Beyond the Buzz: Investigating the Effects of a Notification-Disabling Intervention on Smartphone Behavior and Digital Well-Being

ABSTRACT

Turning off push notifications is a common recommendation to reduce smartphone screen time and to improve daily experiences of digital well-being. As only little empirical evidence exists, this study aims to test the effectiveness of this strategy. Based on a preregistered randomized controlled trial (N = 205), including objectively logged smartphone behavior data and daily mobile diary assessments of subjective experiences, we found that a one-week notification-disabling intervention did not affect smartphone behavior (i.e. checking frequency and screen time). This pattern was not dependent on people’s trait fear of missing out. In addition, we found no effects on perceived control, overuse, smartphone vigilance, productivity, and smartphone-related distraction. The intervention did, however, result in a decrease in perceived checking habit strength, indicating that users experienced their smartphone use to be more intentional. The absence of notifications also led to increased fear of missing out, suggesting that disabling notifications results in drawbacks rather than improvements in digital well-being. Together, these findings challenge the assumption that notifications play a prominent role in driving smartphone use and influencing user experiences. Findings are discussed in light of the complex interplay between device features and digital well-being.

Integrating mobile devices into our everyday lives has resulted in a status quo of anyplace, anytime connectivity (Klimmt et al., 2017). Many people use their smartphones extensively, with average daily screen times adding up to over four hours per day (e.g., Ohme et al., 2021). Complementing these objective statistics, many people also feel that they use their smartphones too much and express a lack of control over their usage patterns (e.g., Büchi et al., 2019; Saad, 2022). Notably, the feeling of using one’s smartphone too much is not fully contingent on people’s actual screen time (Kaye et al., 2020). Recent theoretical advances, therefore, underline the importance of subjective experiences in understanding individuals’ digital well-being (Büchi, 2021; Schneider et al., 2022; Vanden Abeele, 2021).

Digital well-being has been defined as the perceived optimal balance between digital connectivity and disconnectivity, which is contingent upon person-, device- and context-specific factors (Vanden Abeele, 2021). While previous media-psychological research has mainly focused on characteristics of the user (i.e., person-specific factors) to explain smartphone use and subjective experiences (e.g., Bayer et al., 2023; Klimmt et al., 2017), device features might be at least equally important to consider as most digital devices – and the applications that run on these devices – are specifically designed to drive smartphone engagement (Fasoli, 2021; Flayelle et al., 2023; Montag & Elhai, 2023).

One prominent device feature that drives smartphone engagement is notifications. It has been demonstrated that a higher number of notifications correlates with higher daily screen time and checking frequency (Liao & Sundar, 2022). The alerting nature of notifications grabs the user’s attention and triggers an urge to check their phone (Johannes et al., 2019). Receiving notifications can lead to immediate attention shifts that cause distraction from ongoing activities and may impair cognitive performance (e.g., Mendoza et al., 2018; Upshaw et al., 2022). In addition, notifications play a key role in forming pervasive checking habits, i.e., excessively checking one’s phone without any reason (Oulasvirta et al., 2012; Schnauber-Stockmann & Naab, 2019).

Due to these negative consequences of smartphone notifications on daily functioning, disabling notifications is a common recommendation to make one’s phone less intrusive and to “reclaim” one’s time (Center for Humane Technology, 2023). Few experimental studies, however, have investigated the effects of disabling smartphone notifications. In fact, none have yet looked at the effect of disabling notifications on objectively measured screen time. Moreover, the few studies that empirically tested the effectiveness of a notification-disabling intervention on subjective outcomes yielded mixed results. For example, one study found no effects on distraction and productivity (Fitz et al., 2019), whereas another found that the intervention resulted in less perceived smartphone distraction and higher perceived productivity (Pielot & Rello, 2017). Crucially, both studies found socio-emotional drawbacks of the intervention, such as increased anxiety about missing notifications (Fitz et al., 2019; Pielot & Rello, 2017).

In sum, at this point, we lack knowledge about the effects of turning off smartphone notifications on objective smartphone behavior, as well as conclusive evidence about the effects on subjective experiences of smartphone use. Moreover, previous research has not considered any individual differences as moderators of smartphone intervention effects, such as fear of missing out (FoMO). The present study addresses these gaps in the current literature by testing a notification-disabling intervention, employing a randomized controlled trial (RCT) design with objectively logged smartphone tracking data together with daily mobile diary assessments of subjective experiences.

Theoretical Background

The Impact of Notifications on Objective Smartphone Behavior

Notifications are theorized to affect smartphone use in two ways, specifically by increasing both checking frequency and screen time. First, the visual, auditory, and/or vibrating alert that signals the notification’s arrival serves as a direct reminder to check one’s phone. Notifications can thus trigger the user to initiate a smartphone session, even in situations in which they did not plan to use their phone (Heitmayer & Lahlou, 2021; Montalvo et al., 2021). A notification rarely goes unnoticed, as people carry their smartphones with them nearly always and, for instance, often place the device on their desk while working or studying (e.g., Heitmayer & Lahlou, 2021). Although notifications may be ignored, people often attend to them within minutes (Heitmayer & Lahlou, 2021; Mehrotra et al., 2016). Consequently, receiving many notifications during the day is associated with a high frequency of phone-checking behavior (M. Kim et al., 2021; Liao & Sundar, 2022).

