Social networking sites (SNS) are regularly used by billions of people worldwide (Facebook, 2021; Tencent Global, 2022; Twitter, 2021). Reflecting the feature-rich nature of these platforms (Tarafdar et al., 2020), prominent motives for using SNS are diverse and include meeting social needs, seeking entertainment and information gathering (Krasnova et al., 2017; Ku et al., 2013; Kuss & Griffiths, 2011; Lai & Yang, 2014; Pertegal et al., 2019). For the minority of users who may be considered at risk of SNS addiction, another motive has been identified: mood modification (Griffiths et al., 2014; Kuss & Griffiths, 2011).

Mood modification refers to the way in which addictive substances and behaviours may be turned to as part of a dysfunctional coping strategy to regulate emotional states (Griffiths, 2005). It features in negative-reinforcement theories of addiction (Baker et al., 2004; Koob, 2013; Koob & Le Moal, 2008) and in two prominent, models frequently applied to technology addictions, the Components Model of Addiction (Griffiths, 2005) and the Interaction of Person-Affect-Cognition-Execution (I-PACE) model of internet-based addictions (Brand et al., 2016, 2019). The components model, which builds on the work of Brown (1993) and was evidenced through Griffiths’ work on gambling disorder, suggests that all addictions share the following features: salience, mood modification, tolerance, withdrawal, conflict and relapse (Griffiths, 2005). In studies of uncodified behavioural addictions, risk of addiction is often identified through the use of self-report questionnaires that build upon this model (Andreassen et al., 2012; Böthe et al., 2018; Carbonell & Panova, 2017; Fuster et al., 2016; Mennig et al., 2020; Pontes et al., 2014). The I-PACE model, on the other hand, focusses on the interplay between cue reactivity, executive function and predisposing factors (Brand et al., 2016, 2019) and is often presented as a theoretical foundation for experimental behavioural addiction research (Antons & Brand, 2018; Brand et al., 2019; Guo et al., 2024; Pekal et al., 2018; Tikhonov et al., 2024; Trotzke et al., 2020). Mood modification as an aspect of both of these models reflects a broader understanding of addiction as a “goal-directed choice under negative affect” (Hogarth, 2020, p.1). Although there are a number of motives to engage in addiction activities, dysphoric mood is considered a reliable trigger for the enhancement of subjective feelings of craving and, ultimately, engagement in an addiction behaviour (Field & Cox, 2008; Field & Powell, 2007). This effect may have its foundations in outcome expectancy (Goldman et al., 1999) as well as the associations between addiction and deficits in both emotional regulation and inhibition (Brand et al., 2016). In states of dysphoric mood, individuals who already struggle with emotional regulation and inhibitory control are more likely to crave and then engage in activities that they associate with mood change as an emotional coping strategy (Brand et al., 2016, 2019). Although inhibition is a core executive function essential to everyday life, in the context of addiction, deficits in inhibition are considered a risk factor for excessive engagement in addictive behaviours and failed attempts to control usage (Smith et al., 2014). In line with codified addictions (Estevez et al., 2017; Feil et al., 2010; Ivanov et al., 2008), risk of SNS addiction is associated with greater deficits in emotional regulation (Drach et al., 2021; Hormes et al., 2014; Liu & Ma, 2019) and inhibitory control (He et al., 2021; Reed, 2023). Moreover, like gambling and other proposed behavioural addictions, proneness to boredom has been identified as a risk factor for the development and maintenance of SNS addiction (Blaszczynski, 1990; Camerini et al., 2023; Chaney & Chang, 2005; Wegmann et al., 2018; Yang et al., 2020). Camerini et al. (2023) suggest that that individuals with greater boredom proneness may utilise SNS in order to alleviate boredom, which, over time, creates outcome expectancy and a habitual pattern of engagement that may ultimately result in a reliance upon SNS as a coping mechanism for boredom (Camerini et al., 2023). As addiction progresses, gratification from digital media use may wane while compensatory use grows alongside reduced inhibitory control and heightened craving (Brand et al., 2016, 2019).

SNS addiction is not, however, an addiction that has been formally codified by either the World Health Organisation (2019) or the American Psychiatric Association (2013). Regardless, it has been and is the subject of legislation, lawsuits and other legal actions, some of which, if implemented, could impact the accessibility of SNS for everyday users (Consortium Plaintiffs vs. Meta, Snap, ByteDance & Google, 2023; European Commission, 2024; Klobuchar, 2022; SMART Act, 2019; Social Media Use for Minors, 2024). As such, it is imperative that researchers provide evidence for each component of addiction as they may (or may not) manifest for this emerging disorder.

