Received: 25 July 2022 Accepted: 16 January 2023

DOI: 10.1002/jaba.975

RESEARCH ARTICLE

Reducing problematic social media use via a package intervention

Lesleigh Stinson | Jesse Dallery

Department of Psychology, University of Abstract

Florida, Gainesville, FL, United States
Problematic social media use can be characterized as that which interferes with
Correspondence
relationships, work, school, or sleep. Currently, there are no empirically
Lesleigh Stinson, University of Florida, 1908 supported treatments for reducing problematic social media use. We tested a
Stadium Road, Gainesville, FL 32611, USA. package intervention to reduce the daily duration of social media use measured
Email: [email protected]
by a smartphone application with nine undergraduate students who scored as
“addicted” to social media via a version of the Internet Addiction Test. The pack-
Editor-in-Chief: John Borrero
Handling Editor: Bethany Raiff age intervention included contingency management, automated notifications of
application use, and the selection of alternative activities. The package interven-
tion was effective at reducing the daily duration of social media use to goal levels,
or below, for all participants. Eight out of nine participants showed a decrease in
their Internet Addiction Test scores from pre- to postintervention, and overall,
participants did not show an increase in the time spent engaged in their selected
alternative activities. These findings demonstrate that social media use is amena-
ble to behavioral treatment.

KEYWORDS
contingency management, problematic social media use, social media addiction

The prevalence of social media use is increasing; the aver- media and technology (Panova & Carbonell, 2018;
age social media user increased their mean daily use from Varona et al., 2022), we have chosen to use the term
136 min per day on social networking sites in 2017 to problematic social media use throughout this paper.
146 mins in 2021 (Global Web Index, 2022). As of 2020, Prevalence estimates of problematic social media use
230 million Americans were active social media users vary, likely because it does not yet have a formal defini-
(Kemp, 2020). Social media can be defined as websites or tion. A 2014 study of undergraduate students at a univer-
applications that allow users to share information, ideas, sity in the northeast United States found that 9.7% of
videos, or other content electronically (Oxford University respondents engaged in problematic social media use,
Press, 2020). An individual’s social media use may become defined as endorsing three or more modified Diagnostic
problematic if it negatively interferes with other daily activi- and Statistical Manual of Mental Disorders-IV-TR cri-
ties such as work, school, sleep, or relationships (Chahal teria for alcohol abuse and dependence (American Psy-
et al., 2013; Owusu-Acheaw & Larson, 2015; Woods & chiatric Association, 2000; Hormes et al., 2014).
Scott, 2016; Zivnuska et al., 2019). Shensa et al. (2017) found that 44% of their sample of
Researchers have used different terms to discuss the young adults in the United States could be categorized
overuse of social networking sites (e.g., social media as problematic users, which they defined according to a
addiction, disordered social media use, problematic social modified Bergen Facebook Addiction Scale. A more
media use, and social media dependence). Although some recent study, also using a modified Bergen scale, found
of the differences between terms are due to distinctions in that only 3.4% of their sample of 18–25-year-olds in
theoretical conceptualizations (i.e., addiction versus the United States, recruited via an online survey com-
cognitive-behavioral paradigms; Varona et al., 2022), the pany, were considered to engage in problematic social
varying terms largely encompass the use of social media media use (Moreno et al., 2022). Other studies on
sites that result in negative consequences for the user, problematic social media use have used the Social
both interpersonally and psychologically. Given the dis- Media Disorder scale to diagnose problematic use
agreement over whether one can be addicted to social (Paakkari et al., 2021; Zendle & Bowden-Jones, 2019)

J Appl Behav Anal. 2023;56:323–335. wileyonlinelibrary.com/journal/jaba © 2023 Society for the Experimental Analysis of Behavior (SEAB). 323
 19383703, 2023, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jaba.975 by BALL STATE UNIVERSITY MUNCIE, Wiley Online Library on [03/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
324 STINSON AND DALLERY

or a modified version of the Internet Addiction Test media use as a target response (Stanley et al., 2022). Stan-
(Al-Menayes, 2015). ley et al. (2022) implemented a brief CM intervention for
Research has shown that the duration of social media smartphone use where participants were asked to reduce
use is associated with depression and other psychiatric their smartphone use by 30% of baseline with two thirds
diagnoses such as mood and personality disorders of that reduction being driven by a reduction in social
(Andreassen et al., 2014; Keles et al., 2020; Riehm media use. They found that, on average, participants’
et al., 2019; Rosen et al., 2013; Shannon et al., 2022). social media use was reduced in the CM treatment con-
Riehm et al. (2019) found that adolescents in the dition compared with baseline, although it was not
United States who spent more than 3 hr per day on social effective for all participants. The researchers analyzed
media had an increased likelihood of reporting high the behavior of the participants at the aggregate
levels of internalizing problems (e.g., feeling negative level. The current study extends the prior research by
emotions or experiencing sleep issues), which may targeting social media use as the primary behavior of
heighten the risk of mental health difficulties. Similarly, interest rather than as a component, focusing on
Levenson et al. (2016) found that increased daily social behavior at the level of the individual and collecting
media use was correlated with significantly greater odds social validity data.
of experiencing sleep disturbances in a sample of young For most target behaviors that have been effectively
adults in the United States. Furthermore, social media treated by CM, such as cigarette smoking or opioid use,
use can pose a safety hazard if individuals engage in it the goal is abstinence. However, with problematic social
while driving a vehicle. Evidence shows that social media media use, a reduction in the duration of use may be an
use while driving impairs a driver’s attention allocation acceptable goal. Reductions in duration, however, may
and reaction time to red lights and car stops (Hashash be more difficult to discriminate than abstinence for the
et al., 2019). Finally, a Finnish study found that individ- users. For example, it is likely easier for the person
uals who engaged in problematic social media use daily engaging in social media use to discriminate between
were more than 30% more likely to become victims of 10 min of social media use and no use versus 10 min of
cybercrimes than individuals who never engaged in prob- social media use and 8 min of social media use. There-
lematic use (Marttila et al., 2021). fore, we included application limits, which are notifica-
Contingency management (CM) is a behavioral tions that alert users when they have been using an
intervention that is effective for treating a variety of sub- application for a set amount of time. The rationale
stance use disorders (Prendergast et al., 2006) and for behind this inclusion was to provide a stimulus that
promoting a wide variety of health-related behaviors such would signal when the response has reached a certain
as weight loss, medication adherence, and exercise (Ellis duration and, potentially, the loss of a scheduled rein-
et al., 2021; Kurti & Dallery, 2013; Rosen et al., 2007). forcer for continued responding.
CM has also proven to be an effective treatment across a We also included a third intervention component
range of responses that are not topographically similar based on prior behavioral economic literature. The avail-
and are often maintained by multiple sources of reinforce- ability of alternative reinforcers has been posited by
ment (Burrows et al., 2020; Lewis & Lockmuller, 1990). behavior analysts and behavioral economists alike to
Generally, CM involves the delivery of incentives, often in reduce problematic behavior (Acuff et al., 2019;
the form of vouchers that are redeemable for goods or ser- Vollmer & Iwata, 1992). Hunt and Azrin (1973) devel-
vices, contingent upon meeting a prespecified behavioral oped the community reinforcement approach, a treat-
goal (e.g., abstinence). A primary component of CM is the ment for alcohol abuse in which individuals with alcohol
verification of the behavior by some objective measure. As problems contacted reinforcement through the commu-
social media use can be objectively measured through nity for engaging in behavior other than substance use.
smartphone and computer applications, it is a good candi- For instance, participants were provided with a venue to
date for treatment with CM. Objective verification of invite friends for alcohol-free recreational activities and
behavior is an important element of CM because, in its were provided with advice and assistance in securing a
absence, the contingency is placed on the self-report job. Subsequently, Higgins et al. (1991) combined the
response and may not adequately control the target community reinforcement approach with CM to effec-
response. Additionally, self-report may not be reliable. tively reduce cocaine use. More recently, Murphy, Skid-
Indeed, Verbeij et al. (2021) found correlations ranging more, et al. (2012) developed an intervention called the
from r = .49–.65 between self-report of social media use substance-free activity schedule, which was designed as a
and digital trace methods of measuring social media use. supplement to brief motivational interventions for alco-
Although such estimates may be acceptable in some situa- hol abuse. One component of the substance-free activity
tions, they are not accurate enough to determine a treat- schedule involves providing participants with feedback
ment’s effectiveness at reducing social media use. regarding activities other than substance use that the par-
Two recent studies assessed the effects of CM on ticipant may enjoy doing. Thus, we included a selection
smartphone use (Stanley et al., 2022; Williams-Buttari of alternative activities as an intervention component
et al., 2022), and one of those studies included social to increase the likelihood that participants would
 19383703, 2023, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jaba.975 by BALL STATE UNIVERSITY MUNCIE, Wiley Online Library on [03/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
REDUCING PROBLEMATIC SOCIAL MEDIA USE 325

