Explaining Social Media Use Reduction As an Adaptive Coping Mechanism: The Roles of Privacy Literacy, Social Media Addiction and Exhaustion

ABSTRACT

This study integrates privacy calculus and stress-coping theories to understand social media (SM) use reduction. Findings using a survey of 255 SM users show that privacy concerns are a stressor that drives adaptive coping mechanisms and later use reduction. Further, it was demonstrated that the translation of privacy calculus considerations into adaptive coping depends on privacy literacy. Moreover, adaptive coping’s impact on use reduction decreased in users addicted to, but not exhausted by SM use.

KEYWORDS:

• Social media use reduction
• social media addiction
• social media exhaustion
• privacy concerns
• privacy literacy
• coping strategies

Introduction

Social media sites, such as Facebook, Instagram, and TikTok, have been widely used by billions of individuals worldwide looking to entertain themselves, connect with the world, and find information. In 2023, there were about 4.95 billion social media users, representing 61.4% of the world’s population (DataReportal, 2023). Despite the benefits, several negative effects of excessive social media use have been observed. These include fatigue, exhaustion, and personal, family, and professional friction, among others (Luqman et al., 2020; Niu et al., 2022; Shokouhyar et al., 2018). Consequently, there has been a growing trend of corrective behaviors toward social media sites, aimed at reducing the negative effects of excessive use (Cao & Sun, 2018; Hsieh et al., 2011; Lin et al., 2021; Maier et al., 2015; Rahrovani & Addas, 2019). Social media corrective behaviors represent self-imposed actions that users of social media platforms may take to lessen the unpleasant outcomes of their use patterns (Islam et al., 2022; Osatuyi & Turel, 2020). Some examples of corrective actions include quitting social media use, reducing it, taking a “vacation” from the use, or simply switching to another social media platform (Maier et al., 2015; Neves et al., 2023; York & Turcotte, 2015).

While quitting, vacationing, and switching have already been covered in numerous studies, the study of SM use reduction is still rare (Neves et al., 2023). Factors such as realization (when the user realizes the severity and problematic nature of the current levels of social media usage), peer reduction of use (when the users’ peers decide to reduce the social media levels of usage), social media addiction (Osatuyi & Turel, 2020), information and social overload, fatigue (Niu et al., 2022; Shokouhyar et al., 2018), and dissatisfaction (Fu & Li, 2020) have been identified as motivators of social media use reduction. However, there can be additional perspectives that can explain why people decide to self-regulate and reduce their social media use. It is important to study such perspectives because social media use reduction impact users, as well as social media service providers.

Here, we follow Neves et al. (2023) to propose that the privacy calculus theory can serve as a novel lens for understanding social media use reduction. Neves et al. (2023) demonstrated that privacy concerns (institutional or peer) drive social media use reduction. Extending this view, we seek to (1) examine how users weigh privacy concerns against the benefits of use to develop corrective behaviors, and to (2) unpack the mechanisms through which such privacy calculus considerations translate into use reduction intentions. We propose that such mechanism has to do with privacy literacy, which captures one’s understanding of data protection and the capacity to use techniques that provide data protection and privacy safety (Bartsch & Dienlin, 2016). Privacy literacy may influence the decision process of users when weighing concerns against benefits, since knowledge about risks generates more awareness of problems and possible violations that may happen. As such, the first research gap that this study aims to address relates to how different knowledge levels of privacy protection techniques may impact privacy calculus effects on corrective behaviors.

A second research gap involves the limited understanding of how coping strategy choice can motivate use reduction (Lin et al., 2021). Coping strategies (behavioral and cognitive methods used to control crises and situations that are evaluated as stressing) are important determinants of changing one’s technology use pattern (Chen et al., 2019; Jiang et al., 2013) and corrective behaviors (Bae, 2023). Here, we intend to explore the motivating role of adaptive coping strategies and boundary conditions that govern their effects. Adaptive coping strategies represent positive techniques that users may adopt to deal with stressful situations. In the context of SM use, adaptive strategies can be characterized by actively taking steps to avoid the stress or directly decrease its impact (Gaudioso et al., 2017). An opposite approach would be neglecting the stressor and ignoring it (Lin et al., 2021). In this study, we will bring a new stressor to the literature of SM use reduction, examining the privacy perspective as a stressor and assessing its influence on adaptive coping technique adoption and, in turn, the decrease in social media use. While the majority of the studies use this theory to investigate the use of social media (Bae, 2023; Jozaghi et al., 2016), we will use it to understand the use reduction.

The third gap we aim to address relates to the limited understanding of the factors that can influence the process through which adaptive coping can (or fail to) motivate SM use reduction. Such contingency factors can highlight when and how the use reduction process will be more or less potent. While most studies use mediators such as fatigue, regret, and guilt to explain the adoption of social media corrective behaviors (O. Turel, 2014, 2016), the moderating power of users’ psychological states has been overlooked. Here we suggest that SM addiction and exhaustion are important factors that can affect the success of coping (Lin et al., 2021; Maier et al., 2015). SM addiction is a maladaptive dependency of SM “to an extent that various behavioral addiction symptoms arise, including salience, withdrawal, conflict, relapse and reinstatement, tolerance, and mood modification” (Venkatesh et al., 2019, p. 905). Since addiction decreases the ability of users to control and rationally manage social media usage levels, it should affect the motivation potential of adaptive coping. Moreover, we note that SM exhaustion is one of the most common consequences of excessive social media use, representing the situation when the users develop a sensation of being overloaded and “drained” (Sheng et al., 2023). Here, we posit that whether one’s addiction is accompanied (vs not) by exhaustion can affect the translation of coping strategies into SM use reduction. By filling this last gap, our study can lead to not only a better understanding of the process through which privacy reflections motivate SM use reduction, but also provide initial insights into some of the boundary conditions of this process.

