartphone use becomes
Am Psychiatric Assoc. problematic: Application of the alt-sr model to study the predicting role of
Elhai, J. D., Levine, J. C., Dvorak, R. D., & Hall, B. J. (2016). Fear of missing out, need for personality traits. Computers in Human Behavior, 119, Article 106731.
touch, anxiety and depression are related to problematic smartphone use. Computers Noë, B., Turner, L. D., Linden, D. E., Allen, S. M., Winkens, B., & Whitaker, R. M. (2019).
in Human Behavior, 63, 509–516. Identifying indicators of smartphone addiction through user-app interaction.
Elhai, J. D., Yang, H., Fang, J., Bai, X., & Hall, B. J. (2020). Depression and anxiety Computers in Human Behavior, 99, 56–65.
symptoms are related to problematic smartphone use severity in Chinese young Nyström, T., & Stibe, A. (2020). When persuasive technology gets dark?. In European,
adults: Fear of missing out as a mediator. Addictive Behaviors, 101, Article 105962. mediterranean, and middle eastern conference on information systems (pp. 331–345).
Elhai, J. D., Yang, H., McKay, D., & Asmundson, G. J. (2020). Covid-19 anxiety symptoms Springer.
associated with problematic smartphone use severity in Chinese adults. Journal of Oinas-Kukkonen, H., & Harjumaa, M. (2009). Persuasive systems design: Key issues,
Affective Disorders, 274, 576–582. process model, and system features. Communications of the Association for Information
Eyal, N. (2014). Hooked: How to build habit-forming products. Penguin. Systems, 24, 28.
Fogg, B. J. (1998). Persuasive computers: Perspectives and research directions. In Olson, J. A., Sandra, D. A., Chmoulevitch, D., Raz, A., & Veissière, S. P. (2022). A nudge-
Proceedings of the SIGCHI conference on Human factors in computing systems (pp. based intervention to reduce problematic smartphone use: Randomised controlled
225–232). trial. International Journal of Mental Health and Addiction, 1–23.
Fogg, B. J. (2002). Persuasive technology: Using computers to change what we think and Olson, J. A., Sandra, D. A., Colucci, É. S., Al Bikaii, A., Chmoulevitch, D., Nahas, J.,
do. Ubiquity, 2002, 2. Raz, A., & Veissière, S. P. (2022). Smartphone addiction is increasing across the
Fogg, B. J. (2009). A behavior model for persuasive design. In Proceedings of the 4th world: A meta-analysis of 24 countries. Computers in Human Behavior, 129, Article
international conference on persuasive technology (pp. 1–7). 107138.
Hamari, J., Koivisto, J., & Pakkanen, T. (2014). Do persuasive technologies persuade?-a Orji, R., & Moffatt, K. (2018). Persuasive technology for health and wellness: State-of-
review of empirical studies. In International conference on persuasive technology (pp. the-art and emerging trends. Health Informatics Journal, 24, 66–91.
118–136). Springer. Oulasvirta, A., Rattenbury, T., Ma, L., & Raita, E. (2012). Habits make smartphone use
Heitmayer, M., & Lahlou, S. (2021). Why are smartphones disruptive? An empirical study more pervasive. Personal and Ubiquitous Computing, 16, 105–114.
of smartphone use in real-life contexts. Computers in Human Behavior, 116, Article Oyibo, K., & Vassileva, J. (2019). The relationship between personality traits and
106637. susceptibility to social influence. Computers in Human Behavior, 98, 174–188.
Horwood, S., & Anglim, J. (2019). Problematic smartphone usage and subjective and Parry, D. A., Davidson, B. I., Sewall, C. J., Fisher, J. T., Mieczkowski, H., &
psychological well-being. Computers in Human Behavior, 97, 44–50. Quintana, D. S. (2021). A systematic review and meta-analysis of discrepancies
Huang, S., Lai, X., Xue, Y., Zhang, C., & Wang, Y. (2021). A network analysis of between logged and self-reported digital media use. Nature Human Behaviour, 5,
problematic smartphone use symptoms in a student sample. Journal of Behavioral 1535–1547.
Addictions, 9, 1032–1043. Przybylski, A. K., Murayama, K., DeHaan, C. R., & Gladwell, V. (2013). Motivational,
Kampik, T., Nieves, J. C., & Lindgren, H. (2018). Coercion and deception in persuasive emotional, and behavioral correlates of fear of missing out. Computers in Human
technologies. In 20th international trust workshop (co-located with AAMAS/IJCAI/ Behavior, 29, 1841–1848.
ECAI/ICML 2018) (pp. 38–49). Stockholm, Sweden: CEUR-WS. . (Accessed 14 July Rosenquist, N. J., Scott Morton, F. M., & Weinstein, S. (2021). Addictive technology and its
2018). implications for antitrust enforcement. Available at SSRN 3787822.