readily have other activities, and therefore reinforcement, Social media use via smartphone was measured using
available in lieu of social media use. the Screen Time application. Screen Time is an applica-
The primary purpose of the present study was to assess tion for iPhone and Mac computers, built into the iOS soft-
a package intervention including CM, application limits, ware, that records the duration spent in an application. We
and the selection of alternative activities to reduce the daily independently verified the accuracy of the Screen Time
duration of social media use for young adults who engage measure (see Screen Time Validity Assessment).
in problematic levels of use. Young adults were included As a secondary outcome measure, we also collected
because they may be particularly vulnerable to problematic data on social media use via computer, when possible.
social media use and have higher rates of social media use We did this through the Screen Time application for Mac
than other age groups (Shannon et al., 2022; Shensa computers, and for the two participants who owned a
et al., 2017; Smith & Anderson, 2018). We also evaluated PC, we used Work Examiner. Work Examiner is com-
the acceptability of the intervention package and explored mercially available software that monitors computer
whether time spent engaged in alternative activities would activity on PCs and produces a variety of reports, includ-
increase with reductions in social media use. ing the duration spent on each application or website.
For privacy purposes, only participants had full access to
these reports and screen-shared relevant parts of them
METHOD with the researchers upon request.
Additional dependent variables included the self-
Participants, setting, and materials reported time spent engaged in alternative activities,
which was reported on a Behavior Allocation Survey.
We recruited nine participants via Facebook posts in sev- The Behavior Allocation Survey includes a list of com-
eral university-affiliated undergraduate Facebook groups. mon activities and asks participants to report daily dura-
The posts specified that the researchers were seeking partici- tions spent engaged in those activities (see Appendix).
pants who wanted to reduce their social media use. Partici- The Behavior Allocation Survey also contains an “other”
pants were undergraduate students aged 18–25 who owned section where participants were asked to report activities
an Internet-enabled iPhone and either a PC with a Win- that they engaged in that were not specifically listed and
dows operating system or a Mac computer with an operat- for how long they engaged in those activities.
ing system of Monterey or newer (to ensure compatibility We also evaluated scores pre- and postintervention on
with the software mentioned below). We assessed social a version of the Internet Addiction Test modified to
media use via both smartphone and computer because assess social media use. The Internet Addiction Test is a
research suggests that users most often access social media reliable and valid test developed to assess internet addic-
through smartphones (Dixon, 2022) and we were interested tion (Widyanto & McMurran, 2004; Young, 1998) that
in whether social media use via a computer would increase has been modified in previous literature to assess social
concurrently with decreases in social media use via smart- media addiction (Al-Menayes, 2015). It includes 20 ques-
phone. Further inclusion criteria were a score of 31 or tions with responses on a 5-point Likert-type scale
higher on a modified Internet Addiction Test (which indi- (1 = never to 5 = always). Total scores on the Internet
cates at least “mild addiction”; Young, 1998) and reliable Addiction Test range from 0–100, with scores ranging
access to the Internet for the duration of the study, meaning 31–49 being categorized as “mild” addiction, 50–79 as
that they anticipated having at least daily Internet access on “moderate” addiction, and 80–100 as “severe” addiction
their iPhone and computer. All procedures were approved (Young & Abreu, 2011). We modified each of the ques-
by the University of Florida’s Institutional Review Board, tions of the survey by replacing the terms “Internet,”
and all meetings were conducted virtually via Zoom. “online,” and “e-mail” with “social media.” Questions on
the modified Internet Addition Test related to the partici-
pants’ patterns of social media use. Some examples
Dependent variables and measurement include “Do you choose to spend more time on social
media over going out with others?” and “The last thing I
The primary dependent variable was the duration of do at night before I go to sleep is check social media.”
social media use via smartphone, defined as minutes per Last, we included an Intervention Acceptability Ques-
day spent on social networking sites on the participant’s tionnaire (Dallery et al., 2020) that asked participants
smartphone. To determine the daily social media use about their experience with specific aspects of the
duration, we summed the minutes spent on each social intervention.
networking application. We defined social networking
sites as any website or application where individuals can
post or send messages, images, and videos to others, Screen time validity assessment
excluding sites or applications primarily used for one-on-
one messaging. Text messaging, FaceTiming, and e-mailing To verify the validity of the Screen Time application, we
were not considered social media use. observed an individual (one of the authors) open various
 19383703, 2023, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jaba.975 by BALL STATE UNIVERSITY MUNCIE, Wiley Online Library on [03/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
326 STINSON AND DALLERY

social media applications in person, and an observer intervention. All other monetary rewards were added to a
timed the duration spent in each application with a stop- reloadable, prepaid debit card mailed or delivered to the
watch. We then compared the observer’s recorded dura- participant’s mailbox shortly after enrollment. Earnings
tion with the Screen Time reported durations over were added to the cards each Friday, although partici-
10 observations ranging from 9 s to 8 min 11 s. We found pants had the opportunity to request that their earnings
that Screen Time rounds the total time spent on an appli- be added to their cards at any time.
cation down to the minute. For example, if an applica-
tion was open for 5 min 45 s, Screen Time would round
down to 5 min. In this way, Screen Time is a close esti- Baseline
mation of application and website use. However, the one
condition in which we found Screen Time to be inaccu- During baseline, participants were asked to submit
rate was when a user has multiple applications open at screenshots of their daily social media use reports each
once on a computer. Screen Time continued recording morning for the prior day by text message. This was to
the time that an application was open, even if another ensure that all instances of social media use in a 24-hr
application was open in the foreground. This may lead period were recorded. If participants did not send in their
to overestimations of application or website use in this data by a certain time each day (preselected by each par-
specific scenario on a computer. ticipant in the enrollment meeting), we would send a
reminder text message. Participants earned $1 daily for
submitting screenshots of both their phone and computer
Experimental design social media use. We also asked participants to fill out
the Behavior Allocation Survey for the first and second
We used a concurrent multiple baseline design. Partici- days of baseline. We did this by sending the survey the
pants were assigned to one of three groups by order of following morning and asking participants to fill it out
enrollment, and each group of three went through the based on their social media use for the day prior. For
protocol simultaneously. For one group of participants, a participants to move into the intervention condition, their
withdrawal was also necessary to demonstrate adequate baseline data could not show a decreasing trend for the
experimental control. last three data points. However, due to the delay between
the initiation of the intervention condition (i.e., setting up
a meeting with participants) and the start of the interven-
Procedures tion condition (i.e., the implementation of contingencies
for social media use duration), this was not always
Enrollment the case.