To achieve the proposed goals, we hypothesized a model based on the coping theory (Folkman & Lazarus, 1988) where first we posit that privacy concerns, as weighed against perceived benefits, can serve as a stressor that motivates adaptive coping behavior, which in turn, drive intended use reduction. We then theorize how privacy literacy, SM addiction and SM exhaustion can affect this process. We tested the model with a sample of 275 Instagram users.

Ultimately, this paper provides four main contributions. First, it extends the current body of knowledge on corrective behaviors on SM, specifically SM use reduction. We show that privacy calculus informs the adoption of adaptive coping strategies, which in turn, motivates SM use reduction. Second, we show that privacy literacy makes the privacy calculus more potent, and consequently the use of adaptive coping strategies is more likely. In essence, knowledge about privacy makes users more aware of the risks and consequently more predisposed to adopt safe behaviors. Third, we observed a three-way interaction of SM addiction and SM exhaustion on adaptive coping effects, demonstrating that adaptive coping strategies are not equally efficacious in all users. The findings specifically suggest that highly addicted users who are not exhausted by SM use are least motivated, compared to others, to reduce their SM use. Last, the results of this study can provide a basis for practical interventions on SM platforms and SM communities by understanding the barriers and drivers for SM use reduction.

Theoretical background and hypotheses

Social media use reduction drivers

Social media use reduction is a type of corrective behavior on social media. It reflects a motivated reduction of SM usage levels, often aimed at improving users’ wellbeing and allowing them to focus on other tasks and goals. It allows users to curb the problems generated by excessive SM use, such as guilt, anxiety, and exhaustion (Fu et al., 2020; Nawaz et al., 2018). Other known corrective behaviors are quitting (stopping SM use), vacationing (taking a temporary break from SM use) and switching (changing to another SM). These behaviors have the same objective of curbing the negative consequences of SM use (Luqman et al., 2020; Maier et al., 2015; Turel, 2016). Since quitting removes the problem’s root cause, it can be seen as an active endeavor to eliminate all problems generated by social media use (Luqman et al., 2020; Osatuyi & Turel, 2020). However, it also ends all the advantages that social media use may offer. A comparable approach is to temporarily stop utilizing social media. While this allows for a short solution to the problems, these are likely to continue once use is resumed (Ravindran et al., 2014; York & Turcotte, 2015). Similarly, switching to another social media might solve the negative consequences generated by its use, however, it also eliminates any benefits that using that specific social media platform could have and the problems may surface again, if the level of use is not contained (Maier et al., 2015). As such, use reduction appears to be the only corrective behavior that can mitigate the social media problems generated from excessive use, while still providing the benefits of using this technology. By reducing social media use levels, users can decrease negative feelings (e.g., guilt, anxiety, family friction, work friction) and still benefit from all the advantages of these platforms. It is therefore considered in the literature as more realistic and practical compared to other corrective behaviors (Osatuyi & Turel, 2020).

Despite the practicality of this approach, the literature on SM use reduction is still in the embryonic stages. It suggests that drivers of this behavior include SM fatigue and exhaustion (Fu & Li, 2020; Islam et al., 2022; Niu et al., 2022; Shokouhyar et al., 2018), information, technology and social overload (Fu & Li, 2020; Niu et al., 2022; Shokouhyar et al., 2018), peer reduction and realization (Osatuyi & Turel, 2020), and privacy concerns (Neves et al., 2023). Theories that have been used to explain SM use reduction include Lazarus and Folkman’s theory of stress and coping (Islam et al., 2022), Social identity theory (Osatuyi & Turel, 2020), Social cognitive theory (Fu & Li, 2020; Osatuyi & Turel, 2020), Stimulus-Organism-Response (Niu et al., 2022) and the Privacy-calculus theory (Neves et al., 2023).

While such studies make the first strides toward understanding SM use reduction, they have not fully unpacked the privacy angle. Only one study demonstrated the role of privacy in the adoption of SM use reduction (Neves et al., 2023). That study focused on understanding the role of two facets of privacy in the adoption of SM use reduction. Our study intends to extend the privacy calculus perspective by integrating it with the stress coping theory and considering boundary conditions for privacy consideration effects. Notably, the theory of stress and coping has been used solely to study this behavior by Islam et al. (2022). However, the focus of this research was on evaluating the stressors induced by a global pandemic in the adoption of SM use reduction. Our study will test a new perspective on the role of privacy reflection in use reduction decisions, and account for factors that can affect this process. These include privacy literacy, SM addiction and SM exhaustion. Notably, SM addiction and SM exhaustion have been used as direct or indirect factors in models of SM use reduction (Fu & Li, 2020; Islam et al., 2022; Niu et al., 2022; Osatuyi & Turel, 2020). However, here, we posit that they serve as boundary conditions (moderators) for the processes the underlie the translation of privacy reflection into use reduction decisions.