Second, notifications are expected to result in higher daily screen time. Once the notification has made the user pick up their phone, the user may not only act upon the notification itself but may also be triggered by other smartphone cues (e.g., Bayer et al., 2016). For example, salient red badges in the corner of app icons motivate individuals to “clean” these other unaddressed notifications (Burchell, 2015). These visual cues may cause the user to engage in a series of other smartphone activities, resulting in a reactive, pinball-like usage session (Bayer & LaRose, 2018; Bayer et al., 2016). Notifications can thus increase usage session duration by acting as a gateway to other smartphone activities (Oulasvirta et al., 2012; Vanden Abeele, 2021). Correspondingly, it has been shown that the number of notifications received is highly correlated with one’s daily screen time (Liao & Sundar, 2022).

As research shows that notifications are associated with higher checking frequency and screen time, it seems plausible that disabling notifications would curb smartphone behavior. However, to our knowledge, only one study looked at objectively measured smartphone behavior as an outcome of a notification-disabling intervention. A two-week intervention found that daily phone checking frequency was not significantly different in the group who turned off notifications compared to the group who received notifications as usual (Fitz et al., 2019). This finding contradicts widespread assumptions that disabling notifications can help people regulate their smartphone use, highlighting the need to replicate these findings.

In addition, no study has yet examined the unique effect of a notification-disabling intervention on daily screen time. Although one study tested the effect of a one-week smartphone intervention on daily screen time, participants in this study also adhered to other instructions to curb smartphone behavior, such as keeping their smartphone out of reach during important tasks (Wasmuth et al., 2022). This study showed a significant reduction in objectively measured screen time in the experimental group. While this study provides preliminary evidence that a notification-management intervention can impact screen time, this effect may not be uniquely attributed to the lack of notifications. The same uncertainty applies to so-called “detox” interventions (i.e., time-outs from smartphone use or specific apps; for an overview, see Radtke et al., 2021), as some of these interventions involve, but are not limited to, the removal of notifications.

Based on theoretical reasoning as well as empirical literature linking the number of notifications to checking frequency and screen time, we proposed that:

H1:

During the intervention week, participants in the treatment group (i.e., notifications disabled) will have lower (a) checking frequency and (b) screen time compared to the baseline week.

Although the effectiveness of smartphone interventions likely depends on user characteristics, no study has yet looked at individual differences when testing a notification-disabling intervention. Notifications have a social dimension (i.e., relating to or originating from other people), and most smartphone notifications come from messaging applications (Pielot et al., 2018). Therefore, it is expected that individuals who strongly crave staying up to date with their social environment will react differently when not receiving notifications. We, therefore, argue that individuals’ so-called fear of missing out (FoMO) is an important factor to consider when evaluating the effectiveness of a notification-disabling intervention.

FoMO is defined as a “pervasive apprehension that others might be having rewarding experiences from which one is absent,” characterized by “the desire to stay continually connected with what others are doing” (Przybylski et al., 2013, 1841). As smartphone notifications provide immediate updates on this kind of information, individuals high in FoMO may be particularly sensitive to notifications (Johannes et al., 2019; Rozgonjuk et al., 2020). Accordingly, people high in FoMO feel more disrupted by smartphone notifications during their everyday activities (Rozgonjuk et al., 2019). Interestingly, a cross-sectional study found that individuals higher in FoMO spend more time on their phone and use it more frequently when in silent mode compared to other modes (Liao & Sundar, 2022). For individuals high in FoMO, disabling notifications may, therefore, not be as effective in curbing smartphone behavior because it may actually trigger them to check their phones more frequently to see whether they missed out on any updates. We hypothesized that:

H2:

The hypothesized intervention effects on (a) checking frequency and (b) screen time will be less strong for individuals with higher levels of trait FoMO.

The Impact of Notifications on Subjective Smartphone Experiences

Although the expected intervention effect on objective smartphone usage behaviors could be desirable, such usage quantification disregards individuals’ subjective experiences of digital well-being. It has been argued that individuals can have different subjective perceptions about their smartphone use, irrespective of the actual time spent. For example, two people who use their phones for the same amount of time might report differently on whether they are overusing their phone (Kaye et al., 2020; Vanden Abeele, 2021). Accordingly, perceived smartphone overuse is a qualitative experience of time spent on the device, where time spent is experienced as “non-meaningful and dissatisfactory a posteriori” (Fasoli, 2021, p. 1410). Perceived smartphone overuse thus occurs when one has exceeded a subjective and relative sense of optimal use amount (Büchi et al., 2019; Fasoli, 2021).

Notifications may lead to perceived overuse by triggering an individual to use their phone at any time. If the resulting phone session was unplanned (Bauer et al., 2017; Lukoff et al., 2018) or when using the smartphone implies suspending valuable offline activities (e.g., Reinecke et al., 2017), spending time on the smartphone may cause dissatisfaction afterward. In addition, although conveying a sense of urgency (Bartoli et al., 2022), notifications often turn out as unimportant once they have been checked (Neff & Wintersberger, 2022), having unnecessarily distracted the individual.

Similarly, notifications challenge individuals’ perceived control over their smartphone use through their push-and-pull mechanics (Montalvo et al., 2021). By delivering a notification, the smartphone “pushes” updates by actively reaching out to the user, thus controlling the flow of information, regardless of the user’s needs (S.-K. Kim et al., 2016). This is the opposite of a user-initiated way of use, in which the user “pulls” information by intentionally engaging in phone use (Mehrotra & Musolesi, 2018). Therefore, these push-and-pull functions of smartphones relate to reactive versus proactive smartphone behavior, respectively (Vorderer et al., 2016). Notifications thus induce a reactive mode in the user, limiting one’s sense of being in control (Bartoli et al., 2022; van Koningsbruggen et al., 2017).

