Associations of early social media initiation on digital behaviors and the moderating role of limiting use

Abstract

Little is known about the effects of social media initiation on digital behaviors from middle childhood to early adolescence, a critical developmental period marked by peer influence and inaugural access to mobile devices. Participants from middle schools in the Northeast U.S. (N=773; 11-15 years, Mean = 12.6) completed a cross-sectional survey about social media initiation, digital behaviors, and parental restrictions on digital use. Descriptive results demonstrated that overall early adolescents more frequently engaged in positive digital behaviors compared to negative ones. Results from structural equation models showed that initiating social media platforms, namely Instagram or Snapchat, in later childhood (10 years or younger) was significantly associated with problematic digital behavior outcomes compared to either tween (11-12) and/or teen (13+) initiation, including having online friends or joining social media sites parents would disapprove of, more problematic digital technology behaviors, more unsympathetic online behaviors, and greater likelihood of online harassment and sexual harassment victimization. Additionally, there is evidence to show that childhood initiators demonstrated a greater tendency to engage in supportive or civically-engaged online community behaviors compared to older initiator counterparts. Parental restriction of mobile phone use and a less frequent checking of social media ameliorated some of the negative effects.

1.1. Introduction

Little is known about effects of early initiation to social technologies on psychosocial and behavioral health outcomes in early adolescence (ages 11-15), even though 95% of teens aged 13-17 have access to a smartphone, 72% use Instagram, and 69% use Snapchat (Anderson & Jiang, 2018). Popular social media sites such as Instagram and Snapchat all require a minimum age of 13 to register. Despite this federally mandated age limit due to Children’s Online Privacy and Protection Act (COPPA; Federal Trade Commission, 2002) and sporadic regulatory practices to deter under age use by social media companies, it is estimated that 3.6 million of Facebook’s 153 million monthly visitors in the U.S. are under the age of 12 (Richtel & Helft, 2011). As of 2016, a UK study found that 73% of 10-12 year olds had signed up for a social media site despite the age restrictions (Coughlan, 2016). More recently, 1.7 million Instagram users were under the age of 12 (Clement, 2019). Due to the vulnerability of this age and the prevalence of early adolescents’ social media use, it is of practical, social, and scientific significance especially in the age of Open Science Communication to explore (a) associations of early social technology initiation with psychological and digital health outcomes; and (b) whether family rules around social technology use moderate these associations.

1.1.1. Potential concerns of early social technology adoption

The federally mandated age minimum of 13 for social media use set by COPPA originates from a governmental entity (US Federal Trade Commission) rather than from science. It stems from a need to protect children from commercial interests and collection of their personal data without their knowledge rather than from a developmental rationale – a dilemma which Open Science practices can address through sharing of interdisciplinary research findings with industry partners. Behavioral scientists have hypothesized that potential consequences of early adoption could include a dependency on technologies for social interaction, particularly with increasing access to smartphones in younger groups (Elliott & Urry, 2010). Still, many federal age restriction mandates are based on societal fears about the long-term implications of Internet use with negative consequences such as early sexual priming (Subrahmanyam et al., 2004) or aggression via gaming (Irwin & Gross, 1995). There is a scarcity of studies on effects of early social media use, disentangling its potentially unique influences from other social technology use (e.g., mobiles, internet). We urgently need empirical research that examines psychosocial and behavioral health impacts of early adoption of different types of social technologies, in order to inform federal mandates with scientific evidence rather than market, policy, or industry-based motivations.

We define social technologies as digital media technologies dealing with social interactions that occur beyond the solitary browsing of internet sites. They may include text-based or internet-connected devices (e.g., iPads, smartphones) or platforms (e.g., Discord, Youtube channels, multi-user online games) designed to allow for interaction with other users. Social media has been defined as encompassing social networking sites, micro blogs (e.g., Tumblr), content sharing sites (e.g., Snapchat), blogs, Wikis, and interactive gaming sites that provide opportunities for users to share and co-construct content (Kuss & Griffiths, 2017). We define social media initiation as the period of time when teens sign up for their own account and create an online profile inviting others to join their social network. Often this can entail misrepresenting one’s birthdate or age in order to sign up. There are ongoing public conversations about when to introduce media technologies to children, with developmental psychologists warning against screen-time under the age of 2 (Zimmerman et al, 2007) and Internet use among pre-teens and teenagers (Greenfield & Yan, 2006). The American Academy of Pediatrics has developed more nuanced guidelines, recommending that for children aged 6 and over, limits placed on time spent and types of media used are predicated on whether it displaces time spent getting adequate sleep, physical exercise, and other essential healthy behaviors (American Academy of Pediatrics, 2016). However, scientific evidence and guidance is lacking on the developmental implications of initiating social technologies in the critical period of early adolescence.