To date, relatively few studies have examined the use of SNS as a mood modifier, and findings are not uniform. For example, Sagioglou and Greitemeyer (2014) found that users may expect that using Facebook would improve their affective state, but, when measured, affective state often becomes more negative following usage. Negative outcomes of SNS use were also identified in the work of Boursier et al. (2020), who found that loneliness predicts excessive SNS use, but the outcome of SNS use is often increased anxiety. Conversely, Drach et al. (2021) found positive affect (PA) as well as negative affect (NA) to be associated with decreased subjective urges to use SNS, while usage was found to result in increased PA. The above studies did not, however, account for the possible role of addiction in the affective motivation and outcomes of SNS use.

Not surprisingly, time spent on SNS (and smartphones) predicts the risk of SNS addiction (Hong et al., 2014; Marengo et al., 2022), and that time is predicted by users’ positive affective forecast of SNS use (Baum & Baumann, 2021). When put into the context of addictive behaviours, such affective motivation may be more important than affective outcome. In their (2020) study, Tarafdar et al. present a theoretical link between SNS stress and SNS addiction in which users turn to SNS as a distraction technique to cope with stress even when that stress is brought on through SNS use. In a similar cycle of usage centred around NA, trait boredom has been found to predict problematic SNS use, which, in turn, predicts enhanced situational boredom (Donati et al., 2022). While these studies demonstrate how states of NA may lead to SNS use among problematic users, they do not empirically address the full experiential pattern of dysphoric mood, followed by enhanced craving, followed by usage that is typical of addiction (Field & Cox, 2008; Field & Powell, 2007).

In examining mood modification as a component of SNS addiction, the present study tests whether the experience of boredom results in greater feelings of subjective craving, lowered behavioural inhibition in the face of SNS-related cues and ultimately SNS use. This is achieved by utilising the Go/No-Go paradigm and creating situations of boredom. The Go/No-Go task, which is often employed as a measure of behavioural inhibition, presents participants with two cues and asks participants to respond to one cue but not the other. Reinforcing the effectiveness of this task as a measure of inhibitory control, fMRI data has demonstrated heightened activity in the anterior cingulate cortex (ACC), the area of the brain associated with inhibition, in No-Go trials (Braver et al., 2001). In studies applying the Go/No-Go paradigm using neutral stimuli, for example, letters or numbers, differences in inhibitory control have been identified between the general population and individuals facing a range of addictions, including internet addiction (Dong et al., 2010; Kamarajan et al., 2005; Yin et al., 2016), which may reflect broader deficits in inhibitory control among individuals at risk of addiction, but when a combination of neutral and addiction-related cues are employed, cue-specific differences in ACC response and behavioural inhibition within the task can be identified (Detandt et al., 2017; Kreusch et al., 2014; Luijten et al., 2016). This strengthening of ACC response and behavioural differences reflects how, through the addiction process, a relationship forms between addiction-related cues and executive functioning (Dawe et al., 2004; Field & Cox, 2008; Franken, 2003). Yet, where the Go/No-Go task has been used for studying SNS addiction, findings, to date, have not shown significant behavioural differences in behavioural inhibition responses to SNS-related cues (Gao et al., 2017; Moretta & Buodo, 2021a; Turel et al., 2014; Wegmann et al., 2020). However, data from associated ERP measurements indicate that, compared to more typical users, problematic users demonstrate greater attentional allocation to SNS-related cues (Gao et al., 2017). Additionally, reduced behavioural inhibition has been identified when participants are presented with natural and SNS-related rewards (Moretta & Buodo, 2021a), demonstrating potential for the use of this paradigm.

An additional consideration within this study will be any potential impact of style of SNS usage. SNS usage can be split into active and passive styles. Those using an active style, for example, liking, sharing and commenting, may spend more time on SNS than passive users (Quiroz & Michelson, 2021). However, research concerning the relationship between style of usage and risk of SNS addiction is limited and contradictory, with some studies indicating that active usage is associated with SNS addiction (Brailovskaia & Margraf, 2022) and other studies reporting the opposite (Fioravanti, 2020). As an exploratory measure in this study, differences related to style of use were examined for all hypotheses. In total, four hypotheses were tested:

  1. (1)

    Following a boredom manipulation, those at greatest risk of SNS addiction will report greater craving for SNS compared to those at lesser risk of SNS addiction and those not assigned to the boredom condition.