More specifically, notifications may disrupt goal-directed behavior by causing the user to engage in unplanned smartphone use (Du et al., 2019; Meier, 2022). The resulting task-irrelevant engagement with one’s phone is likely to conflict with situational goals and responsibilities (Halfmann & Rieger, 2019; Hofmann et al., 2017; Ytre-Arne et al., 2020). Supporting this reasoning, research has found that individuals who feel more disturbed by notifications are more likely to fail to control their smartphone use at the expense of other important goals (Du et al., 2019; Halfmann & Rieger, 2019).

In addition, notifications may challenge autonomy by influencing behavior in a more indirect way, as they play an important role in smartphone-checking habit formation (Anderson & Wood, 2020; Bayer et al., 2022; M. Kim, 2014; Oulasvirta et al., 2012; Schnauber-Stockmann & Naab, 2019). Notifications are reasoned to drive checking behaviors, since the checking action is perceived as “successful” when a new notification is detected (Duke & Montag, 2017; Schnauber-Stockmann & Naab, 2019), or in some cases when the absence of new notifications or messages is revealed (Bayer et al., 2016). As behavioral habits are known to develop as a result of both repetition of the behavior and unpredictable rewards resulting from the behavior (Wood & Neal, 2007), checking for new notifications is reasoned to quickly become a habit for smartphone users (Bayer et al., 2022; Duke & Montag, 2017; Oulasvirta et al., 2012). In fact, smartphone-checking habits are found to be very prevalent in individuals’ lives (Heitmayer & Lahlou, 2021; Oulasvirta et al., 2012).

In short, notifications seem to promote individuals’ perceived lack of smartphone control and cause higher levels of perceived overuse and perceived checking habits, which suggests that removing notifications may lead to opposite effects. However, little research has investigated how a notification intervention impacts these constructs. One study found that people who disabled notifications for two weeks did not report less perceived overuse or more control over their phone, compared to the control group (Fitz et al., 2019). However, participants did perceive checking their phone as more intentional rather than habitual (Fitz et al., 2019). Similarly, Halfmann and Rieger (2019) found that switching notification sounds off, together with other strategies such as placing the phone out of sight, resulted in higher perceived control.

Although the available literature is limited and inconclusive, based on theoretical reasoning and other empirical findings, it was hypothesized that:

H3:

During the intervention week, participants in the treatment group (i.e., notifications disabled) will report higher levels of (a) perceived control, and lower levels of (b) perceived overuse and (c) habitual checking, compared to the baseline week.

The permanent connectedness provided by mobile devices has been argued to result in a mind-set of ongoing awareness and alertness toward online communication and information (Klimmt et al., 2018). This mind-set has been termed online vigilance (Klimmt et al., 2017), or, in the case of smartphones, smartphone vigilance (Johannes et al., 2019), and is composed of (1) the cognitive orientation toward one’s smartphone (salience); (2) the attentional sensitivity and motivational readiness to act upon smartphone cues (reactibility); and (3) the tendency to keep track of updates and ongoing events on the smartphone (monitoring; Klimmt et al., 2018; Reinecke et al., 2018). Although online vigilance denotes a trait-like disposition, it also shows situational fluctuation and within-person variability (Freytag et al., 2021; Gilbert et al., 2023).

Notifications may evoke high levels of situational smartphone vigilance. The cognitive salience of one’s smartphone could be reflected in thinking about notifications or, more indirectly, about messages or events that could be signaled by notifications. Monitoring refers to proactively checking for potential notifications, and reactibility reflects reactive handling of received notifications (Freytag et al., 2021; Johannes et al., 2021; Koessmeier & Büttner, 2022). Supporting this idea, experimental work has demonstrated that participants’ smartphone vigilance was higher when their smartphone was on the table during a cognitive performance test (Johannes et al., 2019; Koessmeier & Büttner, 2022), and even higher when participants received notifications but were not allowed to check them (Johannes et al., 2019).

In sum, as notifications constantly remind individuals of their smartphone, it seems likely that once no notifications are received for a certain period of time, an individual’s cognitive pre-occupation with their smartphone might be reduced. Therefore, we hypothesized:

H4:

During the intervention week, participants in the treatment group (i.e., notifications disabled) will report lower levels of smartphone vigilance compared to the baseline week.

Notifications have been found to distract individuals from primary tasks (Neff & Wintersberger, 2022; Throuvala et al., 2021), specifically in study contexts (Deng, 2020; Le Roux & Parry, 2021) or during performance tests (Johannes et al., 2019; S.-K. Kim et al., 2016). This distracting effect of notifications can be explained by the limited capacity model of mediated message processing (Lang, 2000), which proposes that people have finite attentional resources, so they are only able to process a limited amount of information at a given moment (Elhai et al., 2021; Wasmuth et al., 2022). As such, notifications can interrupt flow experiences (Duke & Montag, 2017; Elhai et al., 2021; Montag et al., 2016).

Furthermore, due to such distractions, notifications can hinder one’s general productivity experiences. Even when a task interruption is brief, such as a quick phone check, returning to one’s primary activity requires additional cognitive effort and time (Brumby et al., 2013). Switching back to a task may thus result in redundant work, such as reprocessing of information (Brumby et al., 2013). Consequently, notifications have been found to impair cognitive performance (e.g., Kaminske et al., 2022; Mendoza et al., 2018; Upshaw et al., 2022; but see; Johannes et al., 2019), even when people do not check the received notifications (Stothart et al., 2015).