Once the participants were deemed eligible, a one-on-one

enrollment meeting was held through video chat in which Intervention
we asked the participants to turn on the Screen Time
application on their phones. If participants owned a Mac The day before the start of the intervention, we held a
computer, we also asked them to turn on the Screen Time brief meeting with each participant where they decided
application for their computer. If participants owned a on a goal for duration of social media use. We did not
PC, we assisted them in downloading and activating suggest a fixed goal for all participants because of the
Work Examiner. Participants also filled out a demo- high variability in the baseline durations within and
graphic questionnaire that asked about their age, gender across participants. Additionally, to increase the social
identity, ethnicity, and race. validity of the intervention, we wanted participants to
If participants already had the Screen Time applica- determine the duration of social media use that they felt
tion turned on at the time of the enrollment meeting, we was appropriate for them. To inform their decision, par-
asked them to show us their Screen Time application via ticipants were provided with their mean baseline daily
video chat or send screenshots of their daily duration of duration of social media use and findings from relevant
social media use for the past 7 days. These data were used literature. Specifically, we reviewed the findings of
as a comparison with the baseline data to determine if Levenson et al. (2016) regarding daily social media use.
there were any reactivity effects due to researcher Levenson et al. concluded that a duration above 30 min
observation. was associated with increased odds of sleep disturbances,
and Riehm et al. (2019) found that adolescents with more
than 3 hr of daily social media use were at increased risk
Compensation of mental health problems.
We also asked participants to select three activities
Participants earned a $5 Amazon gift card for completion that they wanted to engage in as alternatives to social
of the surveys at baseline and at the end of the media use. This procedure was adapted from the
 19383703, 2023, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jaba.975 by BALL STATE UNIVERSITY MUNCIE, Wiley Online Library on [03/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
REDUCING PROBLEMATIC SOCIAL MEDIA USE 327

alternative activity component of the substance-free did not meet their goal but still submitted their screenshots,
activity schedule (Murphy, Skidmore, et al., 2012) used they earned $1. The contingency was resetting, meaning
to reduce drinking by college students. We asked partici- that if a participant missed their goal or did not submit their
pants to select three activities to maximize the likelihood data, they would return to the $7 earning level the following
that they would engage in those activities by having mul- day. This contingency was only applied to social media use
tiple options. We used the list of activities on the Behav- via smartphone; no consequences were programmed
ior Allocation Survey as a prompt but told participants for social media use via computer. This was because
they could choose any activity, even if it was not listed. most participants reported using social media on their
Once the participants selected three alternative activities, phones much more frequently than on their laptop
we discussed methods of promoting engagement in those computers. The earnings scale was determined by pre-
activities in their daily lives. For instance, if a participant liminary findings with pilot participants suggesting
selected reading, we discussed how it would be helpful to that $3 to $5 earnings were not effective incentives for
choose a book and carry it with them in their bag to reducing social media use and by prior CM research
increase the likelihood of reading (reading on a smart- treating substance abuse (Dallery et al., 2020; Packer
phone was considered an acceptable alternative activity if et al., 2012; Pfund et al., 2021).
it was not through a social media website or application). We also verified that participants kept the Screen
We discussed a potential strategy for each of the activities Time feature on by checking that the previous days’ data
that the participant selected. These conversations were remained in the screenshot. When Screen Time is turned
designed to be similar across participants and very brief, off, all prior data for the week disappears from the
lasting approximately 5 min. graph.
Next, we asked participants to set application limits At the end of the intervention, participants were
on their phones. On iPhones, application limits are set- asked to complete the Behavior Allocation Survey for the
tings that will send an alert to the user when a prespeci- last 2 days of the intervention. We also administered the
fied amount of time has passed using an application. Intervention Acceptability Questionnaire at the end of
When a user receives an application limit alert, they can the intervention period.
either select “OK,” which will close the application,
“Ignore for 1 min,” “Ignore for 15 min,” or “Ignore for
today,” any of which will allow them to continue using Withdrawal
the application. Once a user reached their limit on the
application, they continued to receive reminders that they Three participants experienced a withdrawal phase. The
have reached their limit each time they open the applica- day before we withdrew the intervention, we held a brief
tion until the next day. Although the application limits virtual meeting with these participants. We instructed the
were required, participants were allowed to choose which participants that they could continue earning $1 each day
specific social media applications they applied them to for sending in their screenshots but that they would no
and for how long, as long as the total duration was under longer be able to earn the larger amounts for meeting
their goal. For example, if, during baseline, a participant their goals beginning the following day. We also asked
spent 100 min per day on Instagram, spent 20 min per participants to send us screenshots verifying that they
day on Facebook, and set a duration goal of 90 min per had turned their application limits off at the start of the
day of social media use, they could choose to set an appli- withdrawal.
cation limit for 80 minutes on Instagram. Alternatively,
they could also set up application limits so that their
allowable time was split between applications, such as RESULTS
40 min on Instagram and 40 min on Facebook. We visu-
ally verified that each participant turned on the applica- Table 1 displays participant demographics and partici-
tion limits at the start of the intervention period. pant earnings. Participants were aged 18–22 and earned
Additionally, application limits are displayed on the between $87 and $192. Participants kept the Screen Time
screen that reports daily application use, so we ensured feature on for the duration of the study and submitted all
that they remained on for the duration of the screenshots, except for Participant 13, who did not sub-
intervention. mit one screenshot toward the end of the intervention.
For the first day of the intervention, participants Participants used a variety of social media platforms over
earned $7 for meeting their goal duration of social media the course of the Study.1
use. Earnings increased by $1 per day for each consecu- Figures 1 and 2 show the daily duration of social
tive day a participant met their goal, up to $9 per day. media use via smartphone for all participants. During
Participants also earned $1 for submitting screenshots
regardless of the daily duration of social media use. If a 1
Social media platforms used by participants include Discord, Facebook,
participant met their goal and submitted their screen- Facebook Messenger, Hinge, Instagram, LinkedIn, Pinterest, Reddit, Snapchat,
shots, they could earn between $8 and $10 daily. If they TikTok, Tinder, Twitter, and YouTube.
 19383703, 2023, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jaba.975 by BALL STATE UNIVERSITY MUNCIE, Wiley Online Library on [03/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
328 STINSON AND DALLERY

TABLE 1 Demographic information and total earnings for each participant

ID Age Gender Race, Ethnicity Earnings Goal duration SMU

P2 20 Female White, not Hispanic/Latino $155 70 min, 50 min

P5 21 Female White, not Hispanic/Latino $139 75 min,
P6 21 Female White, not Hispanic/Latino $192 50 min
P11 20 Other: not specified Middle Eastern, not Hispanic/Latino $123 45 min
P12 19 Female White, Hispanic/Latino $107 150 min
P13 18 Male Vietnamese $87 100 min
P14 20 Female White, not Hispanic/Latino $88 150 min
P15 22 Female White, not Hispanic/Latino $130 60 min
P16 19 Female White, not Hispanic/Latino $115 200 min
Note. ID = Participant identification number, SMU = Social media use. Participant 2 has two goal durations because she further reduced her goal duration in the second
intervention phase. Participant 11 self-identified as Middle Eastern when asked about their race.