  2. (2)

    Following a boredom manipulation, those at greatest risk of SNS addiction will experience less inhibitory control in the face of SNS-related cues than those at lesser risk of SNS addiction and those not assigned to the boredom condition. Specifically, these participants are predicted to commit more commission errors when SNS-related cues are No-Go stimuli and have faster response times when SNS-related cues are Go stimuli.

  3. (3)

    When presented with a “real world” boring situation, i.e. waiting, those at greatest risk of SNS addiction will be more likely than other participants to use SNS.

  4. (4)

    Any use of SNS in a “real world” boring situation will result in less boredom and cravings for SNS than those who did not use SNS.

Method

Participants

In a convenience sample, first-year psychology students from the University of Warwick were invited to participate in this study in exchange for course credit. After reading the information sheet, each participant gave written informed consent before the experiment. The experimental procedure was in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the Department Psychology at the University of Warwick.

Boredom Inductions

Mood Induction Videos

Building on the work of Markey et al. (2014), boredom was induced using video (approximately five-minutes in length) of a man talking about his job, while control participants viewed a video clip of similar length taken from the BBC Earth documentary series. The job video featured a plain environment with a central character providing a detailed account of his typical day at work within a company that orders office supplies (Markey et al., 2014; Markey, 2020). The BBC Earth clip, entitled, “Beautiful Hummingbirds in Slow Motion”, featured tranquil music and scenes of hummingbirds drinking from a feeder and from a fountain along with brief commentary from the videographers involved (BBC Earth, 2017). The most frequent adjective in the comments section of this video was “beautiful” (BBC Earth, 2017).

Real World Boring Situation

At what participants believed was the end of the study, they were told that the debrief needed to be printed and were asked to wait in their closed cubicles. Each participant waited for approximately five minutes before being asked to complete an additional round of surveys and then come out of the cubicle to be given the debrief.

Measures

Social Media Craving Scale

In order to measure craving for SNS use, participants were asked to estimate their current urge to use SNS on a 10-point visual analogue scale, with 0 representing “no urge” and 10 representing “a very strong urge”. Such a measure of craving has recently been used in Wegmann et al.’s (2021) study of SNS addiction.

Modified I-PANAS-SF

Boredom was assessed as an item added to the 10-item international Positive and Negative Affect Schedule (I-PANAS-SF) (Thompson, 2007), which measures the degree to which participants report feelings of positive affect (PA), i.e. feeling active, alert, attentive, determined and inspired, and feelings of negative affect (NA), i.e., feeling afraid, ashamed, hostile, nervous and upset. In the original validation of this version of the PANAS, internal consistency was found to stand at 0.80 for PA and 0.74 for NA (Thompson, 2007), and in the present study Chronbach’s alpha was ≥ 0.72 in measures of PA and ≥ 0.70 in measures of NA. As with the original PANAS (Watson et al., 1988), a five-point Likert scale measured the extent of participants’ affective state, (1 = “very slightly or not at all”, 5 = “extremely”). Beyond the added measure of boredom, the original items used within this measure were totalled to create two affective scores: one for PA and one for NA.

Bergen Social Media Addiction Survey

Risk of SNS addiction was assessed using the Bergen Social Media Addiction Survey (BSMAS), a six-item measure, which utilises a five-point Likert scale (1 = “very rarely”, 5 = “very often”) to assess participants’ relationship with SNS over the past year with items reflecting the components of addiction presented by Griffiths (2005).The BSMAS, which was first developed for Facebook addiction before being used for SNS in general (Andreassen et al., 2012, 2016) is widely used around the world, and benefits from an acceptable to very good internal consistency, α ≥ 0.75 (Dadiotis et al., 2021; Lin, et al., 2017; Monacis et al., 2017; Shin, 2022), also found in the present study α = 0.79. Utilising the scoring method proposed by Bányai et al. (2017), participants who scored ≥ 19 on this survey were considered to be at risk of addiction.