Overall, we find that SM use reduction did not receive sufficient attention despite its potential and importance. By considering the privacy perspective, adding a new lens through the stress coping theory, and evaluating different moderators, we hope to better understand this behavior.

Stress coping theory

Stress coping theory explains how individuals deal with stress through the adoption of coping strategies (Folkman & Lazarus, 1988). This theory has been used in different contexts such as psychiatry, medicine, psychology, and information systems (IS). In the IS context, the coping theory has been used for explaining technostress (Ayyagari et al., 2011; Islam et al., 2022; Maier et al., 2015). Technostress is associated with the stress generated by the use of technology, such as social media, producing a set of negative effects for users, including a sense of overload, invasion, and insecurity. Stress captures a reaction caused by a pressure that is higher than what individuals can support (Islam et al., 2022). To deal with that situation, individuals can engage in coping strategies that try to mitigate the stress or its sources. A coping strategy, therefore, aims to reduce or eliminate stressors or stress as a means to improve the individual’s state. Typical coping strategies are broadly aggregated into two types: emotional-focused, and problem-focused (Bae, 2023). Emotional-focused coping happens when an individual tries to deal with emotions generated from a stressful situation (i.e., the symptoms). For example, a common emotional-focused strategy is expressive-support seeking (search for help or social approval). Problem-focused coping, in contrast, is characterized by the creation of a plan to solve the stress source. For example, avoidance (distance from the source of stress) is a common problem-focused strategy. Since problem-focused coping methods (adaptive coping) are the most effective type of coping, we concentrate on them in our study (Gaudioso et al., 2017). These strategies are the antithesis of quitting or dismissing the issues; they involve handling and confronting the problem. Learning to manage SM use is an example adaptive coping strategy.

The stress coping theory has been used as a foundation of many studies about technostress (Ayyagari et al., 2011; Chen et al., 2019; Gaudioso et al., 2017; Hsiao et al., 2017; Ma & Turel, 2019; Tarafdar et al., 2020). Specifically, in the social media corrective behaviors context, the literature already used the stress coping theory to mainly explain the adoption of discontinuance decisions (Lin et al., 2021; Maier et al., 2015). More recently, Islam et al. (2022) brought this theory to explain the effect of COVID-19 on the adoption of social media use reduction behavior. Given that social media use is a source of stress for users (Lin et al., 2021; Turel, 2016), the stress coping theory plays an important role when trying to explain adaptive coping strategies. However, there is still limited knowledge about the role that coping plays in motivating use reduction, and about factors that govern the translation of privacy considerations into coping strategy choices, and of these choices into use reduction decisions.

Here, we seek to bridge the abovementioned research gaps. We posit that users who perceive stressors related to privacy will develop an adaptive coping strategy, which will affect their SM use reduction intention. The applicability of this theory in our study is consistent with previous research that has used stress coping theory to examine how technostress translates through coping strategies into corrective behaviors (Islam et al., 2022; Lin et al., 2021; Maier et al., 2015). Our study extends the use of this theory to the SM use reduction phenomenon, by using the privacy lens to explain the stress coping process.

Hypotheses

Privacy calculus theory suggests that SM users weigh the privacy concerns and benefits when deciding to use the technology. As such, there is an important balance between risks and rewards when using SM, with a specific emphasis on privacy risks, which are prevalent in SM (Neves et al., 2023). Privacy concerns are defined as worries and risks about the possible misuse of personal information that is shared through SM (Choi et al., 2018; Malhotra et al., 2004). The substantial volume of data that can be used by others raises privacy challenges for SM users, who can often envision their information exposed without their permission (Bright et al., 2015). While some research uses and defines two facets of privacy (S. Choi, 2023; Epstein & Quinn, 2021; Neves et al., 2023): institutional privacy concerns (worries about the misuse of private information by the social media provider) and social (worries about the misuse of private information by the user´s peers), here, we focus on the institutional perspective as a first step. We made this choice given that institutional privacy concerns have been commonly studied and acknowledging the need to also consider social privacy concerns in the future.

Due to the social media exposure, users might feel stressed and fear to use SM. As such, it is reasonable to view privacy concerns as a stressor while using SM. As a stressor, it motivates the employment of coping strategies. In contrast, perceived benefits are the positive aspects that the SM user feels while operating with the technology. It is therefore a destressing factor that demotivates the employment of coping strategies.

Here, we focus on adaptive (problem-focused) coping strategies, because they are the most efficacious form of coping (Gaudioso et al., 2017). These manifest in users deciding to apply techniques to mitigate stressful situations. Adaptive coping strategies are the opposite of denying the problems or giving up. These strategies include active coping, by dealing with and facing the problem. Some examples are learning how to control the use and planning the use of features to limit technology use (Gaudioso et al., 2017). Ultimately, we expect privacy concerns to motivate engaging in adaptive coping strategies and perceived benefits to demotivate engaging in adaptive coping strategies. As such, we hypothesize that:

H1:

Privacy concerns increase engagement in adaptive coping strategies.

H2:

Perceived benefits reduce engagement in adaptive coping strategies.

Extending this view, we posit that the abovementioned associations can be influenced by one’s levels of privacy literacy. Privacy literacy is defined as the knowledge about data protection and the ability to apply strategies that guarantee privacy safety and data protection (Bartsch & Dienlin, 2016). An individual with higher levels of privacy literacy will likely have a more informed and risk-averse privacy calculus, reflected in putting higher weights on privacy concerns and lower weights on privacy benefits, compared to someone low in privacy literacy. As such, we hypothesize that:

H3:

Privacy literacy moderates (strengthens) the positive relationship between privacy concerns and adaptive coping strategies.

