Momentary motivations for digital disconnection: an experience sampling study

Abstract

A much-discussed solution for undesirable (over-)use of mobile technologies lies in digital disconnection. Reasons for why individuals reduce their digital media use have been assessed mostly cross-sectionally without accounting for various disconnection practices across everyday situations. This study focuses on three motivations to disconnect that can vary between situations: to (a) avoid distractions, (b) improve well-being, and (c) be more present. A 14-day experience sampling study with 230 young adults (M_age = 25.31, SD = 4.50) yielded 7,360 situations of disconnective behavior. Multilevel regression analyses show that motivations to avoid distractions and to be more present were relevant for disconnection on the situational level. However, a person’s average level of these motivations did not predict disconnective behavior. The well-being motivation was not associated with disconnection either between or within participants. Additional analyses explore variations across time and different levels of digital disconnection.

With mobile communication technologies’ ubiquitous presence and relevance in our lives, balancing digital media use and non-use has become a central challenge for many (Vanden Abeele, 2021). From digital detoxes (e.g., Radtke et al., 2022) to regulations that afford workers the “right to disconnect” (Pansu, 2018, p. 100), one avenue to (re-)establishing balance is to temporarily disconnect from digital media. Increasingly, individuals report a high personal relevance of reducing or otherwise changing their use of digital media. For example, around two thirds of 14- to 29-year-olds in Germany plan to reduce time spent with digital devices and services (Beisch & Koch, 2022). The appealing nature and salient connection cues of digital media clearly drive their ubiquitous usage in daily life (Bayer et al., 2016; Meier et al., 2023), and digital media are central in an interconnected society (Nguyen & Hargittai, 2024). However, accumulating evidence suggests that this ubiquity of mobile connection can increase negative experiences, such as goal conflicts (Vanden Abeele, 2021), phubbing (Nuñez & Radtke, 2023), and stress (Dragano & Lunau, 2020). Experiencing these downsides to constant connectivity in everyday life may lead individuals to be motivated to temporarily disconnect from digital media.

Research has identified several motivations for why individuals engage in digital disconnection (for an overview see Nassen et al., 2023; Nguyen et al., 2022; Vanden Abeele et al., 2022). For example, individuals frequently report the motivations to prioritize face-to-face encounters over digital communication (Agai, 2022), to avoid threats to their well-being, such as exposure to harmful content like hate speech or unrealistic body imagery (Stahel & Baier, 2023; Vandenbosch et al., 2022), or to avoid being distracted by digital media (Nguyen et al., 2022). Nassen and colleagues (2023) systematize motivations into six categories: perceived overuse, social interactions, psychological well-being, productivity, privacy, and perceived usefulness. The salience and intensity of these motivations for digital disconnection may vary both in the short and the long term (Nguyen, 2023), from daily situations to important life transitions like job changes (Nguyen et al., 2021, 2022). Yet, to our knowledge, studies have addressed motivations exclusively at single time points, i.e., in interviews or surveys asking users to reflect retrospectively about their average or general disconnection motivations and behaviors. To address this gap, we focus on three motivations for digital disconnection that likely vary situationally: to (1) avoid distractions, (2) improve well-being, and (3) be more present in the moment (see Nassen et al., 2023; Nguyen et al., 2022).

Previous studies have captured how users reflect their disconnective behavior and underlying motivations in general, yet often ignored situational factors that are central to whether and how people engage in disconnection in daily life (Bozan & Treré, 2023; Geber & Nguyen, 2023; Vanden Abeele et al., 2022). While some choose to abstain entirely from certain types of digital media, many disconnection practices in everyday life are likely more ephemeral (Fast, 2021), targeting immediate or short-term outcomes such as well-being or productivity. Rather than missing all benefits from digital technologies (Radtke et al., 2022; Vanden Abeele, 2021), individuals may be more likely to engage in temporary, short-term disconnection behaviors targeted at specific, personally undesirable aspects of digital technology.

To understand digital disconnection behaviors and the motivational forces behind it, we need to know how much the identified motivations matter in everyday life situations (i.e., within-person effects) or whether they are mainly based on relatively stable personal characteristics (i.e., between-person effects). Furthermore, we aim to explore whether motivations are enacted within the same situation or whether they motivate disconnection behavior over time as well (i.e., lagged effects). We investigate motivations for and practices of digital disconnection using the experience sampling method (ESM) with N = 230 participants who report T = 7,360 disconnection episodes across two weeks. In doing so, this study aims to answer the central question of how important three previously identified motivations for digital disconnection—the motivation to avoid distractions, to improve well-being, and to be present offline—are for situational disconnection behavior in everyday life.

Conceptualizing situational digital disconnection

Digital disconnection has been understood in various ways. Usually, three aspects are central to its definition: digital disconnection is conceptualized as (a) the voluntary reduction in usage of one or more aspects of digital media, (b) across one or more digital devices, apps, features, (types of) interactions, and messages, (c) over a certain period of time (Meier & Reinecke, 2021; Nassen et al., 2023; Vanden Abeele et al., 2022).