Pre-BL BL App Limits BL App Limits per day and 143.1 min per day for prebaseline and base-
150 + CM + CM line conditions, respectively. When the intervention was
initially implemented, daily duration decreased to at or
100 below goal levels for all participants, and the group
means of daily social media use equaled 65.2 min. For
eight of nine participants, the daily duration of social
50 media use remained at or below the goal levels for almost
all days of the intervention. However, after 5 days in the
intervention condition, Participant 13’s daily duration of
0
P6 social media use returned to baseline levels. Participant
13 failed to submit the second-to-last screenshot, which is
Duration Social Media Use (min)

300 why the data path for Participant 13’s data is not
connected at the end of the intervention condition.
We chose to implement a withdrawal for the first
200 three participants because of the notable overlap between
some of the data points in baseline and the application
limits plus CM condition (intervention) for Partici-
100
pants 6 and 2. When the intervention was withdrawn,
the daily duration of social media use via smartphone
0 increased although the latency to change was slower
P2
for Participant 5 than it was for the other two partici-
250 pants. When the intervention was reimplemented, the
daily duration of social media use returned to below
200 the goal levels for all three participants. Notably, Par-
150 ticipant 2 decreased her daily goal by 20 min (from
70 to 50 min) when the intervention was reimplemen-
100 ted. We did not feel that the number of overlapping
data points for the other cohorts was sufficient to war-
50
rant an additional withdrawal phase and instead con-
0 tinued with the planned multiple baseline across
P5
subjects design for the final two cohorts.
0 10 20 30 40 Some participants’ data showed patterns that are
Day worth noting. Participants 2 and 6 reported that all or
part of their initial baseline data were unusual because of
F I G U R E 1 Daily duration of social media use via smartphone for
the company staying with them (Participant 6) or visiting
participants 6, 2, and 5; BL = Baseline, App = Application,
CM = Contingency management. The dotted horizontal lines indicate a partner for a weekend (Participant 2), so they did not
the goal duration of social media use. want to set a lower daily goal for the intervention
period than what they selected. Participant 11’s prebaseline
prebaseline and baseline conditions, the daily duration of data were notably lower than their baseline data. When
social media use was variable for all participants. The asked about this at the end of the study, the participant
group means of daily social media use were 199.9 min reported that after the intake meeting, they redownloaded
 19383703, 2023, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jaba.975 by BALL STATE UNIVERSITY MUNCIE, Wiley Online Library on [03/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
REDUCING PROBLEMATIC SOCIAL MEDIA USE 329

Pre-BL BL App Limits + CM Pre-BL BL App Limits + CM

300 150

200 100

100 50

0 0
P13 P15
Duration Social Media Use (min)

150 600

100 400

50 200

0 0 P16
P11

400 500

400
300
300
200
200
100
100

0 0
P12 P14
0 10 20 30 0 5 10 15 20 25
Day Day

F I G U R E 2 Daily duration of social media use via smartphone for participants 11–16; BL = baseline, App = application, CM = contingency
management. The dotted horizontal lines indicate the goal duration of social media use.

several social media applications they had previously deleted Pre-BL and BL
to “truly see if the intervention would be effective” for them, Intervention
100
Percentage of Days At or Below Goal

which resulted in an increase of daily social media use during

baseline.
We also collected social media use via computer data
for seven of the nine participants. For participants 2, 5,
6, 11, 12, and 16, social media use via computer was 50

infrequent and did not change systematically across

phases. The median value for social media use via com-
puter in each phase was 0 for these six participants. For
Participant 13, social media use via computer was high 0
and variable across all phases (Mdn baseline = 320 min P2 P5 P6 P11 P12 P13 P14 P15 P16
per day, Mdn intervention = 359 min per day; these data
are not included in Figure 1). After the intervention, F I G U R E 3 Percentage of days at or below the goal duration of
social media use; Pre-BL = prebaseline, BL = baseline.
when asked about his social media use via computer, Par-
ticipant 13 reported that he regularly left Discord, a chat
application, open in the background of his computer
throughout the day while he engaged in other activities. Figure 3 shows the percentage of days below the goal
He self-reported that he actively used Discord for about duration for each participant. The prebaseline and base-
10–20 min per day. line percentages of days below the goal duration are
 19383703, 2023, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jaba.975 by BALL STATE UNIVERSITY MUNCIE, Wiley Online Library on [03/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
330 STINSON AND DALLERY

T A B L E 2 Mean difference scores in minutes between baseline and 10

end of intervention BAS surveys
ID SMU Selected alternative activities

P2 45 25

Endorsement
P5 82.5 33.3 5
P6 32.5 5
P11 35 13.8
P12 250.5 16.5
P13 -- --
0
P14 90.5 54.2

Pr ers w S ed
on ec ac s
w om tion

U er d

on lp
ie o l e

N O re s
t M Le info Ne g
t A Ea tan ility

pa to Sat Go e

et A ful Xs

or arn le eds

ac a lls

rn
ou In ne y
ec ila ss

te ed
ris R isf al

c it
th en

n in
ng s
P15 80 3.3

C e
ch se dab

ce
iv on ki
sp va ne
ifi bil
vi f U

P e n
se T

ew w st

y lH
s ib

e N ar
/o m
er ex
nd Fl
P16 48.5 18.75

t
U
Note. BAS = Behavior Allocation Survey, ID = Identification number,

ab
s

M
om s
la

C lin e

n
SMU = Social media use; Positive scores indicate more time was spent engaged in

fu

ar

an
ke
p

Le
el

W
Li
the activity during the intervention phase than in baseline. P13 did not complete

H
any end of intervention surveys.
F I G U R E 5 Individual participants’ endorsement for each item on
the Intervention Acceptability Questionnaire. Each dot represents one
80
Preintervention Score individual’s response. The horizontal lines represent the mean response
Postintervention Score for that questionnaire item.