H4:

Privacy literacy moderates (weakness) the negative relationship between perceived benefits and adaptive coping strategies.

Adaptive coping strategies, as defined before, are a technique that SM users adopt to deal with stressful situations originated by the SM use. These strategies are more likely than others to point to the need to control one’s level of SM use in order to mitigate negative feelings and stress. As such, adaptive coping strategies can motivate SM use reduction. SM use reduction is a natural choice here as a coping mechanism because it allows stress reduction while allowing users to still enjoy the SM benefits. As such, we hypothesize that:

H5:

Adaptive coping strategies increase SM use reduction intentions.

Next, we argue that this relationship might be affected by factors such as SM addiction or exhaustion. This is because such factors can affect the ability and motivation of users to reflect on their states and choices (Naranjo-Zolotov et al., 2021; Qahri-Saremi et al., 2021; Serenko & Turel, 2022). Addiction to technology use manifests in less rational reflection on the data available to the individual; people who are addicted put a stronger emphasis on immediate rewards and ignore long term consequences (Turel et al., 2011; Vaghefi et al., 2020a). Consequently, users addicted to SM use often find it difficult to limit their usage levels. As such, they are less motivated to follow through with adaptive coping strategies they might have considered (Sriwilai & Charoensukmongkol, 2016). This type of user typically adopts maladaptive coping strategies such as denial and ignorance of a problem (Qahri-Saremi & Turel, 2020). As such, addicted individuals’ motivation to reduce SM use will be less sensitive to adaptive coping strategies, compared to others. In essence, it is more difficult to sway them from continuing to use SM excessively. As such, we hypothesize that:

H6:

SM addiction moderates (reduces) the relationship between adaptive coping strategies and SM use reduction intentions.

Lastly, it is reasonable to assume that the abovementioned moderated association is also dependent on the SM exhaustion levels. This idea is rooted in the notion that addicted SM users with high exhaustion levels are more motivated than others to follow through with their adaptive coping strategies. SM exhaustion is an important driver of SM use reduction behavior (Fu & Li, 2020; Fu et al., 2020). Here, we suggest a novel mechanism through which it operates. If addicted individuals are not exhausted by SM use, they are not motivated to act in the form of reduced use. This manifests in addiction reducing the translation of adaptive coping strategies on use reduction. In contrast, when addicted individuals are exhausted, they are more motivated to act, despite being addicted. This manifests in addiction having low or no effect on the translation of adaptive coping into use reduction. As such we hypothesize that:

H7:

There is a three-way interaction among (a) adaptive coping strategies, (b) SM addiction, and (c) SM exhaustion. The negative moderation effect of addiction on the translation of adaptive coping strategies into SM use reduction intentions will diminish as users become more exhausted by SM use.

The proposed hypotheses are summarized in the following research model (Figure 1).

Figure 1. Research model.

Methods

We created an online survey with items adapted from validated scales (see Table 1). It is important to note that the way privacy concerns has been conceptualized and measured varies greatly among studies (Bartol et al., 2021; Li, 2011). In our study, we based the privacy concerns items on previous validated research (Smith et al., 1996; Zhu et al., 2021) used many times and proved to be reliable.

Table 1. Measurement instrument.

Construct	Item	Source
Privacy concerns (PC)	Pc1. Using Instagram, I’m afraid my privacy will be violated. Pc2. Using Instagram, I’m worried that my personal information will be over collected. Pc3. Using Instagram, I’m worried that my personal information will be accessed without authorization. Pc4. Using Instagram, I’m worried that my daily life will be tracked and monitored.	Smith et al., (1996); Zhu et al., (2021)
Perceived benefits (PB)	Pb1. The use of Instagram helps me communicate with others. *(deleted) Pb2. The use of Instagram helps me to feel good. Pb3. The use of Instagram can effectively help me achieve my personal goals. Pb4. The use of Instagram entertains me. Pb5. In general, the use of Instagram is beneficial to me.	Zhu et al., (2021)
Privacy literacy (PL)	Pl1. I know how to delete or deactivate my Instagram account. *(deleted) Pl2. I know how to restrict access to Instagram profile information such as hobbies, interests. Pl3. I know how to make my Instagram profile not accessible via Google. Pl4. I know how to control if others tag my name on Instagram pictures. *(deleted) Pl5. I know how to restrict access to my Instagram postings. Pl6. I know how to restrict access to my contact information (e.g. name, address).	Bartsch & Dienlin, (2016)
Adaptive coping strategies (ACS)	Acs1: I have been concentrating my efforts on doing something about the privacy stressful situations on Instagram. Acs2: I have been taking action to try to make the privacy stressful situation better. Acs3: I have been getting help and instrumental support from other people regarding dealing with privacy stressful situations on Instagram. Acs4: I have been trying to get instrumental support or help from other people about what to do regarding privacy stressful situations on Instagram. Acs5: I have been trying to come up with a strategy about what to do regarding privacy stressful situations on Instagram. Acs6: I have been thinking hard about what steps to take regarding privacy stressful situations on Instagram.	Gaudioso et al., (2017)
SNS addiction (ADD)	Add1. Using Instagram sometimes interfered with other activities. Add2. I have made unsuccessful attempts to reduce the time I interact with Instagram. Add3. Arguments have sometimes arisen at home because of the time I spend on Instagram. Add4. I think that I am addicted Instagram.	Moqbel & Kock, (2018)
SNS Exhaustion (E)	E1: I feel drained from activities that require me to use Instagram. E2: I feel tired from my Instagram activities. E3: Using Instagram is a strain for me.	Nawaz et al., (2018)
Intention to reduce SNS use (RED)	Red1: I intend to reduce using Instagram in the next 3 months. Red2: I predict I would reduce using Instagram in the next 3 months. Red3: I plan to reduce using Instagram in the next 3 months.	Osatuyi & Turel, (2020)

*Pb1, Pl1, and Pl4 were deleted due lack of indicator reliability.