Prior research in communication science has mostly understood voluntary, self-directed digital disconnection as “deliberate (i.e., chosen by the individual)” (Nassen et al., 2023, p. 13), implying that beyond being self-directed, the behavior always needs some degree of conscious intention (Bozan & Treré, 2023; Nassen et al., 2023). Taking a situational perspective of digital disconnection in daily life, we, however, propose that disconnection may also entail behaviors that were initially enacted in pursuit of a goal, but have become habitualized over time (Williamson & Wilkowski, 2022). Many behaviors associated with digital disconnection are not necessarily dependent on deliberation or conscious decision-making in the moment. Instead, disconnection may often be a response to internalized situational or normative requirements around technology non-use in certain social settings (Geber & Nguyen, 2023; Nguyen et al., 2022), such as in-person conversations (Nguyen et al., 2022). Therefore, through the lens of psychological theorizing on self-regulation (Inzlicht et al., 2021) and media habits (Bayer et al., 2016; Meier et al., 2023), definitions of everyday disconnection need to consider habitualized behaviors. Hence, in this study, we extend prior research by also considering automatically enacted, situationally non-deliberate disconnection behaviors in daily life.

Beyond the question of intentionality, prior research has also responded differently to the question of what it is, exactly, about digital media and computer-mediated communication (CMC) that users wish to disconnect from. The Hierarchical CMC Taxonomy (Meier & Reinecke, 2021) allows us to systematically analyze disconnective behaviors across a variety of communication technologies, their characteristics, and their uses. The CMC Taxonomy is useful to assess digital disconnection because it provides an organizing framework for a systematic, domain-independent assessment of digital media (see Nassen et al., 2023; Radtke et al., 2022; Vanden Abeele, 2021). Through the lens of the taxonomy, disconnection can take place at six levels: the devices, types of applications, branded applications, features, interactions, and messages that users may disconnect from. For example, at the device level, disconnection could entail leaving the smartphone at home, at the feature level turning off notifications, and on the message level muting certain words on social media. On the message level, we specifically focus on message content over other message characteristics such as modality based on previous research that identifies content, e.g., social media content, as a potential source of negative experiences. Theoretical and empirical links between digital media and undesirable outcomes on the content level can be found in phenomena such as cyberbullying, unrealistic body imagery, hate speech, or negative news (Skovsgaard & Andersen, 2020; Stahel & Baier, 2023; Vandenbosch et al., 2022).

Applying the CMC Taxonomy levels to disconnection is useful because studies have so far relied on various inconsistent, not yet integrated conceptualizations and measurements of disconnection (for overviews see Nassen et al., 2023; Radtke et al., 2022). For example, some studies are only concerned with individual social media platforms such as Instagram (e.g., Jorge, 2019), while others focus on certain social media features or mobile devices (e.g., Schmuck, 2020). As reviews have noted, this vastly diverging conceptualization and operationalization may be one explanation for inconsistent findings on the effectiveness of digital disconnection (Radtke et al., 2022). Systematically considering disconnection across all levels of the CMC Taxonomy should mitigate issues with comparability, generalizability, longevity of research, and helps account for the diverse uses of mobile phones in particular (Ellison & boyd, 2013; Flanagin, 2020).

Finally, the definition of digital disconnection considers a temporal dimension and assumes disconnection behaviors to have a start and an end. Crucially, this distinguishes disconnection behaviors from technology use discontinuation, which describes cases of complete cessation of usage (e.g., permanently leaving Facebook). In contrast, the definition acknowledges that everyday disconnection practices take place in temporal cycles and depend on situational factors (Vanden Abeele et al., 2024). A situational perspective may enable understanding of within-person changes of disconnection over time that go beyond between-person differences addressed previously in experimental or cross-sectional studies. To observe disconnection practices as they naturally occur in daily life can also mitigate validity issues that experimental studies face when requiring participants to adhere to multiple-day “detox” interventions (Radtke et al., 2022). Instead, this research focusses on digital disconnection in daily life and zooms in on the situational variation in both motivations for and practices of disconnection.

Why we disconnect

Successful self-regulation in general and of digital media use specifically depends on motivation (Reinecke et al., 2022; Werner & Milyavskaya, 2019). Prior research shows that motivations have a stable component, but their activation and relevance for behavioral guidance are typically highly dynamic and vary based on situational goals and circumstances (Fazio, 1990; Judge et al., 2014). This means, motivations can be relatively stable across situations (i.e., trait motivations) or vary situationally (i.e., state motivations; Bäulke et al., 2021; Wasserman & Wasserman, 2020). Thus, we argue it is crucial to understand whether the strength of a certain motivation in the moment leads to more disconnective behavior.

In their recent review, Nassen et al. (2023) distill six categories of motivations present in the disconnection literature, namely perceived overuse, social interactions, psychological well-being, productivity, privacy, and perceived usefulness. However, the situational variation of these motivations was not explicitly addressed and we therefore discuss them in the following. For investigating motivations in everyday life, we seek to focus on motivations that are likely to show some situational variation.