60
Modified IAT Score

All participants for whom the Behavior Allocation

40 Survey data were available reported a decrease in social
media use, but no participant reported an increase in time
20 spent engaged in all three of their selected activities. For
example, Participant 6 reported a 5-min increase in their
selected activities from baseline to the end of the inter-
0
P2 P5 P6 P11 P12 P13 P14 P15 P16
vention, but this increase was accounted for solely by an
increase in exercise.
F I G U R E 4 Modified Internet Addiction Test scores pre- and Figure 4 displays the changes in the Internet Addic-
postintervention; IAT = Internet Addiction Test. The dotted horizontal tion Test scores between baseline and intervention for all
lines indicate the scoring categories (0–30: normal use; 31–49: mild participants. For all participants, there was a decrease in
addiction; 50–79: moderate addiction; and 80–100: severe addiction).
the Internet Addiction Test scores between pre- and post-
intervention, except for Participant 13 who did not
combined in this figure because neither condition respond to follow-up surveys. Six participants moved
included programmed consequences for social media use. from the “moderate addiction” category to the “mild”
For all nine participants, there was a marked increase in addiction category (Participants 2, 5, 11, 12, 14, and 15),
days meeting the goal duration in the intervention phase one participant moved from the “moderate addiction”
compared with the prebaseline and baseline phases. category to the “normal” use category (Participant 16),
Schoolwork, reading, and exercising were the most and one participant moved from the “mild addiction”
selected alternative activities. Table 2 depicts the mean category to the “normal” use category (Participant 6). No
difference scores between the baseline and intervention participants scored in the range of “severe addiction.” We
Behavior Allocation Survey reports for all participants, also assessed whether there was a correlation between
except for P13 who did not complete any postinterven- Internet Addiction Test scores and the daily duration of
tion surveys. To calculate the difference scores, we sub- social media use. There was a moderate correlation
tracted the values from the first time the Behavior (r = .63, p > .006) between the two variables.
Allocation Survey was administered in the intervention Figure 5 shows the responses to the intervention
phase from the first time the Behavior Allocation Survey acceptability questionnaire. For all questions except “pri-
was administered during the baseline phase and then vacy concern” and “wanted more personal help,” a score
repeated this process using the second time the survey above 5 is considered favorable. Those two questions are
was taken in both phases. Then, we averaged these reverse scored, meaning that a response below 5 is con-
scores. A positive score indicates that the participant sidered favorable. Individual responses were generally
engaged in the activities for a longer duration in the favorable; mean responses for questions that were not
intervention phase than they did in the baseline phase. reverse scored were above 5, except for “learn new skills.”
 19383703, 2023, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jaba.975 by BALL STATE UNIVERSITY MUNCIE, Wiley Online Library on [03/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
REDUCING PROBLEMATIC SOCIAL MEDIA USE 331

For the two questions that were reverse scored, mean use was highly variable for most participants. These
responses were less than 3. findings are consistent with the results of Williams-
Buttari et al. (2022), who conducted a similar study with
CM for smartphone use. In the current study, a visual
DISCUSSION analysis of the daily duration by day of the week did not
reveal any consistent patterns (e.g., based on days of the
The current study assessed the effects of a package inter- week). However, for the first cohort of participants, Days
vention, involving contingency management, application 29 and 30 corresponded with the last weekend of term
limits, and selection of alternative activities on the daily before summer vacation and all three of the participants
duration of social media use via smartphone. The results displayed a shorter duration of social media use on those
show that the intervention was effective at reducing the days. As students may be more likely to engage in alter-
daily duration of social media use via smartphone for all native activities at this time (e.g., packing up their dorms
participants, although this effect was short-lived for one or saying goodbye to friends), these data provide support
participant (P13). These findings are generally consistent for the idea that social media use may be affected by
with those of Stanley et al. (2022), who found that the alternative activities.
daily duration of social media use, on average, decreased The Screen Time application was accurate when used
with the implementation of CM. Additionally, there was via smartphone. Although it rounds to the lowest minute,
not a concurrent increase in social media use via com- Screen Time was still sensitive enough to detect socially
puter. Overall, the intervention resulted in 91.8% of days significant changes in social media use. However, it was
below the target goal, compared with 27.4% during pre- not as accurate when used via computer. The limitation
baseline and baseline. The current study is the first to of Screen Time continuing to record the duration when
empirically demonstrate the efficacy of an intervention another window is open in the foreground of a computer
for problematic social media use. Moreover, the present was an issue that Participant 13 reported happening. In
study adds social media use to the growing number of the future, researchers should seek to find a more accu-
problematic responses that have been effectively treated rate measure of application use for Mac computers
by monetary- or voucher-based CM (Davis et al., 2016; (although future software updates may resolve this issue).
Drebing et al., 2007; Kurti & Dallery, 2013; Raiff We used application limits to aid in the discrimina-
et al., 2016; Washington et al., 2014). Not only did all tion of the daily duration of social media use and thus to
participants experience a reduction in the daily duration stay below the goal for CM. We did not collect data on
of social media use, but eight out of nine participants also whether participants received notifications (i.e., went
scored lower on the modified Internet Addiction Test over the limits). Thus, we cannot address the relative
after the intervention than before and scored in a less effectiveness of the application limits in reducing daily
severe addiction category. These data support the social social media use. Future studies may include a compo-
validity of the intervention in that the participants nent analysis to determine which intervention compo-
reported being less addicted to social media. However, it nents are necessary to produce the desired treatment
is not clear whether the numerical categories that quan- effects.
tify these addiction levels are arbitrary or whether they Notably, we did not observe a general increase in self-
are anchored to any clinical or behavioral outcomes. The reported time spent engaging in the selected alternative
moderate correlation between the Internet Addiction Test activities during the intervention. Although some partici-
scores and social media use daily duration suggests that pants showed an increase in duration engaged in the
researchers should use caution when relying solely on selected activities on average, no participant showed an
self-reported, subjective measures of social media use and increase in all three of their selected activities and some
should instead use objective measures when possible. participants did not report an increase in any of their
The social validity of the intervention is also sup- selected activities. Although it may not be necessary for
ported by participants endorsing most of the items on the participants to engage in all three of their selected activi-
Intervention Acceptability Questionnaire. The lowest- ties at an increased duration to see effects on daily social
rated items (other than the two items for which a “low media use, the fact that the intervention was effective for
score” is considered favorable) were “learn new informa- two participants who did not report an increase in any of
tion” and “learn new skills.” Although it is not clear what their selected activities (P2 and P12) indicates that this
information or skills should be learned in the context of component may not have played a role in the efficacy of
an intervention for social media use, it may be useful to the intervention. Alternatively, it could indicate that the
explore these in future interventions. For example, it may participants may have increased their engagement in a
be that more training around arranging the environment nontargeted activity. More research is needed on this inter-
to promote engagement in alternative activities may vention component to determine its relative contribution,
result in reduced social media use. especially given the findings of other researchers using simi-
The variability in the participants’ data warrants lar methods. Specifically, researchers in the field of alcohol
additional discussion. The daily duration of social media misuse found that the substance-free activity schedule,
 19383703, 2023, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jaba.975 by BALL STATE UNIVERSITY MUNCIE, Wiley Online Library on [03/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
332 STINSON AND DALLERY