1.1.2. Social technology use and early adolescence

Since the early years of mobile communication, adolescents have helped drive its mass adoption and remain the most active users today, developing their own social spheres in interactive, multimedia rich environments (boyd, 2007). In a major shift from when Facebook was the most popular social media site among teens (aged 13-17), in 2018, sites such as YouTube (85%) , Instagram (72%), and Snapchat (69%) emerged as the most frequently used sites. (Anderson & Jiang, 2018). In a recent poll, teens were asked which social media sites they preferred to be contacted for brand and commercial product engagement and 70% of them preferred Instagram, 40% preferred Snapchat, and only 10% chose Facebook (Sheetz, 2018). Social media sites are increasingly and intricately tied to adolescent daily behavior. Forty-five percent of adolescents aged 13-17 go online constantly, 44% go online several times a day, and 95% of teens from all racial backgrounds have access to a smartphone (Anderson & Jiang, 2018). However, the vast majority of studies on social technologies examine college-aged or high school populations. Rarely do studies focus on early adolescence when children begin to use social technologies more frequently.

The early adolescent years (between 10-15) are marked by pubertal development, cognitive maturation, school transitions, social identity redefinitions, and the emergence of sexuality. Early adolescence is a particularly vulnerable age due to heightened awareness of peer status, approval, and rejection (Steinberg & Silverberg, 1986). It is associated with a drop in self-esteem, weaker academic performance, and increased anxiety and competition with others (Brinthaupt & Lipka, 2002). Social technologies can facilitate the defining of a peer community or “friendship network” that can invite both positive and negative influences (Charmaraman et al, 2018; James et al., 2017), such as inspiration and closeness as well as disconnection and distress (Weinstein, 2018). Early adolescence can also include a temporary increase in parent-child conflict around autonomy and control over mobile devices, particularly when parents lag behind their adolescent’s understanding of how to access and use social technologies (Charmaraman & Grossman, 2014; Mesch, 2006). Therefore, early adolescence is a critical developmental period to study parent and peer influences on social technology use.

1.1.3. Problematic digital relationships with peers

Due to young teens’ limited capacity to self-regulate and susceptibility to peer pressure (O’Keeffe & Clarke-Pearson, 2011), the potential for engaging in unsafe online and social networking behaviors (e.g., having social media accounts their parents don’t know about, entering into online relationships with unfamiliar adults) may increase as adolescents seek autonomy and separation from their familial networks (Madden et al., 2013). Teens are particularly vulnerable to what they witness on social technologies (Brinthaupt & Lipka, 2002; Ellison et al., 2007). Exposure to risky content posted by friends can cultivate behavioral norms that are then spread through online networks and contribute to the adoption of risky attitudes and behaviors (Moreno, 2011).

In a national study of over 1500 students aged 10-15, 33% experienced online harassment in the past year, of which 9% were directly linked to a social media site (Ybarra & Mitchell, 2008). Most studies about cyberbully victimization report between 20-40% victimization rates in childhood and adolescence (Aboujaoude et al., 2015), which can vary by subgroup (e.g., gender, ethnicity, grade level, etc.). Negative peer online interactions can result in increased levels of cyberbullying via establishment of a peer group culture that rewards bullying behavior, such as through posting hurtful pictures (Dodge et al., 2006). The transient nature of Snapchat can come with both freedom of expression but also digital relationship challenges (Vaterlaus et al., 2016).

Contrary to media hype reports, adolescents who receive unwanted online sexual solicitation are typically harassed by other youth and not adults. Online peer sexual harassment (SH) has been reported in the range of 3-11.96% depending on how one defines it (Jones et al., 2012; Klettke et al., 2014). Often it is difficult to prove these online transgressions due to the ephemeral affordances of a site like Snapchat (Handyside & Ringrose, 2017). Instead of focusing on exploration of sexuality through exchange of digital content or “sexts” (Rice et al., 2018) that is more normative during adolescence, our focus in the current study is on unwanted or unwelcomed sexual comments, which have been attributed to a larger peer climate of dismissive attitudes about sexual harassment in middle school (AUTHORS, 2016).

1.1.4. Fear of Missing Out (FoMO) and problematic digital technology behaviors

The fear of missing out is a pervasive need to stay connected with other people’s activities for fear of being absent from those plans (Przybylski et al., 2013). This fear of being left behind is potentially heightened among tweens and teens using social technologies given the pervasive access to social media feeds, displaying the social content of their peer network’s lives. Several studies have proposed that FoMO serves as a mediator variable connecting psychological needs to social media engagement. For instance, FoMO mediates the link between social media engagement and indicators of wellbeing (e.g., mood, life satisfaction, online vulnerability) (Buglass et al., 2017; Przybylski et al., 2013). FoMO is also a predictor of smartphone addiction and maladaptive mobile phone use (especially for females) (Oberst et al., 2017), more frequent Facebook use, and feeling unpopular and socially isolated on Facebook (Beyens et al., 2016). Barry and colleagues (2017) posit that adolescents (aged 14-17) who experience FoMO approach social media as a way to avoid being isolated from others, as opposed to a desire to connect with others. They argue that social media use is intricately related to internalizing problems with youth who are preoccupied with being excluded. We considered whether, in addition to driving how teens interact with social technologies, FoMO may also be an unintended consequence of social media use by examining whether early initiation of social media use (i.e., middle childhood) is linked with experiences of FoMO in early adolescence.