Relatively stable motivations

Perceived digital overuse describes the experience when digital media use “surpasses an individual standard” (Büchi et al., 2019, p. 2). As perceived overuse constitutes an “accumulated, abstracted consequence” of media use (Büchi et al., 2019, p. 2), a motivation to reduce overuse is likely a cross-situationally held, relatively stable view that follows the reflection on (over-)use of digital media. This is also demonstrated by the finding that the perception of digital overuse is commonplace across individuals (Gui & Büchi, 2021; Vanden Abeele et al., 2022).

Similarly, perceptions of usefulness as well as concerns over privacy can be considered cross-situational and relatively stable because they are centered on comparatively slowly changing (socio-)technological characteristics and affordances of digital media (Bayer et al., 2020). For example, privacy was the most prevalent motivation for permanent account deletion for former Facebook users (Stieger et al., 2013). Hence, motivations based on usefulness and privacy perceptions are aggregated over time and are unlikely to vary much from moment to moment (e.g., Tsay-Vogel et al., 2018).

Situationally varying motivations

Yet, it is underexplored how motivations that vary situationally (i.e., over short time frames) relate to disconnective behavior. While acknowledging the importance of additional motivations identified in the disconnection literature, this study thus zooms in on motivations that are likely to matter on a situation-by-situation basis. The three remaining motivation categories outlined by Nassen and colleagues (2023), targeting productivity, psychological well-being, and social interactions, likely vary substantially in everyday life. From this analysis, we deduce three situationally relevant motivations, namely, (a) to avoid distractions, (b) to improve well-being, and (c) to be more present in the moment.

Avoiding distractions

Productivity is typically defined as a ratio between resource input and desired output (Jensen & Voordt, 2017). For example, in the workplace this translates to the ratio between worked hours and completed tasks (e.g., Schmitt et al., 2021). As such, productivity is a cumulative measure. However, the overall ratio of input to output depends on situational factors such as external and internal distractions. Indeed, distractions are central for productivity (e.g., Aaron & Lipton, 2018). Additionally, distraction is a relatively accessible and specific phenomenon compared to more complex aggregated phenomena like productivity. In situational designs such as ESM research, it can be beneficial to assess specific states (e.g., distraction) compared to abstract concepts (e.g., productivity) as specific experiences may be more accessible to participants, reduce response burden, and increase validity (Gabriel et al., 2019).

Essentially, digital technologies are frequently viewed as disruptive in everyday life, distracting from important activities and tasks like social interactions, work, and studying (Nassen et al., 2023; Nguyen et al., 2022; Russo et al., 2019). Repeatedly giving in to digital distractions likely leads to negative consequences, such as task delay and attentional shifts (e.g., procrastination and multitasking; Carrier et al., 2015; Wiwatowska et al., 2023). To mitigate the interference of digital media with other tasks and activities, many thus try to reduce the salience of digital media in daily life by disconnecting. For instance, users may be motivated to avoid distractions by reducing app notifications or by avoiding certain tempting social media content, such as entertaining short-form videos. As Nguyen and colleagues (2022) note, avoiding distractions is the most reported motivation for digital disconnection. Russo et al. (2019) analyzed LinkedIn comments and found that many users aimed to disconnect to avoid distractions, for example, by using “technological solutionist” (Kuntsman & Miyake, 2022, p. 10) approaches like the do-not-disturb mode that limits notifications.

Therefore, we hypothesize that the motivation to avoid distractions makes it more likely that people employ digital disconnection practices. Specifically, we assume this hypothesis holds both within and between individuals.

H1: A higher motivation to avoid distractions is associated with more disconnective behavior (a) within participants and (b) between participants¹.

Improving well-being

Nassen and colleagues (2023) highlight that the desire to improve well-being is one of the central motivations for disconnection. Digital technologies, and particularly mobile and social media, have been repeatedly viewed as drivers of deteriorating mental health. Whether this view is accurate is intensely debated (see Meier & Reinecke, 2021), yet, beliefs that social media are “bad for you” (Ernala et al., 2022, p. 4) or “ruining our lives” (for a critical discussion see Appel et al., 2020, p. 60) are frequently expressed. For example, digital media are often perceived as stressors in daily life (Wolfers & Utz, 2022), a phenomenon so ubiquitous that it has been coined “digital stress” (Reinecke et al., 2017, p. 90). Disconnection is, furthermore, often motivated by associating (certain) digital technologies such as social media or games with negative effects for personal mental health, or by assuming that one’s well-being may improve when one lives a more technology-free life (see Nassen et al., 2023; Nguyen et al., 2022). Accordingly, users may be motivated to disconnect to improve their well-being both in general and in the here and now (Nguyen et al., 2022; Vanden Abeele, 2021). Because effects of digital media on well-being vary in everyday life (Marciano et al., 2022) and constructs like wanting to feel better in the moment are tangible and specific in everyday situations (Gabriel et al., 2019), well-being likely constitutes an important situational motivation for digital disconnection:

H2: A higher motivation to improve well-being is associated with more disconnective behavior (a) within participants and (b) between participants.