The survey was distributed to university students who use Instagram, because Instagram is the most popular social networking site among students (O. Turel, 2016). We tested our survey with a pilot test using 50 Instagram users. No major changes were made after the pilot.

For data collection, we used a random sample of university students in Portugal, registered on a research database. The students were enrolled in a program taught in English. Thus, all were fluent in English. Of the 390 students we approached, 255 submitted valid answers (65.4% response rate). The calculation for the minimum sample size was performed using the sample size formula for an infinite population, specifically (n = Z^2p * q/d^2), where Z is the standard normal distribution for the (1 − α/2) level, d is the precision, p is the prevalence, and q = (1 − p). Regarding the prevalence (p), we have resorted to statistical data on the adoption of social media corrective behaviors in Portugal, showing a 20% penetration rate (Marktest, 2022). A level of precision (d) of 5% was also assumed. Given this, the minimum sample size needed was 246 respondents, which was achieved.

The questionnaire preserved the subject’s anonymity and included a statement of confidentiality. Moreover, all questions were indirect through measurement items tap manifest from latent variables. Additionally, we reassured respondents that the survey was voluntary, and that withdrawal was available at any time. These measures reduced the risk of dishonest and social-desirable responses (Kwak et al., 2019).

In the survey design it is also important to consider potential cofounding variables. As such, in line with prior studies, we controlled for common confounders of user behavior on social media: age, gender (Osatuyi & Turel, 2020), average SM use time per day (minutes), and SM experience (years using SM) (Vaghefi et al., 2020b).

The resultant sample included 51.3% males, the average age was 24 years old, the average time (in minutes) spent per day on Instagram was 45.6 minutes, and the average number of years using Instagram was 7.2. We examined common method bias risk using a marker variable (level of familiarity/knowledge about the university where this study was conducted) (Lindell & Whitney, 2001). Results showed 1.1% maximal shared variance with other variables, which is indicative of low risk for common method variance (Johnson et al., 2011).

Preliminary analysis

The measurement model was tested using SPSS 29.0 and PLS 3.0. Internal consistency and convergent reliability were validated with composite reliability (CR) scores higher than 0.7, average variance extracted (AVE) scores higher than 0.5 (see Table 2), and loadings values higher than 0.5 (see Table 3). Sufficient discriminatory validity was demonstrated by having the square root of AVE of each construct is higher than the corresponding correlations (see Table 2), loadings are higher than the cross-loadings (see Table 3), and heterotrait-monotrait ratio (HTMT) lower than 0.9 (see Table 4). Also, a confirmatory factor analysis (CFA) revealed appropriate fit indices (χ2/df = 2.47, RMSEA = 0.07 (90% CI [0.06, 0.08]), SRMR = 0.07, CFI = 0.90, IFI = 0.90).

Table 2. Descriptive statistics, correlations, composite reliability (CR), and average variance extracted (AVE).

Constructs	Mean	SD	CR	PPR	PB	PL	ACS	ADD	E	RED
Perceived Privacy Risk (PPR)	4.29	1.60	0.950	0.909
Perceived Benefits (PB)	4.28	1.08	0.830	−0.109	0.743
Privacy Literacy (PL)	3.63	1.53	0.765	−0.107	0.106	0.730
Adaptive Coping Strategies (ACS)	2.26	1.36	0.946	0.276	0.048	0.202	0.862
SM Addiction (ADD)	3.09	1.46	0.858	0.283	0.226	−0.103	0.398	0.776
SM Exhaustion (E)	2.93	1.66	0.938	0.246	−0.115	−0.044	0.431	0.488	0.914
Intention to reduce SM use (RED)	3.61	1.94	0.966	0.215	−0.041	−0.079	0.257	0.522	0.513	0.952

Values in diagonal (bold) are the AVE square root; SD = standard deviation.

Table 3. Loadings and cross-loadings.