Passive Active Social Media Measure

The Passive Active Facebook Measure (PAUM) (Gerson et al., 2017) was used to assess the style of social media use. The PAUM is a validated, 13-item measure of active and passive SNS usage behaviours designed initially for Facebook users with acceptable internal consistency, α ≥ 0.70 (Gerson et al., 2017; Trifiro & Gerson, 2019). In the current study, the internal consistency was found to better for measures of active use α = 0.77 than for passive use α = 0.61. Using a five-point scale, the PAUM measures how often users engage in active social, active non-social and passive usage behaviours. For this study, which includes users of a diverse range of SNS applications, the scale was modified by replacing the term “Facebook” with the term “social media”. In analysing responses, participants were categorised as a more active or more passive users by subtracting mean scores for items related to passive usage from mean scores for items related to active usage before creating a median split between all participants with the lower half of users being the more passive and the upper half being more active. The PAUM has previously been found to have acceptable internal consistency, α ≥ 0.70 (Gerson et al., 2017; Trifiro & Gerson, 2019), which was also found in the current study measures of active α = 0.77 and passive α = 0.61 use.

SNS Use Question

Using a simple yes/no question, participants were asked if they had used SNS during the waiting period.

Go/No-Go Task

Behavioural inhibition was measured using a Go/No-Go task created using PyschoPy software (Peirce et al., 2019) and featuring SNS-related imagery taken from four popularly used social networking sites in the UK, Facebook, Instagram, Tik Tok and Twitter (Ofcom, 2021). The stimuli consisted of images of a phone, 8 cm high and 4 cm wide, presented on a gray background (RGB: 128, 128, 128). The image showed a typical smartphone layout with sixteen blue boxes (RGB: 91, 155, 213) representing apps and a multi-coloured background (see Fig. 1A for an example display). One box contained either the SNS-related image or a control image (see Fig. 1B).

Fig. 1
figure 1

Example Visual Stimuli (A) and Other Matched Stimuli Appearing in the Display (B)

Full size image

At the start of the Go/No-Go task, participants were asked by the experimenter to select the app that they use the most out of Facebook, Instagram, Tik Tok and Twitter. Corresponding to this selection, they were then presented with a Go/No-Go task featuring images of a phone with the Facebook, Instagram, Tik Tok or Twitter app and control images of a phone with a different app that shared low level visual information, i.e. colour and other visual features.

Building on the work of Gao et al. (2019), each trial began with a fixation cross appearing in the centre of the screen for 200 ms. Next, a phone image appeared in the centre of the screen for 1,500 ms. To increase task difficulty, app images appeared in one of six locations on the phone. Participants were instructed to either press the up arrow key or withhold their response depending on the type of stimuli (SNS-related or control) seen on the phone. Participants completed four blocks in total with 25 trials in each block with feedback provided after each trial. Each block featured 20 Go and 5 No-Go trials. Half of the blocks featured SNS-related stimuli as the Go target with control images seen in No-Go trials, and half featured control imagery as the Go targets with SNS-related images seen in No-Go trials. Block orders (SNS Go targets followed by control Go targets or vice versa) were counterbalanced between participants. Reaction times in Go trials, commission error rates (responding in No Go trials) and omission error rates (failing to respond in Go trials) were recorded for trials with SNS-related and control stimuli.

Experiment Procedure

The experimental procedure is summarized in Fig. 2. Following the provision of a participant information sheet and the completion of a consent form, participants were asked to provide basic demographic information (age, gender and indication of visual impairment) as well as baseline self-reports of mood and craving via the modified I-PANAS-SF and the social media craving scale. Next, they watched either the boredom induction or control video before once more completing the modified I-PANAS-SF and the social media craving scale. All participants then completed the Go/No-Go task, followed by another round of completing the modified I-PANAS-SF and the social media craving scale. At this point, participants were told that the debrief needed to be printed and were instructed to wait in the small, closed room (cubicle) where they had completed the other tasks. Following a period of approximately five minutes, the researcher entered the room and informed them that the debrief had been printed, but there was still one remaining survey. Participants were then asked to complete the modified I-PANAS-SF and the social media craving scale a final time before completing the PAUM, BSMAS, and SNS use question. Following completion of these surveys, participants were given a debrief.

Fig. 2
figure 2

Experiment Procedure

Full size image

Results

Data Analysis

Statistical analyses were performed using IBM SPSS Version 29 (IMB Corp, 2022). After removing incomplete responses and responses from those who had failed attention check questions, the final sample consisted of 114 participants, 85.09% female, mean age = 18.61 (SD = 0.71). 47.37% of this sample was considered at risk of SNS addiction. Results from Chi-square tests for independence demonstrated no significant associations between assigned boredom group and risk of addiction, X2(1, n = 114) = 0.63, p = 0.43, phi = -0.07, no significant association between assigned boredom group and style of SNS use, X2(1, n = 114) = 1.30, p = 0.25, phi = -0.11, and no significant association between risk of SNS addiction and style of SNS use, X2(1, n = 114) = 1.75, p = 0.19, phi = -0.12.