Being present offline

The final motivational category proposed by Nassen and colleagues (2023) relates to social interactions. Social reasons may motivate individuals to be present in face-to-face situations, rather than engage in digital media use (Nguyen, 2023). Disconnective behavior for more presence in the situation may target a range of social outcomes, for instance, adhering to disconnection norms (Geber et al., 2023), wanting to relate to others (Deci & Ryan, 2012; Halfmann, 2021), and preferring unmediated connection (Agai, 2022). We view the motivation to be present offline as central to social aspects of digital disconnection because it highlights instances when face-to-face connection may be more relevant than digital connection (see Nguyen et al., 2022).

Being present in the moment is an inherently situational motivation; many discussions of disconnection uncover a frequently desired notion of an unmediated experience that is lived “in the moment” (Syvertsen & Enli, 2020, p. 1274). Individuals want to live authentically, and some believe this to be hampered by digital media (Lomborg & Ytre-Arne, 2021; Meier & Reinecke, 2023). Being connected to others via messengers or to content on social platforms may come with a sense of being disconnected from non-mediated, face-to-face interactions (e.g., McDaniel et al., 2021; Thomas et al., 2022). For instance, individuals who engage in digital media use during offline activities may be more focused on online interactions than offline surroundings, sometimes leading to interference with social interactions such as phubbing (Nuñez & Radtke, 2023). Relatedly, individuals frequently report that they disconnect because they do not want to “displace important […] offline activities” (Nguyen et al., 2022, p. 9). This leaves many with a desire to reduce technology use to be more present offline (e.g., Enli & Syvertsen, 2021; Nguyen et al., 2022). Empirically, Nguyen et al. (2022) find that being present offline was the second most relevant motivation for digital disconnection after the motivation to avoid distractions. Therefore, we hypothesize:

H3: A higher presence motivation is associated with more disconnective behavior (a) within participants and (b) between participants.

Finally, to assess which of the three motivations outlined are more relevant across the levels of digital disconnection, we will explore which motivations predict disconnection from five levels of the CMC Taxonomy: devices, applications, features, interactions, and message content (RQ1).

Goal proximity of disconnection motivations

One central psychological mechanism that can influence whether a motivation for digital disconnection is enacted is temporal goal proximity (Bandura & Schunk, 1981). Digital disconnection entails behavior that is oriented toward a goal, i.e., to reduce digital media use or certain aspects of it (Nguyen & Hargittai, 2024). Goals are central to motivations, as goal characteristics influence motivations and drive how motivated behavior is enacted (Louro et al., 2007; Wallace & Etkin, 2018). Crucially, goals vary in their proximity, meaning how close in time, abstract, and complex a goal is (Lee & Bong, 2019). More proximal goals are achievable soon, are concrete, and consist of only a small number of steps. Drawing on insights from psychological research on goal setting, more proximal goals should be easier to achieve, as their results are more immediately perceptible, and thus motivations are more likely to be enacted (Lee & Bong, 2019; Locke & Latham, 2002). Distal or higher-order goals, in contrast, are relevant to sustain motivation over longer periods of time (Lee & Bong, 2019; Locke & Latham, 2002).

For momentary digital disconnection, both proximal and more distal goals are likely relevant to motivate disconnective behaviors. For example, improving well-being may be a motivation with a comparatively distal goal, as it could stem from the perception of multiple, sometimes unavoidable impacts of digital technology use that unfold over time, such as stress due to many emails. In contrast, avoiding distractions may be more proximal, as it is linked more directly to specific aspects of digital technologies such as app notifications as well as to more concrete tasks that might compete with digital media use.

In the context of outcomes of digital media use more generally, the relevance of outcome proximity has been considered, such as for well-being (Büchi & Hargittai, 2022). Yet, to our knowledge, the role of temporal goal proximity has not yet been explored for (motivations of) digital disconnection. We suggest that an understanding of how motivations relate to digital disconnection across time depending on goal proximity could elucidate the psychological processes that may lead to disconnection. Motivation toward more long-term vs. immediate outcomes could be one such mechanism that explains temporal patterns in disconnective behavior. Therefore, we will explore how motivations to disconnect are enacted across time (RQ2).

Method

To answer the research questions and test hypotheses, a two-week experience sampling study was conducted from November 2022 to March 2023 with up to five daily measurements at quasi-random intervals. The study is preregistered at https://osf.io/285hs.

Procedure

Participants were recruited via the online access SoSci Panel (Leiner, 2016) and through two email lists of students enrolled at two German universities. After filling out a pre-survey, participants installed an Android-only smartphone app (movisensXS) which delivered ESM probes. Daily measurements started on the day after participants had finished the pre-survey. Participants received five questionnaires per day at random times from 9am to 9pm at least 2:15 hours apart for 14 consecutive days (i.e., max. 70 probes). Questionnaires could be accessed for 15 minutes after the initial notification. Participants who filled out at least 60% of the probes were rewarded with a 15€ voucher. As additional incentives, participants could optionally receive a day-level summary of their ESM responses after study completion and a sample-level summary of overall results once the study was concluded. We set the cutoff for compliance relatively low at 60% compared to similar studies (Schnauber-Stockmann & Karnowski, 2020) for conceptual reasons, namely to avoid interference with mobile (non-)usage behavior or habits. Participants were further instructed that they should not change their disconnective behavior to comply with the study, e.g., to keep notifications off when in do-not-disturb mode or when they were not using a device.