While notifications are shown to affect distraction and productivity, studies testing a notification-disabling intervention have yielded mixed results regarding these two outcomes. Although one study found that a one-week notification-disabling intervention did not result in improved concentration or productivity (Fitz et al., 2019), another study showed that disabling notifications for one day did result in less distraction and higher productivity (Pielot & Rello, 2017). Other studies that tested multiple strategies concurrently within one intervention found similar results (Myers et al., 2022; Wasmuth et al., 2022). In addition, disabling the sounds of notifications for one week led to lower inattention and higher productivity (Kushlev et al., 2016). Therefore, we hypothesized:

H5:

During the intervention week, participants in the treatment group (i.e., notifications disabled) will report higher levels of (a) productivity and lower levels of (b) smartphone distraction compared to the baseline week.

While the notification-disabling intervention is expected to improve individuals’ digital well-being, not receiving notifications may also result in social-emotional drawbacks in the form of increased situational fear of missing out (Przybylski et al., 2013). Not receiving notifications has been found to induce a fear of missing out on notifications (i.e., notification-FoMO; Fitz et al., 2019; Pielot & Rello, 2017). In addition, as notifications convey updates about the online sphere, the intervention may raise concerns about not staying up-to-date on what is happening online (Fitz et al., 2019; Wegmann et al., 2017), which we refer to as content-FoMO. Therefore, we hypothesized:

H6:

During the intervention week, participants in the treatment group (i.e., notifications disabled) will report higher levels of (a) notification-FoMO and (b) content-FoMO compared to the baseline week.

Method

The study employed a mixed design in the form of a randomized controlled trial, consisting of a between-factor (control vs. treatment group) and a within-factor (baseline vs. intervention week). For exploratory analyses not reported in this paper, the study included a third post-intervention week. Data was collected using a research app (Murmuras) that tracked participants’ smartphone behavior and sent daily diary surveys, in addition to three larger surveys that were administered via an online survey tool. The study design, study materials, and hypotheses were preregistered on the Open Science Framework¹ and were approved by the Ethics Review Board of the University of Amsterdam (2022-YME-15436).

Procedure

We recruited a sample of young adults (18–30 years old) as they may be particularly in need of effective strategies to curb their smartphone use, considering that they have higher levels of smartphone usage and perceived overuse compared to older adults (e.g., De Marez et al., 2023; Saad, 2022). Participants had to use an Android smartphone since the research app was only available for Android.

In the first phase of the data collection, participants were recruited via the online lab facilities of the university. As the number of students signing up made it seem unlikely that we would reach our target sample size of N = 200 (see OSF), an additional sample was recruited via an online panel company (Prolific). All participants were financially compensated for their participation (15–20 €). After providing informed consent and completing the entry survey (T0), participants were instructed to install the research app. Participation in the study took 25 days, including the three weeks of interest and four transition days. Figure 1 provides an overview of the procedure.

Figure 1. Study procedure.

The first week served as a baseline measure with the same procedure for all participants, followed by the second week in which the intervention took place for the treatment group. During these two weeks, all participants received short daily diary surveys every evening at 8:00 pm to complete before going to sleep.² On the transition days after the baseline week (T1) and the intervention week (T2), participants completed a longer survey that was also sent at 8:00 pm. At the end of the T1 survey, participants were randomly assigned to one of the conditions, with a ratio of 2(treatment):1(control) to account for anticipated attrition within the treatment group. The treatment group was instructed to disable all notifications for all apps, except the research app, in their smartphone settings and keep them disabled for a week. The control group was told that they did not have to change any settings. After the intervention week, participants completed a final survey (T2). All participants were asked to keep the research app installed for another week. In this third week, participants were told they could manage their notifications according to their preferences and no longer had to complete any surveys, while their smartphone behavior was still being tracked. On the final day, a reminder was sent to uninstall the app.

Participants

Five inclusion criteria were preregistered. First, participants had to install the research app after completing the T0 intake survey. Second, participants needed to have at least fourteen days of tracking data. Third, completing the T1 survey was essential, because this survey included the condition assignment and treatment instructions. Fourth, in both weeks, participants had to complete at least four out of seven daily surveys (Griffiths et al., 2022). Fifth, for the treatment group, intervention compliance had to be sufficient (i.e., the tracking data shows that the total number of notifications was reduced by at least 75% in the intervention week compared to the baseline week). See Figure 2 for the number of excluded participants based on these criteria. Applying the preregistered exclusion criteria resulted in a final sample size of N = 205 (M_age = 24.45; SD_age = 3.92; 52% female, n_intervention = 114, n_control = 91). At baseline, participants were highly motivated to reduce their screen time (“I want to reduce the time I spend on my smartphone;” 1 = strongly disagree to 7 = strongly agree, M = 5.37; SD = 1.53).

Figure 2. Flow chart sample exclusion criteria.

Measures

In the T0 survey, demographic information was assessed (age; gender), as well as trait FoMO. In the daily diary surveys, subjective outcome variables were assessed. On average, participants completed 6.79 daily surveys in the baseline week (SD = 0.56), and 6.67 daily surveys in the intervention week (SD = 0.69), resulting in a completion rate of 96%.