Boredom Ratings

Boredom ratings were averaged across participants at each measuring point separately for each mood induction group (see Fig. 3A). A mixed design ANOVA with the within-subject factor Measuring Point (T1, T2, T3, T4) and the between-subject factor Mood Induction (boredom, control) revealed a significant main effect of Measuring Point, F(3,336) = 35.15, p < 0.001, ηp2 = 0.24: Boredom ratings increased from 1.82 before the induction video (T1) to 2.72 after the video (T2); it dropped to 1.93 after the Go/No-go task (T3) and increased again after the waiting period (T4). All these changes were significant in Bonferroni-corrected posthoc tests (all p < 0.001). The main effect of Mood Video Induction was significant, F(1, 112) = 15.65, p < 0.01, ηp2 = 0.06, indicating that, overall, participants were more bored in the boredom than in the control group (2.39 vs. 2.02). There was also a significant interaction between Measuring Point and Mood Induction, F(3,336) = 17.68, p < 0.001, ηp2 = 0.14, due to a higher increase in boredom after the boredom than after the control video. Posthoc t-tests confirmed that rated boredom only differed at T2, t(112) = 5.93, p < 0.001, Cohen’s d = 1.11, and not at the other measuring points (all p > 0.21).

Fig. 3
figure 3

Average Boredom Ratings (left graph), PA (solid lines) and NA (dashed lines) Ratings (right graph) for the Two Mood Induction Groups Watching the Boredom or the Control Video

Full size image

We also explored whether the risk of SNS addiction (at risk, not at risk), the SNS usage style (active, passive), or the SNS use (yes, no) during the waiting period affected the boredom ratings. Findings from t-tests demonstrated that initial levels of boredom were significantly elevated among those at higher risk of SNS addiction (M = 1.98, SD = 0.92) compared to those participants at lesser risk (M = 1.67, SD = 0.73), t (112) = 2.03, p = 0.04, Cohen’s d = 0.38. By the end of the boredom manipulation, however, these differences had dissipated between participants at higher (M = 2.80, SD = 1.26) and lower (M = 2.57, SD = 1.28) risks of addiction, t (112) = 0.96, p = 0.34, Cohen’s d = 0.18, following the increase in boredom felt by participants who had watched the boredom-inducing video. Otherwise, in the three respective Measuring Point x Group ANOVAs, only SNS use had a marginally significant main effect on boredom, F(1, 112) = 3.49, p = 0.06, ηp2 = 0.03: Participants using SNS during the waiting period were slightly more bored throughout the experiment than participants not using SNS (2.30 vs.2.00, respectively); the other two factors did not matter (risk of SNS addiction, both p > 0.25; SNS usage style, both p > 0.28), and SNS use was found to have no significant impact on changes in self-reported boredom between the start end of the waiting period (p = 0.40).

PANAS

Individual PA and NA scores were calculated and averaged across participants separately for each mood induction group (see Fig. 3B). The ANOVA with the factors Measuring Point and Mood Induction on the PA scores revealed a significant main effect of Measuring Point, F(3,336) = 52.35, p < 0.001, ηp2 = 0.32, due to a decrease in PA after the mood induction and the waiting period. The main effect of Mood Induction, F(1, 112) = 11.08, p < 0.001, ηp2 = 0.09, and the interaction, F(3,336) = 5.26, p < 0.001, ηp2 = 0.05, were significant: As shown in Fig. 3B, the participants in the boredom group showed overall less PA than participants in the control group (10.65 vs. 12.77, respectively), and this difference was most pronounced after the mood induction (T2: 9.21 vs. 12.56, respectively). The corresponding ANOVA on the NA scores revealed only a significant effect for Measuring Point, F(3,336) = 15.25, p < 0.001, ηp2 = 0.12, due to the slight decrease (-0.68) in NA after the mood induction.

We also explored the other three grouping factors, with three separate Measuring Point x Group ANOVAs for PA and the equivalent three ANOVAs for NA. SNS Usage Style had a significant main effect on PA, F(1, 112) = 5.52, p = 0.021, ηp2 = 0.047: Participants using a more active style had overall a higher PA score than participants using more passive style (12.56 vs. 11.03, respectively). The other two factors did not affect PA, neither risk of SNS addiction (both p > 0.89) nor SNS Use (both p > 0.29). Risk of SNS addiction had a marginally significant effect on NA, F(1, 112) = 3.04, p = 0.084, ηp2 = 0.026: Participants at risk of SNS addiction had, overall, a slightly higher NA score than participants not at risk (6.58 vs. 5.93, respectively). The other two factors did not affect NA (all p > 0.24). Additionally, differences in PA and NA between the start and end of the waiting period were not found to differ based upon SNS use while waiting (p > 0.35).