Constructs	Items	PPR	PB	PL	ACS	ADD	E	RED
Perceived Privacy Risk (PPR)	PPR1	0.906	−0.062	−0.119	0.256	0.279	0.217	0.237
	PPR2	0.947	−0.112	−0.083	0.276	0.289	0.265	0.200
	PPR3	0.919	−0.136	−0.148	0.228	0.265	0.171	0.170
	PPR4	0.864	−0.090	−0.039	0.239	0.187	0.235	0.170
Perceived Benefits (PB)	PB2	0.011	0.745	0.054	0.051	0.233	−0.102	0.025
	PB3	−0.057	0.805	0.125	0.142	0.213	0.036	0.027
	PB4	−0.158	0.669	−0.020	−0.157	0.160	−0.147	−0.013
	PB5	−0.152	0.745	0.157	0.061	0.008	−0.225	−0.242
Privacy Literacy (PL)	PL2	−0.096	0.146	0.751	0.122	−0.029	−0.070	−0.008
	PL3	−0.083	0.066	0.884	0.207	−0.124	−0.032	−0.090
	PL5	−0.059	0.016	0.501	0.059	−0.035	0.029	−0.075
Adaptive Coping Strategies (ACS)	ACS1	0.295	−0.010	0.186	0.857	0.368	0.429	0.272
	ACS2	0.288	−0.011	0.195	0.857	0.289	0.321	0.199
	ACS3	0.178	0.142	0.157	0.852	0.392	0.354	0.168
	ACS4	0.132	0.134	0.184	0.843	0.378	0.361	0.180
	ACS5	0.219	0.056	0.166	0.897	0.342	0.396	0.264
	ACS6	0.285	−0.032	0.156	0.867	0.303	0.358	0.227
SM Addiction (ADD)	ADD1	0.229	0.099	−0.138	0.231	0.756	0.383	0.457
	ADD2	0.251	0.132	−0.135	0.268	0.845	0.399	0.464
	ADD3	0.213	0.225	0.058	0.416	0.708	0.348	0.317
	ADD4	0.174	0.292	−0.057	0.376	0.788	0.382	0.346
SM Exhaustion (E)	E1	0.236	−0.022	0.042	0.448	0.449	0.911	0.452
	E2	0.213	−0.120	−0.060	0.373	0.478	0.938	0.498
	E3	0.226	−0.173	−0.100	0.362	0.409	0.892	0.453
Intention to reduce SM use (RED)	RED1	0.218	−0.050	−0.110	0.246	0.524	0.538	0.958
	RED2	0.178	−0.043	−0.003	0.252	0.443	0.433	0.928
	RED3	0.216	−0.024	−0.101	0.237	0.516	0.484	0.969

Table 4. Heterotrait-monotrait ratio (HTMT).

Constructs	PPR	PB	PL	ACS	ADD	E	RED
Perceived Privacy Risk (PPR)
Perceived Benefits (PB)	0.160
Privacy Literacy (PL)	0.142	0.207
Adaptive Coping Strategies (ACS)	0.289	0.171	0.231
SM Addiction (ADD)	0.325	0.322	0.185	0.490
SM Exhaustion (E)	0.268	0.220	0.105	0.469	0.579
Intention to reduce SM use (RED)	0.226	0.126	0.114	0.270	0.590	0.551

Structural model

After establishing that the measurement model is satisfactory, we test our model in blocks, using PLS 3.0. The variable entry order was: (block 1) control variables, privacy concerns, perceived benefits, privacy literacy, adaptive coping strategies; (block 2) SM addiction moderation; (block 3) SM exhaustion three-way interaction with SM addiction and adaptive coping strategies. Figure 2 contains standardized path coefficients, significance levels, and explained variance.

Figure 2. Structural model.

Our model explains 21.2% of the variation in adaptive coping strategies and 43.5% of the variation in SM use reduction intention. Privacy concerns have a positive and statistically significant effect on adaptive coping strategies ($\hat{\beta }$ = 0.278, p < .01). Thus, H1 is supported. Perceived benefits are not a statistically significant predictor of adaptive coping strategies ($\hat{\beta }$ = -0.006, p > .10). Thus, H2 is not supported, at least at average levels of privacy literacy. Privacy literacy moderates the relationship between privacy concerns and adaptive coping strategies ($\hat{\beta }$ = 0.147, p < .01) and the relationship between perceived benefits and adaptive coping strategies ($\hat{\beta }$ = 0.159, p < .01). Thus, H3 and H4 are supported. Adaptive coping strategies have a positive and statistically significant effect on SM use reduction intention ($\hat{\beta }$ = 0.236, p < .01). Thus, H5 is supported. SM addiction moderated the relationship between adaptive coping strategies and SM use reduction intention ($\hat{\beta }$ = -0.110, p < .01). As such, H6 is also supported. Last, there is a statistically significant three-way interaction among adaptive coping strategies, SM addiction, and SM exhaustion in the expected direction ($\hat{\beta }$ = 0.090, p < .01). Thus, H7 is supported.

Discussion

This work demonstrated that privacy concerns can be viewed as a stressor that increases engagement in adaptive coping strategies (H1). This result supports Lazarus’ overarching theory of stress dynamics and coping and extends it to privacy in the SM context. Contrarily, benefits did not influence engagement in adaptive coping strategies (H2). The benefit of using SM is a hygiene factor that is needed for continued use, but it is not potent in reducing safety-focused behaviors. This aligns with previous studies that prove that benefits are a driver for the continuance usage of a technology (Cheng et al., 2021; Turel & Serenko, 2012). The work suggests that behaviors that promote user safety are influenced primarily by risk factors such as privacy concerns (worries of privacy violation during social media use), and less so by benefit considerations (such as entertainment, enjoyment, and satisfaction with social media use).