in conjunction with a brief motivational intervention, can intervention to plan a meeting, the contact may have
reduce alcohol consumption and alcohol-related problems altered their social media use. Similar patterns have been
(Murphy, Dennhardt, et al., 2012; Murphy, Skidmore, observed in other studies implementing contingency man-
et al., 2012). In the present study, we only asked participants agement interventions (e.g., Participants 2 and 12 in
to choose alternative activities. In the alcohol research, Dallery et al., 2020). Last, Participant 14, 15, and 16’s
researchers prompted participants to think about their career prebaseline data were slightly higher, on average, than
and personal goals and how their current substance use their baseline data. This may reflect a reactivity effect.
could interfere with the achievement of those goals. As a part However, because it did not occur with participants in
of the substance-free activity schedule, participants also other cohorts, the higher prebaseline values are more
received personalized suggestions for leisure activities that likely an artifact of the data being collected during the
they may enjoy. It is possible that these additional elements final week of winter break before classes began.
included in the alcohol use interventions could have Future research may also explore methods for promot-
enhanced the effects of selecting alternative activities. ing maintenance for the lowered daily duration of social
It is worth pointing out that not all social media use is media use. When the intervention was withdrawn for the
harmful; moderate social media use may be beneficial. For first cohort of participants, the daily duration of social
instance, research has shown that exposure to social media media use returned to baseline levels for all three partici-
may help adolescents develop empathy and enhance their pants. In a similar study, Stanley et al. (2022) found that
self-esteem (Guinta & John, 2018; Vossen & Valkenburg, the effects of CM on social media use did not maintain, on
2016). This is not dissimilar from some addictive commodi- average, for participants three weeks after the intervention
ties, for which moderate use may be acceptable and not dis- was withdrawn. These data suggest that procedures should
ruptive for an individual (e.g., alcohol, cannabis). This is, in be implemented specifically to target response maintenance.
part, why our intervention was designed to reduce, and not For instance, Silverman et al. (2019) implemented a proce-
eliminate, social media use. dure in which samples were collected at increasing intervals
A limitation of this study is the small number of male contingent on the samples meeting the incentive criteria.
participants. This could reflect the recruitment methods, These procedures were effective at promoting medication
or it could reflect a self-selection bias as participants adherence for individuals with HIV for up to 2 years
volunteered to participate. Several studies have found (Silverman et al., 2019) and effective at maintaining smok-
that women are more likely to engage in problematic ing abstinence for up to 6 weeks (Dallery et al., 2020).
social media use than males (Andreassen et al., 2012, These outcomes across different populations and responses
2014; McAndrew & Jeong, 2012), so it is possible that suggest that these procedures may also be effective in main-
our sample accurately represented the demographics of taining a lower daily duration of social media use.
individuals who engage in problematic social media use. Finally, the implementation of CM can be quite
A final limitation of this study was that, although the costly. Future research is needed to continue to explore
screenshots of social media use durations were time and test treatments for problematic social media use that
stamped, it is possible that participants could have sub- may be more cost effective and, therefore, accessible to
mitted someone else’s data. However, two features of this more individuals. Deposit contracts, such as those used
study may have decreased the likelihood of this occur- in Williams-Buttari et al. (2022), could be used to miti-
rence. First, application limits were visible in screenshots gate the costs of CM, as well as prize-based CM in which
if the participant used the application that they were set an intermittent schedule of reinforcement is implemented
for, meaning that if a participant was using someone based on goal completion (Petry et al., 2004).
else’s data, that person would have also had to set the The current study provides support for the use of CM
same application limits as the participant. Second, our in conjunction with application limits to treat problem-
advertisements for the study specified that we were atic social media use, although more research is needed
recruiting individuals who were motivated to reduce their to determine whether all intervention components are
daily social media use. Although neither of these proce- necessary and to incorporate methods that support
dures is infallible in preventing participants from using response maintenance. It is the first empirical demonstra-
someone else’s screenshots, they may have reduced the tion of an intervention for problematic social media use
likelihood that it would occur. per se and adds to the growing number of problematic
The data from the final cohort of participants behaviors that researchers have shown to be effectively
(Participants 14, 15, and 16) warrant additional discus- treated by CM. These findings suggest that problematic
sion. Participants 15 and 16 displayed a drop in the daily social media use is a set of responses that may be profit-
duration of social media use for the 2 days before the ably treated as operant behavior.
intervention. All participants were contacted 1 to 2 days
prior to the intervention to set up a time to meet regard- A C K N O W L E D G M EN T S
ing the start of the intervention. Because both partici- Lesleigh Stinson is now at the Department of Health
pants were contacted 2 days before the start of the Education and Behavior, University of Florida.
 19383703, 2023, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jaba.975 by BALL STATE UNIVERSITY MUNCIE, Wiley Online Library on [03/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
REDUCING PROBLEMATIC SOCIAL MEDIA USE 333

F U N D I NG I NF O R M A T IO N systematic review and meta-analysis of randomized controlled tri-