Craving Ratings

Craving ratings were averaged across participants separately for each mood induction group (see Fig. 4A). A mixed design ANOVA with the within-subject factor Measuring Point and the between-subject factor Mood Induction revealed a significant main effect of Measuring Point, F(3,336) = 5.78, p < 0.001, ηp2 = 0.05, due to an increase in cravings after the mood induction (T2) and after waiting period (T4). There was also a significant interaction between Measuring Point and Mood Induction, F(3,336) = 3.90, p < 0.01, ηp2 = 0.03: As shown in Fig. 4A, the craving in the two groups did not differ at the beginning of the experiment (T1); however, after the mood induction (T2-T4), the boredom group reported slightly more cravings than the control group (on average 4.94 vs.4.19, respectively). When we again explored whether the craving ratings were affected by the other three grouping variables, SNS Addiction showed a main effect, F(1,112) = 6.04, p = 0.016, ηp2 = 0.051, indicating that, overall, participants at risk of SNS Addiction reported more craving than participants not at risk (4.99 vs. 3.98, respectively), starting from baseline measurements.

Fig. 4
figure 4

Average Craving Ratings Depending on Participants’ Mood Induction Group (left graph) or Depending on Whether They Used their SNS or Not During the Waiting Period (right graph)

Full size image

Furthermore, as shown in Fig. 4B, SNS Use depended on craving (or vice versa), that is, participants using their SNS during the waiting period had overall more cravings than participants not using their SNS, Group main effect, F(1,112) = 4.04, p = 0.047, ηp2 = 0.035; Note that this craving emerged mainly after the mood induction (T1: 4.28 vs. 4.00; T2-T4 average: 4.89 vs.3.79, respectively), Measuring Point x Group interaction, F(3,336) = 3.18, p = 0.024, ηp2 = 0.028. Finally, SNS Usage and Style did not matter (both p > 0.11) when it comes to reported craving, with no differences between the strength of increase in craving from the start to the end of the waiting period between participants who did and did not use SNS while waiting, p = 0.12.

Grouping Factor Interactions

We explored whether the four grouping factors (Mood Induction, Risk of SNS Addiction, SNS Usage Style, and SNS Use) would interact where they have shown to affect boredom, craving, and PANAS scores. The number of grouping factors (of which 3 were not systematically varied), made it not possible to test the full design involving all four factors, because group sizes were too small and uneven (i.e., 2–12 participants). We therefore decided to test only for 2 × 2 interactions, involving two grouping factors at a time, focusing on those factors that showed significant effects in the result sections above.

We have reported earlier that the boredom ratings depended on mood induction. Three separate 4 × 2 × 2 ANOVAs further explored this effect on boredom, showing that mood induction did not interact with the other three grouping factors, SNS Usage Style, SNS Addiction, or SNS Use (all p > 0.17). We also reported that craving ratings depended on mood induction, SNS addiction, and SNS use. Six 4 × 2 × 2 ANOVAs explored these effects on craving, showing only one significant interaction between Mood induction and SNS Use, F(1, 110) = 4.59, p = 0.030, ηp2 = 0.042: As illustrated in Fig. 5A, overall craving (averaged across T1-T4) was higher in participants using SNS (compared to those not using SNS) – but this effect occurred only in the group that had watched the boredom video. Finally, we reported that the PA scores depended on mood induction and SNS usage style. Five separate 4 × 2 × 2 ANOVAs further explored these effects on PA; however, the two grouping factors did not interact with each other, nor did they interact with the other two grouping factors (all p > 0.16).

Fig. 5
figure 5

The effect of SNS Use on Average Craving Ratings for the Two Mood Induction Groups (left graph) and on Average Reaction Time in Milliseconds with Separate Bars for Block

Full size image

Go/No-go Task

In Go trials, individual mean response times and omission error rates (i.e., misses) were calculated separately for each stimulus type (SNS-related, control). Detection was easier for SNS-related than for control stimuli, as seen in faster response times (489 vs. 517 ms, respectively), t(113) = 8.49, p < 0.001, Cohen’s d = 0.795, and in fewer misses (0.4 vs. 0.8%, respectively), t(113) = 3.97, p < 0.001, Cohen’s d = 0.371.