To further reduce interference with digital disconnection behaviors as well as intrusiveness of our probes, we chose to ask questions about the last 2 hours, a somewhat longer timeframe than typical for ESM studies on media and communication topics (Schnauber-Stockmann & Karnowski, 2020). Participants who responded to a probe were asked whether they disconnected at any levels of the CMC Taxonomy within the last 2 hours. If so, they could respond to rating questions about why they disconnected. If they did not disconnect, items on need satisfaction were displayed instead to keep the length of the questionnaires consistent regardless of response pattern.

Participants

According to the preregistered inclusion criteria, we recruited Android users aged 18 to 35. We set the age limit for participation to 35 because studies suggest that digital media use and disconnection are more central to younger users (see Beisch & Koch, 2022; Oksa et al., 2021). Additionally, a more homogenous sample regarding key characteristics—such as deliberate disconnection from digital media—helps reduce the influence of potential confounds and thus improves a central aspect of data quality (Bornstein et al., 2013).

We determined our minimum sample size, first, by setting it higher than typical studies in communication research at both the person and the situation level (i.e., Mdn = 79 individuals and Mdn = 1,918 situations, Schnauber-Stockmann & Karnowski, 2020). Like similar ESM studies (e.g., Johannes et al., 2021), we determined the desired maximum sample size of 242 participants based on our available funding resources, which was deemed acceptable given the relative lack of prior research on situational motivations for disconnection. Two hundred forty-eight participants reached the compliance threshold of completing at least 60% of probes. We use this compliance threshold for all following confirmatory and exploratory analyses except for the exploration of time lag because missing pairs across time substantially reduce the number of observations (see Table 3). After filtering out participants and probes who did not meet our preregistered inclusion criteria, the final sample consists of N = 230 individuals who answered a total of T = 12,160 probes. In T = 7,360 of these situations, participants reported any form of digital disconnection within the last two hours. Post-hoc sensitivity analyses show that the study is likely adequately powered (see Supplementary material D, https://osf.io/486aq).

On average, participants were 25.31 years old (SD = 4.50). Out of 230 participants in total, 77 participants (32%) identified as men, 156 as women (66%), and four as non-binary or another gender (1%). Forty-four percent of participants had a high school degree or equivalent, 8% a vocational education, 26% a bachelor’s degree, 20% a master’s degree or equivalent, and 2% reported a different type of education. Additional sample characteristics are reported in Supplementary material A in the OSF.

Measures

All situational ESM and pre-survey items can be found in the OSF and the preregistration contains a description of all additional variables that were collected as part of a larger project.

Digital disconnection was measured using a self-developed scale that was quantitatively pilot-tested (N = 15) and assessed qualitatively with the think-aloud method (N = 8; Wolcott & Lobczowski, 2021). The measure consists of five items about digital disconnection at five levels of the CMC Taxonomy (Meier & Reinecke, 2021): device, application, feature, interaction, and message content. We focused on the content facet of the message level of the CMC Taxonomy because we assume that it is central to digital disconnection due to the relevance of digital media content described in previous research, for example, unrealistic body imagery, hate speech, or negative news (e.g., Vandenbosch et al., 2022). The two application levels in the CMC Taxonomy (type of application, branded application) were combined for easier understanding and parsimony. Each item could be answered dichotomously, with yes (1), indicating that the participant disconnected at this level, or no (0), indicating that disconnection did not occur at this level. If participants did not disconnect at any level, they could indicate so on a sixth item. We calculated the measure for disconnective behaviors as the mean across all possible levels per measurement occasion to get an index of average disconnective behavior in a situation. Internal consistency of the measure was good (α_within = .83, α_between = .91, ω_within = .83, ω_between = .93).

Motivations to disconnect were assessed as the three motivations to avoid distractions, improve well-being, and be more present based on items adapted from Nguyen et al. (2022) and motivations as conceptualized in Nassen et al. (2023). If participants indicated that they disconnected, they were asked the degree to which they did so based on each of the motivations, i.e., “I wanted to be less distracted,” “I wanted to feel better, e.g., less stressed,” and “I wanted to be more present in offline life.” Each of the three questions could be answered on a Likert-type scale from 1 (do not agree at all) to 7 (completely agree).