Experiment.com; B.F. Skinner Foundation. als. Behaviour Research and Therapy, 136, Article 103781. https://
doi.org/10.1016/j.brat.2020.103781
Global Web Index. (2022). The biggest social media trends for 2022.
C ON F L I CT O F I NT E R E ST STA T E M E N T https://www.gwi.com/reports/social
The authors declare no conflict of interest. Guinta, M. R., & John, R. M. (2018). Social media and adolescent
health. Pediatric Nursing, 44(4), 196–201.
E T H IC S A P P RO V AL Hashash, M., Zeid, M. A., & Moacdieh, N. M. (2019). Social media
browsing while driving: Effects on driver performance and atten-
All procedures were approved by the University of Flori- tion allocation. Transportation and Research Part F: Traffic Psy-
da’s Institutional Review Board, and all meetings were chology and Behaviour, 63, 67–82. https://doi.org/10.1016/j.trf.
conducted virtually via Zoom. 2019.03.021
Higgins, S. T., Delaney, D. D., Budney, A. J., Bickel, W. K.,
ORCID Hughes, J. R., Foerg, F., & Fenwick, J. W. (1991). A behavioral
Lesleigh Stinson https://orcid.org/0000-0002-5127-5763 approach to achieving initial cocaine abstinence. The American
Journal of Psychiatry, 148(9), 1218–1224. https://doi.org/10.1176/
Jesse Dallery https://orcid.org/0000-0002-2882-1105 ajp.148.9.1218
Hormes, J. M., Kearnes, B., & Timko, C. A. (2014). Craving Facebook?
R E FE RE NC E S Behavioral addiction to online social networking and its associa-
Acuff, S. F., Dennhardt, A. A., Correia, C. J., & Murphy, J. G. (2019). tion with emotion regulation deficits. Addiction, 109(12), 2079–
Measurement of substance-free reinforcement in addiction: A sys- 2088. https://doi.org/10.1111/add.12713
tematic review. Clinical Psychology Review, 70, 79–90. https://doi. Hunt, G. M., & Azrin, N. H. (1973). A community-reinforcement
org/10.1016/j.cpr.2019.04.003 approach to alcoholism. Behaviour Research and Therapy, 11(1),
Al-Menayes, J. (2015). Dimensions of social media addiction among 91–104. https://doi.org/10.1016/0005-7967(73)90072-7
university students in Kuwait. Psychology and Behavioral Sciences, Keles, B., McCrae, N., & Grealish, A. (2020). A systematic review: The
4(1), 23–28. https://doi.org/10.11648/j.pbs.20150401.14 influence of social media on depression, anxiety and psychological
American Psychiatric Association. (2000). Diagnostic and Statistical distress in adolescents. International Journal of Adolescence and
Manual of Mental Disorders (4th ed.). revised text. https://doi.org/ Youth, 25(1), 79–93. https://doi.org/10.1080/02673843.2019.
10.1002/9780470479216.corpsy0271 1590851
Andreassen, C. S., Torsheim, T., Brunborg, G. S., & Pallesen, S. (2012). Kemp, S. (2020, February 11). Digital 2020: The United States of
Development of a Facebook Addiction Scale. Psychological America—DataReportal—global digital insights. DataReportal.
Reports, 110(2), 501–517. https://doi.org/10.2466/02.09.18.PR0. https://datareportal.com/reports/digital-2020-united-states-of-
110.2.501-517 america
Andreassen, C. S., Torsheim, T., Brunborg, G. S., & Pallesen, S. (2014). Kurti, A. N., & Dallery, J. (2013). Internet-based contingency manage-
Predictors of use of social network sites at work—a specific type of ment increases walking in sedentary adults. Journal of Applied
cyberloafing. Journals of Computer-Mediated Communication, Behavior Analysis, 46(3), 568–581. https://doi.org/10.1002/jaba.58
19(4), 906–921. https://doi.org/10.1111/jcc4.12085 Levenson, J. C., Shensa, A., Sidani, J. E., Colditz, J. B., &
Burrows, C., Dallery, J., Kim, S. J., & Raiff, B. R. (2020). Validity of a Primack, B. A. (2016). The association between social media use
functional assessment for smoking treatment recommendations and sleep disturbance among young adults. Preventive Medicine,
questionnaire. The Psychological Record, 70(2), 215–226. https:// 85, 36–41. https://doi.org/10.1016/j.ypmed.2016.01.001
doi.org/10.1007/s40732-020-00375-5 Lewis, M. J., & Lockmuller, J. C. (1990). Alcohol reinforcement: Com-
Chahal, H., Fung, C., Kuhle, S., & Veugelers, P. J. (2013). Availability plex determinant of drinking. Alcohol Health & Research World,
and night-time use of electronic entertainment and communication 14(2), 98–104. https://link.gale.com/apps/doc/A9858781/AONE?
devices are associated with short sleep duration and obesity among u=tall22798&sid=googleScholar&xid=bf5a11ad
Canadian children. Pediatric Obesity, 8(1), 42–51. https://doi.org/ Marttila, E., Koivula, A., & Räsänen, P. (2021). Cybercrime victimiza-
10.1111/j.2047-6310.2012.00085.x tion and problematic social media use: Findings from a nationally
Dallery, J., Stinson, L., Bolívar, H., Modave, F., Salloum, R. G., representative panel study. American Journal of Criminal Justice,
Viramontes, T. M., & Rohilla, P. (2020). mMotiv8: A 46(6), 862–881. https://doi.org/10.1007/s12103-021-09665-2
smartphone-based contingency management intervention to pro- McAndrew, F. T., & Jeong, H. S. (2012). Who does what on Facebook?
mote smoking cessation. Journal of Applied Behavior Analysis, Age, sex, and relationship status as predictors of Facebook use.
54(1), 38–53. https://doi.org/10.1002/jaba.800 Computers in Human Behavior, 28(6), 2359–2365. https://doi.org/
Davis, D. R., Kurti, A. N., Skelly, J. M., Redner, R., White, T. J., & 10.1016/j.chb.2012.07.007
Higgins, S. T. (2016). A review of the literature on contingency Moreno, M., Riddle, K., Jenkins, M. C., Singh, A. P., Zhao, Q., &
management in the treatment of substance use disorders, 2009– Eickhoff, J. (2022). Measuring problematic internet use, internet
2014. Preventive Medicine, 92, 36–46. https://doi.org/10.1016/j. gaming disorder, and social media addiction in young adults:
ypmed.2016.08.008 Cross-sectional survey study. JMIR Public Health and Surveil-
Dixon, S. (2022, April 28). Mobile share of U.S. social media visits lance, 8(1), e27719. https://doi.org/10.2196/27719
2017–2019. Statista. https://www.statista.com/statistics/477368/us- Murphy, J. G., Dennhardt, A. A., Skidmore, J. R., Borsari, B.,
social-media-visits-share/ Barnett, N. P., Colby, S. M., & Martens, M. P. (2012). A random-
Drebing, C. E., Van Ormer, E. A., Mueller, L., Hebert, M., ized controlled trial of a behavioral economic supplement to brief
Penk, W. E., Petry, N. M., Rosenheck, R., & Rounsaville, B. motivational interventions for college drinking. Journal of Consult-
(2007). Adding contingency management intervention to voca- ing and Clinical Psychology, 80(5), 876–886. https://doi.org/10.
tional rehabilitation: Outcomes for dually diagnosed veterans. 1037/a0028763
Journal of Rehabilitation Research and Development, 44(6), 851– Murphy, J. G., Skidmore, J. R., Dennhardt, A. A., Martens, M. P.,
865. https://doi.org/10.1682/jrrd.2006.09.0123 Borsari, B., Barnett, N. P., & Colby, S. M. (2012). A behavioral
Ellis, J. D., Struble, C. A., Fodor, M. C., Cairncross, M., economic supplement to brief motivational interventions for col-
Lundahl, L. H., & Ledgerwood, D. M. (2021). Contingency man- lege drinking. Addiction Research & Theory, 20(6), 456–465.
agement for individuals with chronic health conditions: A https://doi.org/10.3109/16066359.2012.665965
 19383703, 2023, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jaba.975 by BALL STATE UNIVERSITY MUNCIE, Wiley Online Library on [03/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
334 STINSON AND DALLERY