Kaptein, M. C., Markopoulos, P., De Ruyter, B., & Aarts, E. (2010). Persuasion in ambient Rozgonjuk, D., Levine, J. C., Hall, B. J., & Elhai, J. D. (2018). The association between
intelligence. Journal of Ambient Intelligence and Humanized Computing, 1, 43–56. problematic smartphone use, depression and anxiety symptom severity, and
Kaptein, M., Markopoulos, P., De Ruyter, B., & Aarts, E. (2015). Personalizing persuasive objectively measured smartphone use over one week. Computers in Human Behavior,
technologies: Explicit and implicit personalization using persuasion profiles. 87, 10–17.
International Journal of Human-Computer Studies, 77, 38–51. Rozgonjuk, D., Sindermann, C., Elhai, J. D., Christensen, A. P., & Montag, C. (2020).
Khalil, A., & Abdallah, S. (2013). Harnessing social dynamics through persuasive Associations between symptoms of problematic smartphone, facebook, whatsapp,
technology to promote healthier lifestyle. Computers in Human Behavior, 29, and instagram use: An item-level exploratory graph analysis perspective. Journal of
2674–2681. Behavioral Addictions, 9, 686–697.
Kim, S. E., Kim, J. W., & Jee, Y. S. (2015). Relationship between smartphone addiction Rozgonjuk, D., Sindermann, C., Elhai, J. D., & Montag, C. (2020). Fear of missing out
and physical activity in Chinese international students in korea. Journal of behavioral (fomo) and social media’s impact on daily-life and productivity at work: Do
addictions, 4, 200–205. whatsapp, facebook, instagram, and snapchat use disorders mediate that
Kwon, M., Kim, D. J., Cho, H., & Yang, S. (2013). The smartphone addiction scale: association? Addictive Behaviors, 110, Article 106487.
Development and validation of a short version for adolescents. PLoS One, 8, Article Rozgonjuk, D., Sindermann, C., Elhai, J. D., & Montag, C. (2021). Comparing
e83558. smartphone, whatsapp, facebook, instagram, and snapchat: Which platform elicits
Kwon, M., Lee, J. Y., Won, W. Y., Park, J. W., Min, J. A., Hahn, C., Gu, X., Choi, J. H., & the greatest use disorder symptoms? Cyberpsychology, Behavior, and Social
Kim, D. J. (2013b). Development and validation of a smartphone addiction scale Networking, 24, 129–134.
(sas). PLoS One, 8, Article e56936. Smids, J. (2012). The voluntariness of persuasive technology. In International conference
Lanette, S., Chua, P. K., Hayes, G., & Mazmanian, M. (2018). How much is’ too much’? on persuasive technology (pp. 123–132). Springer.
The role of a smartphone addiction narrative in individuals’ experience of use. Textor, C. (2022). Number of undergraduate students enrolled at public colleges and
Proceedings of the ACM on Human-Computer Interaction, 2, 1–22. universities in China from 2010 to 2021. URL: https://www.statista.com/statisti
Lei, L. Y. C., Ismail, M. A. A., Mohammad, J. A. M., & Yusoff, M. S. B. (2020). The cs/227028/number-of-students-at-universities-in-china/.
relationship of smartphone addiction with psychological distress and neuroticism Timmer, J., Kool, L., & Est, R.v. (2015). Ethical challenges in emerging applications of
among university medical students. BMC psychology, 8, 1–9. persuasive technology. In International conference on persuasive technology (pp.
Lewis, P. (2017). ‘our minds can be hijacked’: The tech insiders who fear a smartphone 196–201). Springer.
dystopia. The guardian, 6, 2017. Van Deursen, A. J., Bolle, C. L., Hegner, S. M., & Kommers, P. A. (2015). Modeling
Li, L., Niu, Z., Mei, S., & Griffiths, M. D. (2022). A network analysis approach to the habitual and addictive smartphone behavior: The role of smartphone usage types,
relationship between fear of missing out (fomo), smartphone addiction, and social emotional intelligence, social stress, self-regulation, age, and gender. Computers in
networking site use among a sample of Chinese university students. Computers in Human Behavior, 45, 411–420.
Human Behavior, 128, Article 107086. Waskom, M. L. (2021). Seaborn: Statistical data visualization. Journal of Open Source
Loid, K., Täht, K., & Rozgonjuk, D. (2020). Do pop-up notifications regarding smartphone Software, 6, 3021.
use decrease screen time, phone checking behavior, and self-reported problematic Williams, J. (2018). Stand out of our light: Freedom and resistance in the attention economy.
smartphone use? Evidence from a two-month experimental study. Computers in Cambridge University Press.
Human Behavior, 102, 22–30. Yang, Z., Asbury, K., & Griffiths, M. D. (2019). Do Chinese and british university students
use smartphones differently? A cross-cultural mixed methods study. International
Journal of Mental Health and Addiction, 17, 644–657.

12