Analytical strategy

Analyses were conducted in R (R Core Team, 2023) using the lme4 package (Bates et al., 2023). We employed linear mixed effects models and distinguished person- and situation-level variance with random effects within-between models (Bell et al., 2019). To this end, new variables were calculated for each predictor by (a) person-mean centering each predictor (within component) and (b) sample-mean centering each participant’s mean score (between component). This method allows to control the within-person effect for variance at the person level, for example, due to personality or socioeconomic status (Bell et al., 2019). We constructed null models with random intercepts on disconnective behavior as a baseline and gradually made the models more complex by introducing predictors and controls. We then interpreted the best fitting models’ predictions based on AIC, BIC, ANOVA comparisons, and R² and conducted robustness checks with control variables (Snijders & Bosker, 2012). Beta coefficients were calculated by z-standardizing all variables.

To investigate associations between the motivations and individual levels of the CMC Taxonomy (RQ1), we calculated multilevel correlation coefficients between the relevant variables. As five outcomes were assessed, interpretations were based on Bonferroni-adjusted confidence intervals (99% CIs). To explore whether situational motivations vary concerning the time point at which they are associated with disconnective behavior (RQ2), lagged predictors were introduced to the models. Because the literature does not indicate how long lags for motivations on disconnective behavior typically are, we report lags from 0 to 2 situations. If a motivation is associated with later disconnective behavior, this might indicate a lower goal proximity. We regressed the predictors at t₀ to the outcome variables of the following probe t₁. The following probe was at least 2:15 and at most 3:27 hours later, and we did not lag variables overnight. Note that for the lagged analyses, the compliance threshold (60%) was not used. Instead, all valid probes for all participants who fit inclusion criteria were utilized because introducing lags drastically reduces statistical power due to missing probes and our sampling scheme. This can be observed in Table 3, where with longer lags, the number of complete situation pairs shrinks.

Results

Descriptives

In 59% of probes (T = 7,360), participants indicated that they engaged in digital disconnection from at least one of the CMC Taxonomy levels in the previous two hours. Note that levels are not mutually exclusive. Disconnection at the device level took place in 4,489 (36%) situations, at the application level in 5,740 (46%) situations, at the feature level in 5,896 (48%) situations, at the interaction level in 6,313 (51%) situations, and at the content level in 6,458 (52%) situations. Central descriptive statistics and zero-order correlations are depicted in Table 1.

Situational motivations and digital disconnection

Distraction motivation

To assess whether the motivation to avoid distractions significantly increases the degree to which participants engage in disconnective behaviors (H1), we first fit a null model. Including the motivation to avoid distractions as a predictor improves the model by all preregistered fit criteria. Adding random slopes further improves fit, indicating within-person variation. The coefficients show that the motivation to avoid distractions at the within-person level (β = .10, b = 0.02, 95% CI [0.02–0.02], t(148.5) = 6.92, p < .001) and between-person level (β = .09, b = 0.02, 95% CI [0.00–0.05], t(247.9) = 2.10, p = .036) significantly predicts more disconnection. However, when controlling for age and gender, the between-person effect does not explain a significant amount of variance, whereas the effect for within-person motivation to avoid distractions remains robust (see Table 2). Thus, H1a is accepted and H1b is rejected.

Well-being motivation

The second hypothesis that the motivation to improve well-being would increase disconnective behavior (H2) was tested using the same procedure as for the previous hypothesis. We first specified the null model. Adding the sample-mean centered and the person-mean centered well-being motivation variable improved model fit by all criteria, except BIC. A model with random slopes demonstrates improved model fit. The between-person component of the well-being motivation significantly predicts higher disconnection (β = .12, b = 0.02, 95% CI [0.01–0.04], t(226.9) = 2.63, p = .009,), whereas the within-person component does not (β  =  0.01, b = 0.00, 95% CI [−0.00–0.01], t(161.9) = 0.43, p = .670). However, when including age and gender as control variables, the association of the between-person component and disconnection is no longer significant (see Table 2). Thus, neither H2a nor H2b are accepted.

Presence motivation

Finally, the hypothesis that the motivation to be present would increase disconnective behavior (H3) is assessed. Adding the presence motivation predictors and allowing slopes to vary improves model fit by all preregistered criteria. Both the within-person (β = .14, b = 0.02, 95% CI [0.02–0.03], t(150) = 9.2, p < .001) and between-person component (β = .09, b = 0.02, 95% CI [0.00–0.04], t(246.6) = 2.63, p = .028) of the motivation to be more present predict more disconnective behavior. While this effect of presence motivation on digital disconnection stays robust at the within-person level, the between-person effect disappears when controlling for age and gender (see Table 2). Therefore, H3a is accepted and H3b is rejected.

Simultaneously including all three motivations—distraction, well-being, and presence—at the within- and between-person level and their random slopes in one model shows that the within-person effects of distraction motivation and presence motivation remain significant when accounting for the other motivations in this more complex model (see Figure 1 and Table 2). As an exploratory robustness check, we included the extent of digital media use for leisure or for work, time since study start, time of day, and weekend vs. weekday in the most complex model (see Model 3 in Table 2). The effects of the situational distraction and presence motivations remain robust when accounting for these controls.