Owusu-Acheaw, M., & Larson, A. G. (2015). Use of social media and Smith, A., & Anderson, M. (2018, March 1). Social media use in 2018.
its impact on academic performance of tertiary institution stu- Pew Research Center. https://www.pewresearch.org/internet/2018/
dents: A study of students of Koforidua Polytechnic, Ghana. Jour- 03/01/social-media-use-in-2018/
nal of Education and Practice, 6(6), 94–101. https://doi.org/10. Stanley, T. B., Correia, C. J., & Irons, J. G. (2022). Contingency man-
1504/ijsmile.2016.079493 agement for smartphone and social media use: A feasibility study.
Oxford University Press. (2020). Social media In: Lexico.com. https:// Addiction Research & Theory, 30(5), 323–329. https://doi.org/10.
www.lexico.com/definition/wake [Accessed 7/13/2020] 1080/16066359.2022.2038140
Paakkari, L., Tynjälä, J., Lahti, H., Ojala, K., & Lyyra, N. (2021). Varona, M., Muela, A., & Machimbarrena, J. (2022). Problematic use
Problematic social media use and health among adolescents. Inter- or addiction? A scoping review on conceptual and operational def-
national Journal of Environmental Research and Public Health, initions of negative social networking sites uses in adolescents.
18(4), Article 1885. https://doi.org/10.3390/ijerph18041885 Addictive Behaviors, 134, Article 107400. https://doi.org/10.1016/j.
Packer, R. R., Howell, D. N., McPherson, S., & Roll, J. M. (2012). Investi- addbeh.2022.107400
gating reinforcer magnitude and reinforcer delay: A contingency Verbeij, T., Pouwels, J. L., Beyens, I., & Valkenburg, P. M. (2021). The
management analog study. Experimental and Clinical Psychopharma- accuracy and validity of self-reported social media use measures
cology, 20(4), 287–292. https://doi.org/10.1037/a0027802 among adolescents. Computers in Human Behavior Reports, 3,
Panova, T., & Carbonell, X. (2018). Is smartphone addiction really an 100090. https://doi.org/10.1016/j.chbr.2021.100090
addiction? Journal of Behavioral Addictions, 7(2), 252–259. https:// Vollmer, T. R., & Iwata, B. A. (1992). Differential reinforcement as
doi.org/10.1556/2006.7.2018.49 treatment for behavior disorders: Procedural and functional varia-
Petry, N. M., Tedford, J., Austin, M., Nich, C., Carroll, K. M., & tions. Research in Developmental Disabilities, 13(4), 393–417.
Rounsaville, B. J. (2004). Prize reinforcement contingency man- https://doi.org/10.1016/0891-4222(92)90013-V
agement for treating cocaine users: How low can we go, and with Vossen, H. G., & Valkenburg, P. M. (2016). Do social media foster or
whom? Addiction, 99(3), 349–360. https://doi.org/10.1111/j.1360- curtail adolescents’ empathy? A longitudinal study. Computers in
0443.2003.00642.x Human Behavior, 63, 188–124. https://doi.org/10.1016/j.chb.2016.
Pfund, R. A., Cook, J. E., McAfee, N. W., Huskinson, S. L., & 05.040
Parker, J. D. (2021). Challenges to conducting contingency man- Washington, W. D., Banna, K. M., & Gibson, A. L. (2014). Preliminary
agement treatment for substance use disorders: Practice recom- efficacy of prize-based contingency management to increase activ-
mendations for clinicians. Professional Psychology: Research and ity levels in healthy adults. Journal of Applied Behavior Analysis,
Practice, 52(2), 137–145. https://doi.org/10.1037/pro0000356 47(2), 231–245. https://doi.org/10.1002/jaba.119
Prendergast, M., Podus, D., Finney, J., Greenwell, L., & Roll, J. (2006). Widyanto, L., & McMurran, M. (2004). The psychometric properties of
Contingency management for treatment of substance use disor- the Internet Addiction Test. Cyberpsychology & Behavior, 7(4),
ders: A meta-analysis. Addiction, 101(11), 1546–1560. https://doi. 443–450. https://doi.org/10.1089/cpb.2004.7.443
org/10.1111/j.1360-0443.2006.01581.x Williams-Buttari, D., Deshais, M. A., Reeve, K. F., & Reeve, S. A. (2022).
Raiff, B. R., Jarvis, B., & Dallery, J. (2016). Text-message reminders plus A preliminary evaluation of the effects of a contingency
incentives increase adherence to oral antidiabetic medication in adults management + deposit contract intervention on problematic smart-
with Type 2 diabetes mellitus: A pilot study. Journal of Applied phone use with college students. Behavior Modification, 0(0).
Behavior Analysis, 49(4), 947–953. https://doi.org/10.1002/jaba.337 Advance online publication. https://doi.org/10.1177/
Riehm, K. E., Feder, K. A., Tormohlen, K. N., Crum, R. M., 01454455221113561
Young, A. S., Green, K. M., Pacek, L. R., La Flair, L. N., & Woods, H. C., & Scott, H. (2016). #Sleepyteens: Social media use in
Motjabai, R. (2019). Associations between time spent using social adolescence is associated with poor sleep quality, anxiety, depres-
media and internalizing and externalizing problems among US sion, and low self-esteem. Journal of Adolescence, 51(1), 41–49.
Youth. JAMA Psychiatry, 76(12), 1266–1273. https://doi.org/10. https://doi.org/10.1016/j.adolescence.2016.05.008
1001/jamapsychiatry.2019.2325 Young, K. S. (1998). Internet addiction: The emergence of a new clini-
Rosen, L. D., Whaling, K., Rab, S., Carrier, L., & Cheever, N. A. cal disorder. Cyberpsychology & Behavior, 1(3), 237–244. https://
(2013). Is Facebook creating “iDisorders”? The link between clini- doi.org/10.1089/cpb.1998.1.237
cal symptoms of psychiatric disorders and technology use, atti- Young, K. S., & Abreu, C. N. (2011). Internet addiction: A handbook
tudes, and anxiety. Computers in Human Behavior, 29(3), 1243– and guide to evaluation and treatment. John Wiley and Sons.
1254. https://doi.org/10.1016/j.chb.2012.11.012 Zendle, D., & Bowden-Jones, H. (2019). Is excessive use of social
Rosen, M. I., Dieckhaus, K., McMahon, T. J., Valdes, B., media an addiction? BMJ, 365, article l2171. https://doi.org/10.
Petry, N. M., Cramer, J., & Rounsaville, B. (2007). Improved 1136/bmj.l2171
adherence with contingency management. AIDS Patient Care and Zivnuska, S., Carlson, J., Carlson, D., Harris, R., & Harris, K.
STDs, 21(1), 30–40. https://doi.org/10.1089/apc.2006.0028 (2019). Social media addiction and social media reactions: The
Shannon, H., Bush, K., Villeneuve, P. J., Hellemans, K. G., & implications for job performance. The Journal of Social Psy-
Guimond, S. (2022). Problematic social media use in adolescents chology, 159(6), 746–760. https://doi.org/10.1080/00224545.
and young adults: Systematic review and meta-analysis. JMIR 2019.1578725
Mental Health, 9(4), Article e33450. https://doi.org/10.2196/33450
Shensa, A., Escobar-Viera, C. G., Sidani, J. E., Bowman, N. D.,
Marshal, M. P., & Primack, B. A. (2017). Problematic social
media use and depressive symptoms among U.S. young adults: A
nationally representative study. Social Science & Medicine, 182,
150–157. https://doi.org/10.1016/j.socscimed.2017.03.061
How to cite this article: Stinson, L., & Dallery, J.
Silverman, K., Holtyn, A. F., Rodewald, A. M., Siliciano, R. F., (2023). Reducing problematic social media use via
Jarvis, B. P., Subramaniam, S., Leoutsakos, J. M., Getty, C. A., a package intervention. Journal of Applied
Ruhs, S., Marzinke, M. A., & Fingerhood, M. (2019). Incentives Behavior Analysis, 56(2), 323–335. https://doi.org/
for viral suppression in people living with HIV: A randomized clin- 10.1002/jaba.975
ical trial. AIDS and Behavior, 23(9), 2337–2346. https://doi.org/10.
1007/s10461-019-02592-8
 19383703, 2023, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jaba.975 by BALL STATE UNIVERSITY MUNCIE, Wiley Online Library on [03/08/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
REDUCING PROBLEMATIC SOCIAL MEDIA USE 335

A P P E ND I X A

Behavior Allocation Survey

How many hours did you engage in each of the activi-
ties listed below yesterday? If there is another activity that
you spent time doing that is not listed, please add it to
the “other” section.

• Reading ______hours
• Watching TV or movies ______ hours
• Exercising ______ hours
• Talking on the phone ______ hours
• Texting ______ hours
• Attending class ______ hours
• School work ______ hours
• Job ______ hours
• Religious activities ______ hours
• Club activities ______ hours
• Hanging out with friends ______ hours
• Hanging out with romantic partner ______ hours
• Hanging out with family ______ hours
• Crafts (painting, crocheting, sewing, woodwork, etc.) ______ hours
• Playing a musical instrument ______ hours
• Using social media ______ hours
• Cooking ______ hours
• Playing video games ______ hours
• Surfing the internet (e.g., news sites, Buzzfeed, excluding social media) ______ hours
• Other
• ____________ ______ hours
• ____________ ______ hours
• ____________ ______ hours
• ____________ ______ hours