Four separate 2 × 2 ANOVAs with the within-subject factor Stimulus Type (SNS-related, control) and the between-subject factor Group explored whether any of the four grouping factors (Mood Induction, SNS Addiction, SNS Usage Style, SNS Use) affected the response times in the go-trials. As expected, the within-subject factor Stimulus Type remained significant in all four ANOVAs; however, none of the effects involving Group reached statistical significance (all p > 0.31). The four equivalent ANOVAs for misses showed the same pattern (all p > 0.21).

We split the data into blocks, comparing blocks 1&2 versus blocks 3&4, to test whether the mood manipulation was wearing off over time. Four 2 × 2 × 2 way ANOVA (one for each grouping factor) on the RT revealed a significant interaction between Block (1&2, 3&4) and SNS Use (no, yes), F(1, 112) = 7.95, p = 0.006, ηp2 = 0.066. As can be seen in Fig. 5B, whether participants were going to use SNS or not, affected reaction times in Block 1 and 2 (26 ms), but not in Block 3 and 4.

In no-go trials, participants made overall 16.0% commission errors (i.e., false alarms). Four separate t-tests explored whether any of the four grouping factors (Mood Induction, SNS Addiction, SNS Usage Style, SNS Use) affected the commission errors. Participants in the boredom video group made slightly more errors than participants in the control video group (18.0% vs. 14.3%, respectively), t(112) = 2.13, p = 0.07, Cohen’s d = 0.192. The other three grouping factors did not affect commission errors (all p > 0.38).

When the effect of Stimulus Type as a within-subject factor was examined in 2 × 2 ANOVAs with the between-subject factor Group (Mood Induction, SNS Addiction, SNS Usage Style, SNS Use), findings showed the effect of Stimulus Type as well as the interaction between Stimulus Type and Group were non-significant in all four ANOVAs (p > 0.61). The same was found when responses were broken down by blocks (p > 0.15).

SNS Use

Finally, a Chi-square test for independence was employed to measure whether those at risk of SNS addiction were more likely to engage in SNS usage during the waiting period. Findings showed no significant association between risk of SNS addiction and SNS usage behaviour, X2(1, n = 114) = 0.18, p = 0.67, phi = -0.04.

Discussion

Findings from this study demonstrate the impact of boredom on subjective feelings of craving for SNS, behavioural inhibition and SNS use. Yet, these findings do not necessarily reflect the addictive potential of SNS.

Entering into this experiment, as to be expected, heightened levels of boredom were identified among those at risk of SNS addiction. This heightened boredom aligns with previous research indicating that risk of SNS addiction is associated with boredom proneness (e.g., Wegmann et al., 2018). This finding, as well as the higher level of NA identified among participants at greater risk of addiction at the start of this experiment, may reflect the NA proneness that is typically found in codified addictions (Cheetham et al., 2010). An initial heightened level of craving was also identified among these participants at risk of addiction, which, again, demonstrates how SNS addiction may be similar to codified addictive disorders. Importantly, no significant differences in boredom or craving were identified at baseline between participants randomly assigned to the boredom and control groups. Following a successful boredom induction, however, participants who had seen the boredom-inducing video reported significantly greater levels of both boredom and craving compared to participants who had seen the control video. Contrary to hypothesis 1, this increase in craving occurred regardless of risk of addiction.

When participants went on to complete the Go/No-go task, SNS-related cues were more salient than control cues, that is, they produced fewer omission errors and RTs were significantly quicker. Such enhanced detection of SNS-related Go targets is in line with other studies employing this paradigm (Gao et al., 2019; Turel et al., 2014). These SNS-related boosting effects were, however, contrary to hypothesis 2, not related to risk of addiction or assigned boredom group. However, in no-go trials, it was participants from the boredom group who were found to commit more commission errors, but, again, no differences were found between participants at greater and lesser risk of SNS addiction. Additionally, no differences were found in relation to stimulus type. This suggests that, after an experience of boredom, participants may experience less behavioural inhibition, but this is not cue-related.

As the experiment progressed, participants with greater levels of boredom and craving throughout the experiment were found to be more likely to use SNS during the waiting period, but no differences were identified based upon risk of addiction, leaving hypothesis 3 unsupported. With the waiting period positioned after the Go/No-go task, it is possible that SNS use could have been related to cue-induced craving following exposure to SNS-related imagery. However, craving differences emerged primarily after the mood induction and did not increase following the Go/No-go task.