Trait fear of missing out

To assess FoMO as a trait variable, five items from the FoMO trait subset (Wegmann et al., 2017) of the original Fear of Missing Out scale (Przybylski et al., 2013) were used, with a response scale ranging from 1 (very untrue of me) to 7 (very true of me). Measurement validity was evaluated with a Confirmatory Factor Analysis (lavaan package; version 0.6–16; Rosseel, 2012). However, two factors emerged (see OSF). Therefore, we decided to keep the factor with the most items (n = 3). Internal consistency of the resulting scale was acceptable (Cronbach’s α = .74; psych package; version 2.3.9; Revelle, 2022). A mean scale was computed (M = 3.78, SD = 1.37).

Smartphone behavior

The two indicators of participants’ smartphone behavior, i.e., checking frequency and screentime, were objectively measured by the research app, as well as the daily number of notifications received.

Checking Frequency. Phone checking is defined as the initiation of a phone session, which consists of unlocking the phone, followed by a series of phone activities, until the phone is locked again. Checking frequency refers to the total number of phone checks per day. In the baseline week, participants checked their phone 85.78 times per day on average (SD = 52.74, median = 74, min = 2, max = 401).

Screen Time. The data included some extreme outliers of people using their phone for 22 hours per day. App-specific data showed some apps that typically keep the screen turned on, even when the user does not use their phone (YouTube Vanced; Basic Daydreams). Therefore, these apps were removed from the general screen time data.³ Still, some extreme outliers were present, but these people seemed to use gaming apps extensively and were thus not removed. On average, participants spent more than five hours per day on their phone in the baseline week (M = 327.43 minutes, SD = 175.48, median = 294.87, min = 3.48, max = 1270.00).

Number of Notifications. The daily number of notifications was computed by adding up all logged notification entries per day. We removed notifications that were sent by the phone system or by apps that are not installed by the user (e.g., phone launcher, alarm). On average, participants in the control group received 150.90 (SD = 134.19) notifications in the baseline week and 147.88 (SD = 127.55, Cohen’s d = .07) in the intervention week. The treatment group received 154.00 (SD = 143.28) notifications per day in the baseline week, which was reduced to 8.17 during the intervention (SD = 11.72, Cohen’s d = 1.15), indicating high compliance among participants.

Daily subjective outcomes

In the daily surveys, subjective outcomes were assessed with single-item measures, with 7-point answering scales (1 = not at all to 7 = very much). Perceived overuse was measured with the item “Today, I spent too much time on my phone.” Perceived control was measured with the item “Today, I felt in control of my own smartphone use.” Habitual checking was assessed with one item adapted from the self-report behavioral automaticity index (SRBAI; Gardner et al., 2012): “Today, I mostly checked my phone without thinking.” Smartphone vigilance was measured with one item per dimension, adapted from the Online Vigilance Scale (Reinecke et al., 2018). Salience was measured with “Today, I had a hard time disengaging mentally from online content on my smartphone,” and monitoring was measured with “Today, I constantly monitored on my smartphone what was happening online.” The reactibility dimension was left out as referring to incoming notifications would be inapplicable during the intervention. Perceived productivity was assessed with one item by Fitz et al. (2019): “Today, I was productive.” Smartphone distraction was assessed with the item “Today, I felt distracted by my phone.” Notification-FoMO was measured with the item “Today, I felt I was missing out on important notifications on my phone” (Fitz et al., 2019). Content-FoMO was measured with the item “Today, I felt I was missing out on what was happening online.”

Analytical Strategy

All analyses were conducted in R (version 4.3.2; R Core Team, 2023). Following the preregistration, we tested the hypotheses with Linear Mixed Models (lme4 package; version 1.1–35.1; Bates et al., 2015) with Maximum Likelihood estimation. We estimated two-level models for all outcome variables in which weeks (level 1; baseline vs. intervention week) are nested within participants (level 2). To examine the effect of the intervention, we were interested in the interaction effect of week (baseline vs. intervention) and group (control vs. treatment). This interaction was tested in separate models for the different outcomes of H1, H3, H4, H5, and H6. To test trait FoMO as a moderator (H2), a three-way interaction model was tested. As preregistered, the daily measures were aggregated into week means for all outcomes, resulting in two data points per participant for each outcome. For an overview of all means and standard deviations in both weeks, see Table 1. Zero-order correlations between key variables in the baseline week can be found in Table 2.

Table 1. Overview of means, standard deviations, and LMM results.

	Control group (n = 91)		Treatment group (n = 114)
	Baseline week M (SD)	Intervention week M (SD)	Baseline week M (SD)	Intervention week M (SD)	t	p	Hypothesis
Checking frequency (n per day)	86.11 (43.26)	86.36 (45.00)	85.51 (48.51)	83.78 (53.00)	−0.81	.421	H1a
Screen time (minutes per day)	305.57 (140.86)	312.63 (150.88)	344.86 (152.06)	335.23 (153.88)	−1.69	.094	H1b
Perceived control	4.75 (0.95)	4.70 (0.98)	4.54 (0.92)	4.57 (1.00)	0.66	.509	H3a
Perceived overuse	3.51 (1.00)	3.31 (1.17)	3.66 (0.98)	3.44 (1.16)	−0.08	.933	H3b
Habitual checking	3.48 (1.34)	3.40 (1.30)	3.79 (1.13)	3.28 (1.20)	−3.08	.002	H3c
Smartphone vigilance (salience)	3.05 (1.22)	2.98 (1.16)	3.15 (1.07)	3.13 (1.25)	0.38	.705	H4
Smartphone vigilance (monitoring)	2.79 (1.21)	2.85 (1.27)	3.04 (1.12)	3.14 (1.26)	0.31	.756	H4
Productivity	4.43 (1.00)	4.09 (1.10)	4.08 (0.79)	4.13 (0.97)	3.09	.001	H5a
Smartphone distraction	3.22 (1.02)	3.10 (1.14)	3.44 (1.04)	3.22 (1.23)	−0.64	.523	H5b
Notification-FoMO	2.68 (1.16)	2.69 (1.17)	2.70 (1.04)	4.01 (1.63)	6.39	<.001	H6a
Content-FoMO	2.52 (1.20)	2.61 (1.22)	2.72 (1.00)	3.55 (1.44)	4.66	<.001	H6b

df = 205. Hypotheses confirmed at the p < .05 level are marked in bold.