Exploration: situational motivations and disconnection levels

For exploratory purposes, we assess how the three motivations are associated with different levels of digital disconnection along the CMC Taxonomy. Before doing so, we analyze associations between the levels of digital disconnection. As Figure 2 shows, generally, disconnection levels are more closely related to neighboring levels than to more distant ones both within and between participants. For example, situationally, the content level associations are the strongest with the neighboring interaction and feature levels compared to weaker associations with the app level and no significant associations with the device level.

To explore how the three motivations relate to each individual level of disconnection along the CMC Taxonomy (RQ1), we report multilevel correlations for each of the levels of disconnection separately with the three motivations. Correlation coefficients for all statistically significant associations are reported in Figure 3.

For example, Figure 3 shows that between persons, the motivation to avoid distractions appears to be correlated with disconnection at the device and application level. Disconnection at the content level is correlated with well-being motivation between and within participants. The overall strongest associations seem to be among presence motivation and disconnection at the device level, followed by the distraction motivation and the application level.

Exploration: disconnection motivations across the day

To explore how the enactment of disconnection motivations varies over time (RQ2), multilevel regression models with time lags were constructed. Because using lags leads to substantially fewer measurement occasions, for this exploratory analysis only, we chose to include observations regardless of the 60% compliance threshold from all otherwise valid observations. Table 3 shows the coefficients for the lagged predictors. Within particiants, the motivations to be more present (β = .03, b = 0.01, 95% CI [0.00–0.01], t(5,254) = 2.90, p = .004),to improve well-being (β = .03, b = 0.01, 95% CI [0.00–0.01], t(5,240) = 3.28, p = .001), and to avoid distractions (β = .04, b = 0.01, 95% CI [0.00–0.01], t(5,241) = 3.74, p < .001) significantly predict more digital disconnection in the next situation. This means that when participants indicated that they were motivated to disconnect for presence, distraction, or well-being reasons, they were engaging in slightly more disconnective behaviors around two to three hours later. When using only probes from participants with at least a 60% compliance rate, the within-person associations remain robust.

Discussion

We show that enacting digital disconnection behaviors in a situation depends on how motivated individuals are to avoid distractions and to be present offline (H1a, H3a). While within-person standardized coefficients of β = .10 to .14 are considered small effects, they are above those typically found in much of CMC and media effects research (e.g., Meier & Reinecke, 2021). Within a situation, the motivation to improve well-being was not associated with more disconnective behavior (H2a). None of the motivations included in this study explained disconnection behavior on the person level, at least not beyond the control variables age and gender (H1b, H2b, H3b). The ICCs further show that the majority of variance in digital disconnection can be explained on the situation level. Overall, these findings lend support to the proposed situational, everyday perspective on digital disconnection and its motivations. This study builds on and extends previous research in which digital disconnection was mostly viewed as a phenomenon varying between individuals best captured through experimental or cross-sectional analyses (for overview and analysis see Fast, 2021; Nassen et al., 2023).

Situational motivations to disconnect

Our study supports the assumption that a relevant motivation for digital disconnection is the motivation to be more present. This reflects the common notion that disconnection allows people to enjoy unmediated, face-to-face communication (e.g., Syvertsen & Enli, 2020). Abstaining from using digital media, especially at the device level (see Figure 3), could be a useful tool to achieve this. However, our analyses show that this motivation is not significantly related to digital disconnection at the person level; wanting to be present in the moment may not be an effective trait-like motivation but rather vary situationally. This demonstrates that a situational perspective on the motivation to be present is important for the assessment of digital disconnection behavior (Funder, 2006).

Similarly, the motivation to avoid distractions was found to be predictive of digital disconnection behavior within persons, consistent with previous work that suggests digital disconnection may be used to avoid distractions and interruptions through digital media (e.g., Fast, 2021; Nassen et al., 2023). On the other hand, the distraction motivation did not explain differences in disconnection at the between-person level. However, exploratory analyses on the different disconnection levels document that participants who generally aimed to avoid distractions more were also more inclined to disconnect specifically from entire devices or applications. These associations extend to the situational level; avoiding distractions could be a relevant motive to disconnect situationally, most likely from devices and applications, and certain features such as notifications. Still, future confirmatory analyses of these exploratory findings are needed.

Finally, the motivation to improve well-being was not associated with more disconnection, either between or within individuals. This is somewhat surprising, considering that disconnecting is frequently viewed and suggested as a solution to perceptions of reduced well-being due to (negative forms of) digital media use (Nguyen et al., 2022; Radtke et al., 2022). Even though the exploratory lagged multilevel analyses should be interpreted cautiously due to the lack of robustness of the associations, they give an initial indication that well-being may be a more distal goal of digital disconnection, again underlining the value of a situational approach.

Disconnecting at all levels of the CMC Taxonomy

Previous research did not use a systematic and comprehensive approach to measuring digital disconnection (Nassen et al., 2023). Thus, this study applied the Hierarchical CMC Taxonomy (Meier & Reinecke, 2021) to systematically measure digital disconnection from the device to the message content level. The proposed hierarchical structure of CMC levels is indeed reflected in the present study, as both between and within persons, disconnecting was more closely associated among neighboring levels. Distinguishing between levels of digital disconnection additionally contributed to a clearer picture of the links between the different motivations and digital disconnection. Importantly, the exploratory analyses showed that wanting to increase well-being may be more strongly associated with disconnection from content and interactions, rather than from entire devices or apps per se. While internal consistency for the measure was good, additional validation of our newly proposed disconnection measure is needed.