Finally, it was hypothesised that SNS use would result in reduced feelings of boredom and craving. However, boredom and craving were found to increase between the start and end of the waiting period, with no significant differences identified between the two SNS use groups. Likewise, SNS use was not found to impact self-reports of PA and NA. This finding is contrary to the finding from Drach et al. (2021), who noted an increase in positive affect following SNS use, as well as the findings of Sagioglou and Greitemeyer (2014), who found evidence of mood deterioration following Facebook use.

Throughout this analysis, the effect of style of using SNS was considered. However, the only significant findings were in relation to PA: More active users reported higher levels of PA than more passive users. No association were identified between dominant style of use and risk of addiction. Still, those at higher risk of addiction reported slightly higher NA than those at lower risk.

Conclusion

Overall, findings from this study suggest that SNS usage may follow the typical experiential pattern of dysphoric mood, enhanced craving and usage found with established addictions (Field & Cox, 2008; Field & Powell, 2007), but risk of addiction may not be relevant. Although those at risk of SNS addiction may experience greater boredom proneness and, with that, may crave SNS more, experiences of situational boredom do not appear to have unique effects on these users. Participants who experienced boredom within this experiment by watching a boredom induction video reported increased craving and demonstrated decreased behavioural inhibition. Participants who were more bored were also more likely to use SNS regardless of how they responded to the BSMAS survey. These behavioural results, combined with the differences in RT to SNS-related cues found across participants, indicate that SNS may not be a typical object of addiction. It has been suggested that in internet-use disorders, users have a “first-choice use” not unlike the “first-choice drug” found in substance abuse (Brand et al., 2016). This study has demonstrated that bored participants, rather than those at greater risk of addiction, were more likely to use SNS within these controlled conditions. So, is SNS a “first-choice drug”? It may simply be the “front page of the internet” (Adorjan et al., 2021; Asrese & Muche, 2020; Lagorio-Chafkin, 2018; Organisation for Economic Co-operation & Development, 2019).

Findings from this study fit with past experiments employing the Go/No-go paradigm that have not identified the behavioural differences between participants that would be expected with an addictive disorder (Gao et al., 2017; Moretta & Buodo, 2021a; Turel et al., 2014; Wegmann et al., 2020). These findings also fit with wider literature, showing that, in tasks typically found in addiction research, differences in participant responses to SNS-related cues based upon self-reports related to the risk of SNS addiction have not been consistently found (Du et al., 2020; Moretta & Buodo, 2021b; Thomson et al., 2021). With such research appearing against a backdrop where restrictions of SNS use may be a target in legislation addressing SNS addiction (SMART Act, 2019; Social Media Use for Minors, 2024), there is an urgent need for more experimental research.

It should be noted that the present study is limited to just one type of mood induction. Future studies may consider how other affective states impact craving, behavioural inhibition and SNS use. Of particular interest in the case of SNS addiction is the experience of loneliness as a driver of usage (Boursier et al., 2020; Kolas & von Mühlenen, 2024). However, to induce loneliness could be a challenging ethical choice.

A further limitation of this study is its use of self-report scales. One limitation associated with the use of questionnaires is the risk of social desirability bias (Van de Mortel, 2008). In the case of this study, this was limited by the use of anonymous participant identification numbers. A more pressing limitation associated with the use of self-report measure is found in addiction scale selection. Measures of SNS addiction, as with measures of other novel behavioural addictions, largely follow a transdiagnostic approach in which scales use for novel addictions are based upon extant measures of codified addictions with few meaningful alterations in content beyond the replacement of words related to the target behaviour (Billieux et al., 2015). In the case of SNS addiction, numerous measurements exist, but experimental evidence of the relevance of individual items within these measures is limited. This study employed the BSMAS due to its popularity in SNS addiction research and its items being broadly reflective of the discourse of SNS addiction (Kolas & von Mühlenen, 2024). Although, as an uncodified disorder, it is impossible to formally diagnose SNS addiction, future studies may consider the use of interview with a clinician specialising in codified behavioural addictions to confirm the grouping of participants based upon risk.

Finally, it may be argued that the imagery used in the Go/No-go task could limit the interpretation of findings. In this study images of SNS icons were presented on top of a smartphone image. As SNS is also accessible via computer, this choice in imagery could have presented a confounding variable between participants based upon how they access SNS. However, the online world is primarily via smartphone (China Internet Network Information Center, 2022; OfCom, 2022), and this presentation was designed to create greater ecological validity.