Table 2. Zero-order correlations baseline week.

	1.	2.	3.	4.	5.	6.	7.	8.	9.	10.	11.	12.	13.
1. Number of notifications	-
2. Daily checking frequency	0.32***	-
3. Daily screen time	0.33***	0.19**	-
4. Perceived control	0.03	−0.09	−0.18*	-
5. Perceived overuse	0.16*	0.16*	0.36***	−0.62***	-
6. Habitual checking	0.04	0.13	0.02	−0.40***	0.45***	-
7. SV salience	0.06	0.15*	0.28***	−0.56***	0.72***	0.52***	-
8. SV monitoring	0.08	0.19**	0.24***	−0.44***	0.66***	0.49***	0.81***	-
9. Productivity	0.04	0.08	−0.14*	0.45***	−0.28***	−0.12	−0.20**	−0.16*	-
10. Smartphone distraction	0.11	0.08	0.26***	−0.66***	0.83***	0.57***	0.77***	0.70***	−0.24***	-
11. Notification-FoMO	0.03	0.08	0.11	−0.27***	0.41***	0.44***	0.69***	0.67***	−0.05	0.51***	-
12. Content-FoMO	0.01	0.07	0.15*	−0.36***	0.53***	0.39	0.70***	0.76***	−0.09	0.60***	0.79***	-
13. FoMO (trait)	−0.08	−0.11	−0.03	−0.03	0.07	0.05	0.06	0.06	−0.13	0.06	0.07	0.08	-

Correlations based on the week means of the daily outcome variables in the baseline week. *p < .05, **p < .01, ***p < .001.

Results

Objective Intervention Outcomes

No significant effect of the intervention was found for phone checking frequency, t(205) = −0.81, p = .421, nor for daily screen time, t(205) = −1.69, p = .094. These results indicate that smartphone behavior did not change as a result of the intervention, thus not supporting H1a and H1b. In addition, no support was found for the proposed three-way interactions with trait FoMO as a moderator, neither for checking frequency, t(205) = −0.35, p = .729, nor screen time, t(205) = −0.30, p = .766. Therefore, no support was found for H2a and H2b.

Subjective Intervention Outcomes

Contrary to expectations, perceived control did not significantly change as a result of the intervention, t(205) = 0.66, p = .509, nor did the intervention affect perceived overuse, t(205) = −0.08, p = .933. Both H3a and H3b were thus rejected. In line with expectations, the intervention resulted in a decrease in perceived habitual checking, t(205) = −3.08, p = .002, Cohen’s d = .43, supporting H3c. In contrast, no significant intervention effect was found for smartphone vigilance, for both the salience dimension, t(205) = 0.38, p = .705, and the monitoring dimension, t(205) = 0.31, p = .756. Thus, no support was found for H4.

A significant effect was found for productivity, t(205) = 3.09, p = .002, Cohen’s d = .43, however, not in the expected direction, as the control group showed a decrease in productivity while the treatment group only marginally showed the hypothesized increase. In addition, no significant intervention effect was found for smartphone distraction, t(205) = −0.64, p = .523. Therefore, H5a and H5b could not be confirmed. Finally, we tested the effects of the intervention on the two state FoMO outcomes. In line with expectations, participants in the treatment group had higher notification-FoMO, t(205) = 6.39, p < .001, Cohen’s d = .89, and experienced significantly more content-FoMO during the intervention, t(205) = 4.66, p < .001, Cohen’s d = .65, supporting H6a and H6b.

Exploratory Analyses

In addition to the preregistered analyses, we added a number of exploratory analyses. An overview of the exploratory results can be found in the online Appendix on OSF.

First, to test the robustness of our findings, we reran all models with the non-aggregated data (i.e., the dataset contained daily values instead of the weekly averages). Here, we also included a random slope for week per person to allow for different slopes over time across individuals. We additionally ran these models with age and gender as control variables. None of these models yielded different findings (see Table A1).

Next, we ran several analyses to see if intervention effects could be revealed when taking into account other potentially relevant factors. First, we tested the assumption that not receiving notifications may reduce the use of social media apps in particular, as this use may be particularly driven by notifications compared to the use of other apps. Since the data included app-specific smartphone use, we tested for an intervention effect specifically on social media screen time, but no significant effect was found, t(205) = −0.37, 95% CI [−16.59, 11.40]. Thus, the intervention also did not affect social media-specific screen time. Second, we tested whether the intervention would be less effective for people who normally have their phone on silent mode most of the time. Although we excluded these people (n = 85) and ran all models again, this did not yield different findings (see Table A2).

In addition, we tested two potential moderators of the intervention effects on smartphone behavior. First, we tested whether people who were more motivated to reduce their smartphone use showed stronger reductions in checking frequency and screen time as a result of the intervention, but this was not the case. Second, we tested whether intervention effects on smartphone behavior would be different for people who normally receive more notifications. We thus added the number of received notifications in the baseline week as a moderator. However, no significant effects were found (see Table A3).