While neighboring levels were more strongly associated, it is possible that individuals disconnect from certain aspects of digital media but replace reduction in one area with increased usage in another. For example, instead of scrolling through social media on the smartphone, they might use a laptop instead. Disentangling the intricacies of disconnecting appears essential for understanding the processes and effects of digital disconnection. Toward this end, for instance, unobtrusive digital trace data combined with survey data could enable new insights.

Motivations and digital disconnection across time

The exploratory findings on lagged disconnective behavior show that while the well-being motivation was not predictive of disconnection in the same situation, it indeed was in the following situation. Yet, the presence motivation might be relevant for disconnection both in the same and the next situation, whereas the distraction motivation may be relevant within the same and in the following two situations. Hence, the proximity of goals may be relevant to when and for how long participants enact a motivation to disconnect, underscoring the complexity of disconnection across time (Locke & Latham, 2002).

For the exploratory analysis of goal proximity with time lags, all valid probes were included due to the accumulation of missing values with time lags and because we weighed statistical power higher than compliance with the ESM protocol for exploratory purposes. Future studies should compensate for missing data due to lags, for example, by including more daily assessments. Follow-up studies could build on these exploratory results by directly measuring or manipulating (temporal) goal proximity.

Limitations and future directions

Measuring non-use of digital technologies with a digital technology may have undesirable side effects, such as measurement reactivity or increased smartphone use due to the study itself. While we have taken steps to reduce this influence, the mobile ESM approach may still have biased our results. However, as the goal was to study short-term digital disconnection in everyday life, the smartphone offers the least intrusive and most available method to assess these behaviors. Further, incentivizing participants to fill out at least 60% of probes may have influenced participants’ behavior. However, this study tried to reduce the potential for interference by setting the cutoff for incentivization lower than usual. By setting the cutoff for compliance at 60% we may have excluded participants who were disconnecting frequently and predominantly at higher levels, such as from their smartphone, specifically. Yet, crucially, our findings remain stable when including all otherwise valid probes, underlining the robustness of the insights. Still, due to the focus on situational digital disconnection, situations from users who refrain from using their smartphone frequently and for long durations are likely under-represented.

Our study assumed that individuals would want to disconnect at least sometimes. However, it should be noted that in some situations individuals may also be motivated to actively seek connections through digital technologies and, therefore, not wish to disconnect (Hampton & Shin, 2022). Some goals could thus lead to more connection, for example, the goal to increase productivity may encourage the use of enterprise social media (e.g., Klingelhoefer & Meier, 2023). Hence, future research should explore in which situations individuals are motivated to seek connection over disconnection, and vice versa.

Within the study, several choices were made to focus on specific aspects of digital disconnection, which may neglect other relevant areas. For example, we focused on younger users, limiting generalizability to this specific group. As we presume that disconnection and its motivations are different across age groups (e.g., Beisch & Koch, 2022), older adults’ and adolescents’ disconnective behavior may be driven by different factors, such as technical barriers (Nguyen et al., 2021). Future studies could investigate whether these everyday motivations are similar for various age groups, as disconnection is likely relevant for older adults and adolescents as well (see Nguyen et al., 2021), but may play out differently in daily life.

Participants were recruited using convenience sampling. While data quality within the non-commercial SoSci panel is sufficiently high (Leiner, 2016), our sample is not representative of characteristics at the country level; for instance, participants were highly educated. The focus on a more homogenous group of young adults has the advantage of improving statistical inference, but limits generalizability to other groups (Bornstein et al., 2013). Furthermore, sampling was likely affected by some level of self-selection and the technical limitation of the ESM app to Android (Nguyen & Hargittai, 2024). However, because our study focuses on within-person associations in everyday life and sampled at quasi-random intervals, it likely represents a valid picture of participants’ everyday disconnection motivations and behaviors.

Previous research identified more than the three motivations selected for this study (Nassen et al., 2023). Our focus was on three situationally varying motivations but additional situational motivations could be relevant or more stable motivations could interact with situational characteristics (Wasserman & Wasserman, 2020). Future studies could assess how additional motivations may be relevant for situational digital disconnection and examine more distal goals across longer time frames. Moreover, a promising future direction is to consider additional situational factors, such as measuring disconnection as well as experiences of connection (Nguyen & Hargittai, 2024). Overall, exploring the interplay between person-level and situational aspects of disconnection and its motivations appears to be a fruitful new avenue for future research that seeks to understand why, when, and to what effect users disconnect from smartphones, social media, and similar digital technologies.

Data availability

The data underlying this article are available in the OSF at https://osf.io/486aq/.

Funding

The authors declare no funding.

Conflicts of interest: The authors declare that there is no conflict of interest.

Notes

References