Regarding the moderating effects, we concluded that privacy literacy increased the positive effect of privacy concerns (H3) on the adoption of adaptive coping strategies (see Figure 3). This means that for users with higher levels of privacy literacy, more aware of risks, and knowledgeable of data protection practices, the positive effect of privacy concerns is higher for the adoption of adaptive coping strategies than for users with low privacy literacy. We also concluded that privacy literacy moderates the relationship between perceived benefits and adaptive coping strategies (H4) (see Figure 4). We demonstrated that for low levels of privacy literacy, perceived benefits have a negative impact on the adoption of coping strategies, while for high levels of privacy literacy, the impact of perceived benefits becomes positive. This means that knowledgeable users, even perceiving the SM benefits as relevant and important, recognize the need to have a healthy use and the risks associated with the SM use. Overall, these situations demonstrate the need to have knowledgeable users with awareness of SM risks in order to users engage in less risky behaviors and mitigate negative stress and consequences of SM use.

Figure 3. Moderation effect 1.

Figure 4. Moderation effect 2.

This work also showed that adaptive coping strategies lead to the adoption of SM use reduction (H5). This means that one possible coping strategy to mitigate the negative balance between risk and benefits that emerges from one’s privacy calculus is SM use reduction. Overall, it is possible to conclude that people that are aware and informed about the privacy risks of using SM more easily engage in adaptive coping strategies and then, adopt SM use reduction behavior, achieving a healthier and less risky SM use.

Additionally, our study concluded a three-way interaction (H6 and H7) between adaptive coping strategies, SM addiction, and SM exhaustion (see Figure 5). We concluded that for highly addictive users, the positive impact of adaptive coping strategies on the adoption of SM use reduction behavior is lower than for low addictive users. We also noticed that in a situation of “fatigue syndrome,” where users experience high levels of addiction and exhaustion at the same time, the power of adaptive coping strategies on SM use reduction is higher than for users that are addicted and less exhausted. This means that exhaustion makes the effect of adaptive coping strategies decrease less than for users more exhausted yet still addicted. Overall, the conclusions shed light on the privacy-coping processes that drive people to reduce SM use, and factors that make it harder for people to reduce use. The findings ultimately point to possible ways to control the negative outcomes of SM use.

Figure 5. Moderation effect 3.

Theoretical implications

Our work explored the adoption of SM use reduction through the application of stress coping theory and privacy calculus theory, making several contributions theoretical contributions.

First, the findings allow us to expand the current body of knowledge about the adoption of SM corrective behaviors, and specifically SM use reduction. We showed that privacy concerns is an important stressor that leads to the adoption of SM use reduction through the implementation of adaptive coping strategies. This work is in consonance with other studies that proved the effect of privacy related stressors as drivers of coping mechanisms (Chen et al., 2019; McHugh et al., 2018). However, our work is different from the existing studies by applying similar logic to the study of social media corrective behaviors.

Our work also confirms the conclusions of Neves et al. (2023) that demonstrated the positive impact of privacy concerns on SM use reduction decisions. In our study, we do not demonstrate separately the effects of the two facets of privacy concerns as Neves et al. (2023) did. However, we extended the ideas by conceptualizing privacy concerns as a stressor that leads to adaptive coping strategies, and later SM use reduction, and by unpacking the boundary conditions (moderators) that govern this process. As such, the current work is a starting point for the exploration of the role of privacy reflections in driving corrective behaviors in the technology context.

Second, our work confirms the power of adaptive coping strategies in motivating SM use reduction. This conclusion is an extended version of prior works that assumed that adaptive coping strategies have a negative effect on quitting SM use (Lin et al., 2021). While adaptive coping strategies negatively impact the discontinuance use of SM, in the case of use reduction, the conclusions are contrary. This shows that users who have in mind a problem-focused strategy to deal with the consequences and stressors of social media use, prefer the adoption of use reduction, thinking about the benefits that they can still enjoy, instead of taking a drastic approach and stopping using social media. This demonstrates the need to study social media use reduction as a standalone phenomenon; models of discontinuance may not always apply to social media use reduction decisions.

Third, we theorized and tested the moderator role of privacy literacy demonstrating that users more knowledgeable about data protection are more aware of the SM risks and more predisposed to adopt safety behaviors in the SM context. This conclusion confirms that literacy is an excellent approach to providing safe social media usage, decreasing the intention to disclose private information, and excessively using social media. As previous works confirmed (Bartsch & Dienlin, 2016; Desimpelaere et al., 2020), knowledge is a driver of privacy safety behavior; and as such, serves as a boundary condition for people’s coping strategy choices. By using privacy literacy as a moderator in our study, we also highlight the need for other possible interactions in these types of studies. Variables that describe different physical and emotional states of SM users can be used as moderators (e.g. negative affect). The current work paves the way for future studies to consider additional moderation effects on the translation of privacy reflections into coping choices.

Last, we discovered a three-way interaction that explains why not all adaptive coping choices motivate use reduction. Specifically, in SM users with simultaneous high levels of addiction and exhaustion there is a better translation than in others, of adaptive coping strategies into SM use reduction intention. This conclusion is in consonance with the work of Kim et al. (2023), that studied the “fatigue syndrome.” This syndrome characterizes users who suffer from SNS exhaustion and present typical addictive symptoms at the same time. We discovered that the efficacy of adaptive coping techniques in reducing SM use is higher in users who are both very addicted and extremely exhausted at the same time than for addicted users but little exhausted. This informs future research as it alludes to the need to combine multiple moderation effects. The three-way interaction discovered here gives an initial nuanced view of why people vary in their motivation to reduce SM use. We call for more research to examine this important question.