Finally, we examined the temporal patterns of effects during the intervention week, as it could be that intervention effects only occurred gradually over time and were not detected by a comparison of the weeks. We therefore tested whether effects followed a linear pattern during the intervention week for the treatment group (see Table A4). We found significant linear decreases during the intervention for perceived overuse, t(651.79) = −2.45, 95% CI [−0.10; −0.01], the salience dimension of smartphone vigilance, t(650.70) = −3.01, 95% CI [−0.11; −0.02], the monitoring dimension, t(651.24) = −4.02, 95% CI [−0.14; −0.05], smartphone distraction, t(651.92) = −2.73, p = .006, 95% CI [−0.12; −0.02], notification-FoMO, t(649.25) = −5.12, p < .001, 95% CI [−0.16; −0.07], and content-FoMO, t(649.76) = −6.18, p < .001, 95% CI [−0.18; −0.10]. This suggests that, although the hypotheses regarding perceived overuse, salience, monitoring, and smartphone distraction were not confirmed, these outcomes were reduced during the intervention.

Discussion

Based on widespread assumptions about the pervasive negative effects of notifications on individuals’ smartphone experiences, one common recommendation is to turn off notifications. However, little empirical evidence existed as to whether disabling notifications is indeed effective in curbing objective smartphone behavior and beneficial for smartphone users’ subjective experiences. The current study, therefore, investigated the effects of a notification-disabling intervention on both objectively logged smartphone behavior and daily subjective experiences. Strikingly, we found that turning off notifications did not affect objective smartphone behavior (i.e., screen time or checking frequency), nor did it substantially improve participants’ subjective smartphone experiences.

No Effects on Objective Smartphone Behavior

Due to limited prior work, the expectations that the intervention would impact smartphone behavior were primarily based on correlational findings indicating that the more notifications one receives, the higher their screen time and checking frequency (Liao & Sundar, 2022). In the baseline week, we replicated these correlations as the number of notifications was positively related to both checking frequency (r = 0.32, p < .001) and screen time (r = 0.33, p < .001; see Table 2). However, our results show that turning off notifications did not lead to reductions in screen time and checking frequency. These findings shed a different light on the cross-sectional associations, as they suggest that notifications might not necessarily drive smartphone use but are rather an indicator of how people use their smartphone. In other words, people who use their smartphone more extensively also receive more notifications as they, for instance, might use more apps and communicate more online.

In finding no effect of the intervention on checking frequency, we replicate the null finding by Fitz et al. (2019). Our finding contrasts, however, with that of Wasmuth et al. (2022), whose intervention did result in a decrease in checking frequency. Yet, as their study employed various intervention strategies (e.g., keeping one’s phone out of sight; Wasmuth et al., 2022), it is likely that their finding is not attributable to the absence of notifications but rather to the combination of strategies.

Overall, the null effects on smartphone behavior support the idea that smartphone behavior has become a highly habitualized behavior that is resistant to change (e.g., Oulasvirta et al., 2012; Schnauber-Stockmann & Naab, 2019). In other words, even when cues that trigger smartphone use (i.e., notifications) are removed, people will still frequently use their phone as it has become an integral part of their everyday lives. Consistent with previous work, this implies that smartphone use is mostly self-initiated rather than driven by notifications (Heitmayer & Lahlou, 2021). The notion that smartphone behaviors are highly habitualized further indicates that a one-week intervention may be too short to affect these deep-rooted habits, suggesting the need for studying longer-term interventions.

Little Effects on Subjective Experiences

Most subjective experiences were also not affected by the intervention. Generally, these findings suggest that notifications play a less prominent and active role in impacting user experiences than expected.

Crucially, no effects of the intervention were found on the two smartphone vigilance indicators (i.e., salience and monitoring), which indicates that, despite the absence of notifications, individuals kept thinking about their smartphones and were inclined to monitor what was happening on their smartphones. This finding suggests that smartphone vigilance is more stable than anticipated, which may be attributed to people keeping their phones in sight (Johannes et al., 2019; Koessmeier & Büttner, 2022). Alternatively, individuals may have internalized the smartphone vigilance mind-set to such an extent that removing device cues does not make a difference. In a broader sense, the finding confirms that being permanently online and permanently connected is not only deeply integrated into individuals’ daily behavioral routines but also into their “thinking routines” (Klimmt et al., 2017).

In fact, perhaps the very absence of notifications triggers speculative thoughts about notifications that one has potentially received without knowing (i.e., salience), leading to the tendency to check if that is the case (i.e., monitoring). The latter explanation fits well with our finding that the intervention resulted in increased situational fear of missing out. People felt more worried about missing notifications and online updates, a finding in line with prior work (Fitz et al., 2019; Kushlev et al., 2016; Pielot & Rello, 2017), which may be driven by a social pressure to immediately respond to incoming messages (Bayer et al., 2016; Lutz, 2023; Mai et al., 2015). Generally, the cognitive (pre)occupation with their phone may have led people to use their phones as much as they normally would, potentially explaining the lack of intervention effects on smartphone behavior (Le Roux & Parry, 2022).

In addition, disabling notifications did not lead to improved productivity or reduced smartphone distraction. On the one hand, this is surprising because earlier work did find these effects (Fitz et al., 2019; Kushlev et al., 2016; Pielot & Rello, 2017), and the strategy is often recommended to improve everyday cognitive performance and efficiency (e.g., Fleming, 2021). On the other hand, however, it is not so surprising as participants’ smartphone behavior remained unchanged during the intervention. This suggests that they were equally absorbed in their phones during studying and working tasks, given the integration of smartphones in such activities (e.g., M. Kim et al., 2021; Mendoza et al., 2018).