Practical implications

The results of the study allow us to provide some practical recommendations, namely from the user’s perspective but also for social media providers. Firstly, our findings suggest that social media providers that may want to prevent use reduction, should reduce users’ privacy concerns. While we did not study ways to do so, the literature suggest many ways to reduce privacy concerns (Lin et al., 2021), such as providing assurances, and implementing technical measures. The efficacy of such approaches should be considered in future research.

Secondly, given that privacy literacy is one of the factors that potentiates a healthy use of social media, we recommend increasing SM users’ privacy literacy and awareness. With more knowledgeable users, the risks of using social media excessively can be minimized, and healthier use can be achieved. This could be obtained by different means. For instance, if social media platforms intend to retain their users, they can help them to achieve a healthy use of the technology by giving recommendations and training on privacy management (e.g. by launching pop-up video and posts or creating quizzes with some reward). This can help them, as adverse outcomes of excessive social media can drive not only use reduction, but also quitting and switching (Maier et al., 2015; Turel, 2014). Moreover, privacy literacy can be taught at educational institutions, and government and non-government organizations, guaranteeing that social media users are aware of all risks when using the technology and have a sense of how they can protect themselves against the dangers. The usefulness of such approaches, though, should be examined in future research.

Another important conclusion of this study was the need to promote the adoption of adaptive coping strategies in unhealthy SM users. It is proven that coping strategies can be trained (Gaudioso et al., 2017). Hence, it is important for organizations to work on employee training and for schools to train students to improve the skills needed for adopting adaptive strategies. The motivation of SM providers is often the reverse; they aim to stop use reduction. Our research leads us to believe that they can do this by lessening privacy threats and helping users control their usage levels to avoid addiction and exhaustion. Based on our findings, it is reasonable to anticipate that if SM providers make the service less likely to violate users’ privacy, individuals will perceive fewer privacy concerns and be less tempted to cut back on their use, at the very least.

Limitations and future research

This work has several limitations that should be addressed. First, the majority of participants were young-adult SM users from a single European country. As a result, generalizability is constrained and has to be strengthened through replication and expansion to more groups in order to consider different sociodemographic characteristics and cultures. Second, Instagram was the sole subject of our investigation. It is important to extend our work to other SM. Third, we did not use or specify the two types of privacy in our research (institutional and social). Future research could apply the same methodology focusing more on the social perspective of privacy (worries about how others may misuse our private information). Third, we limited our analysis of the SM use reduction study to the privacy and coping strategies as drivers. Future research might combine our viewpoint with other drivers and theories to develop a comprehensive model that delineates reduction from several perspectives. Fourth, although several steps were taken to reduce social desirability bias risk, future research may consider additional techniques, such as randomized response (Fisher, 1993) and measuring and controlling for SDB tendencies (Turel et al., 2011). Finally, this study includes a limited set of controls. Additional confounding variables should be considered in future research. Examples include users’ personality traits (Ho et al., 2017), wellbeing levels (Dhir et al., 2018), purpose of use (hedonic vs work), type of platform, and life events (births, deaths) that can affect SM use and motivation to reduce SM use.

Conclusion

This work demonstrated that privacy concerns are a stressor that positively impacts the adoption of adaptive coping strategies. This one, later positively influences the adoption of SM use reduction intention. Our work showed the need to have moderator analysis. We saw that users with high levels of privacy literacy, thus more knowledgeable about data protection, are more aware of the SM risks and more predisposed to adopt safety behaviors in the SM context. Finally, we discovered a three-way interaction model, proving that a significant observed behavior is the “fatigue syndrome.” The effectiveness of adaptive coping strategies in decreasing SM use is less effective for users who are severely addicted and yet a little exhausted.

Acknowledgement

This work was supported by national funds through FCT (Fundação para a Ciência e a Tecnologia), under the project – UIDB/04152/2020 – Centro de Investigação em Gestão de Informação (MagIC)/NOVA IMS.

Disclosure statement

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

Additional information

Notes on contributors

Joana Neves

Joana Neves is a PhD student and Invited lecturer at the NOVA Information Management School (NOVA IMS). Her background is in Information Management, and she earned her bachelor’s degree and master’s from NOVA IMS. She is an invited professor, integrates research projects at the same university, and is an independent consultant in the Information Management area. Currently, her research is focused on social media corrective behaviors and privacy area.

Ofir Turel

Ofir Turel is Professor of Information Systems Management within the School of Computing and Information Systems at The University of Melbourne. He has published over 200 journal papers in leading IS journals as MIS Quarterly, JMIS, Journal of AIS. His research has been also featured in media outlets, including, The Washington Post, The Daily Mail, CBC, TV stations in Japan, South Korea, Canada, documentaries in Germany and France.

Tiago Oliveira

Tiago Oliveira is a Full Professor of Information Management and President of Scientific Council at the NOVA Information Management School (NOVA IMS). His research interests include technology adoption, Sustainable Technologies, and privacy. He has published papers in several academic journals and conferences, including European Journal of Information Systems, Information & Management, Tourism Management, Decision Support Systems, Government Information Quarterly, among others. Tiago has authored more than 250 scientific articles in Journals and conference proceedings. Tiago has more than 29,000 citations (https://scholar.google.com/citations?user=RXwZPpoAAAAJ). Tiago Oliveira was included in the prestigious 2021, 2022, and 2023 edition of the “Highly Cited Researchers” index.