Interestingly, we found that the intervention did significantly reduce individuals’ perceived checking habit, supporting the notion that not receiving notifications creates an impression of intentionality of smartphone use (Fitz et al., 2019; Garaialde et al., 2020). One possible explanation is that a phone without notifications can only be “checked” by opening a specific app to see updates, which indicates a deliberate thought process about what one intended to find out. Perhaps this more conscious smartphone use would eventually result in a decrease in actual checking frequency, again highlighting the need for longer-term intervention studies.

Furthermore, we exploratively tested for linear patterns during the intervention, and found that overuse, online vigilance, and smartphone distraction were reduced over the course of the intervention. Thus, despite the null effects on these outcomes, the intervention seems to have impacted them to some extent. If the decreasing effects are consistent, a longer intervention could thus reveal significant improvements. Moreover, both notification-FoMO and content-FoMO showed linear decreases during the intervention, which suggests that the drawbacks of the intervention fade out when one gets used to the absence of notifications.

Overall, we can conclude that the one-week intervention did not significantly improve everyday user experiences. Interestingly, however, it should also be noted that participants’ digital well-being was not particularly low during the baseline week. For example, participants reported rather high levels of perceived control and low levels of smartphone distraction. Smartphone vigilance means were also rather low, consistent with existing work (e.g., Gilbert et al., 2023; Johannes et al., 2021). Thus, despite the high frequency of smartphone checking (on average 86 times per day) and high screen time (on average over five hours), and participants’ strong motivations to reduce their smartphone use, participants did not report substantial smartphone-related issues. This suggests that the negative subjective experiences associated with smartphones may not be as severe as commonly believed. Alternatively, it could be that people have become accustomed to these experiences and do not perceive them as particularly problematic compared to their everyday lives or the lives of others (e.g., Chang et al., 2023). The deep integration of smartphones may have altered people’s perceptions of what is considered normal (Klimmt et al., 2017).

Limitations and Methodological Challenges

The present study faces several limitations. First, due to the practical considerations of ensuring the brevity of daily administered questionnaires, we relied on single-item measures for all outcome variables. While this approach may reduce measurement reliability, it is a common methodological approach to capture daily experiences without overburdening participants. In addition, recent studies have proven the reliability and validity of single-item measures (e.g., Matthews et al., 2022).

A second shortcoming is that only Android users could participate in our study, which is a common issue in smartphone logging research due to the limited availability of passive logging applications for iOS devices (e.g., Deng, 2020; Johannes et al., 2021). This divide may be relevant to consider as it has been shown that Android and iOS users differ somewhat in personality (Götz et al., 2017). However, it has also been reported that Android users have higher opt-in rates for app notifications than iOS users because of different opt-in regulations (Böhm et al., 2019), suggesting that a notification-disabling intervention yields stronger effects for Android users. Given our null findings, we would thus not expect different outcomes for iOS users. Nonetheless, future researchers could consider data donation approaches instead of smartphone logging apps to obtain a more inclusive sample (e.g., Baumgartner et al., 2023; Ohme et al., 2021).

Finally, although neither the main analyses nor the exploratory analyses point to significant intervention effects, there may still be other relevant variables that were not considered in the present study. For example, social norms around mobile connectivity (e.g., availability norms) potentially explain why people are afraid to miss notifications or updates (e.g., Bayer et al., 2016). We also did not check whether participants used their phones for work purposes, which may explain why some people’s smartphone behavior did not change. In addition, phone tracking data could be supplemented by investigating the use of other devices, as people may compensate for not receiving notifications by, for instance, checking for updates on their tablet or laptop. Future research could also more closely examine whether disabling notifications might change usage patterns throughout the day (e.g., fragmented vs. sticky use; Siebers et al., 2023).

Conclusion

Our study is one of the first to test the effects of a notification-disabling intervention on both objective smartphone behavior and subjective user experiences. Despite high compliance and a sample that was highly motivated to reduce their smartphone use, we found that screen time and checking frequency were unaffected by the intervention, as well as most subjective experiences, including smartphone vigilance. In fact, participants did experience more situational FoMO, suggesting that disabling notifications results in drawbacks rather than improvements. Thus, although smartphone notifications are argued to be “the most ubiquitous feature of the most ubiquitous device on the planet” (Fitz et al., 2019, p. 84), simply turning them off does not seem to make the smartphone much less ubiquitous. Our findings highlight the need for empirical support when formulating recommendations for improving digital well-being. Furthermore, the findings suggest that notifications do not necessarily drive smartphone use, but they may reflect how extensively people use their smartphones.

Open scholarship

This article has earned the Center for Open Science badges for Open Data, Open Materials and Preregistered. The data and materials are openly accessible at https://osf.io/45q72/.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data described in this article are openly available on the Open Science Framework at https://osf.io/45q72/.

Additional information

Funding

This work was supported by the Digital Communication Methods Lab (DigiComLab) of the Amsterdam School of Communication Research (ASCoR).

Notes

1. One additional hypothesis (about compliance) and one research question (about the post-intervention week) will not be discussed in this paper due to limited space but they are reported on OSF.

2. The student sample completed these surveys via the research app, whereas the Prolific sample received reminders from the app to complete the new survey via Prolific.

3. To test robustness, analyses for H1 were also done with data including the screen-on apps but this did not yield different findings.