1.1.5. Online network influences and positive online engagement

Despite ongoing fears of new technologies having negative effects on our youth, emerging research suggests that under the right conditions, social media engagement can provide both instrumental and social support (Odgers & Jensen, 2020). Adolescents with lower social and emotional wellbeing, such as those with severe depressive symptoms, are more likely to report going online to their peers on social media for emotional support (Rideout & Fox, 2018). In fact, high Instagram and Snapchat use in early adolescence has been found to predict greater close friendship competence and perceived peer support (Vannucci & McCauley Ohannessian, 2019). Social technology use also provides potential opportunities to exercise online civic engagement (Van Hamel, 2011) and prosocial online behaviors (Erreygers et al., 2017). A study of 979 11-17 years old youth found that greater levels of online civic engagement was significantly associated with less harassment perpetration and greater likelihood to help as a bystander (Jones & Mitchell, 2016). In addition to examining links between age of social technology initiation and negative digital behaviors, the current study also considers the links between age of social technology initiation and positive digital behaviors such as positive social media use and sympathetic online behaviors.

1.1.6. Associations between parent restrictions and adolescents’ social media use

Parental monitoring has shown protective influences for teens’ social media use, while lack of monitoring increases the likelihood that a student will be an online cyber perpetrator (Hemphill & Heerde, 2014). A longitudinal study (Hemphill & Heerde, 2014) following students from age 10 to 19 showed that predictors of adolescent cyberbullying perpetration included lack of clear parental rule-setting. Given that parents can intervene, discourage, or exacerbate cyberbullying (Spears et al., 2009), parental restriction of adolescents’ online behavior may represent an important prevention target.

One aspect of parental monitoring associated with lower levels of negative social media behaviors is “restrictive mediation,” which encapsulates parent rule-setting to limit teens’ social media use (Collier et al., 2016). A review of research found that restrictive mediation of children’s social media use can decrease children’s and adolescents’ time spent using social media and their exposure to mature content (Collier et al., 2016). Other studies found that parental restrictions on teens’ social media use were associated with lower teen engagement in risky online activities (Lee, 2012; Notten & Nikken, 2016). However, overly restrictive rules after children enter adolescence could also backfire, with tweens reporting anger at parents on imposing strict limits (Evans et al., 2011). With few exceptions (e.g., Badenes-Ribera et al., 2019) most studies of parental restrictive mediation for social media use focus on middle to late adolescence, while the role of parents in teens’ social media use during middle childhood and early adolescence is not well understood.

A recent study demonstrated that obtaining a smartphone at a later age (e.g., 13 instead of 10) was associated with positive behavioral health outcomes, such as increased sleep time and earlier bedtimes (Authors, 2021). Intricately related to social media rule-setting are parental attitudes towards smartphone use (i.e., appropriate age to get a first smartphone), the need for restrictions on their use (Rideout, 2015), and the tensions that can arise when tweens perceive their first smartphone as a way to assert their personal autonomy (Weinstein & Davis, 2015). As the predominant mode of accessing the internet changes from co-viewing tablets with parents to more private use of personal mobile devices, increased conflict with parents restricting tablet use (Beyens & Buellens, 2017) can spillover into the realm of smartphone use. Given growing rates of smartphone addiction among adolescents (e.g., Lee & Lee, 2017), research is needed to address how parents monitor smartphones specifically. One study found that parents who were concerned about their teens’ smartphone use and had efficacy to intervene were more likely to use restrictive mediation to regulate their teens’ smartphone use (Hwang et al., 2017). Studies of Asian teens found that lower levels of restrictive mediation were associated with higher smartphone addiction (Chang et al., 2019; Lee & Kim, 2018). Additionally, more frequent social media use has been associated with more frequent parent-child conflict and social isolation within families, which may be exacerbated during the early adolescent years due to the negotiating of social media boundaries at this critical developmental stage (Dworkin et al., 2018; Vannucci & McCauley Ohannessian, 2019).

The effectiveness of parents’ restrictions in preventing negative consequences may in part depend on the manner in which parents communicate their restrictions to their children and adolescents. For instance, discussing media content with the intention of respecting a child’s autonomy and sense of agency (as opposed to being controlling) can lead to higher acceptance of the value of the rules themselves (Fikkers et al., 2017; Valkenburg et al., 2013) and more prosocial online behavior (Meeus et al., 2018). Further research is needed to understand when is the most effective time for parents to establish rules for social media use, which may be intricately tied to social media frequency. In the current study, we consider whether parental restrictive monitoring, defined as setting limits on the amount of time teens are allowed to use their phone or internet each day, ameliorated the links between social technology initiation and problematic digital behaviors.

1.1.7. Theoretical background

Social technology behaviors are not individual acts in isolation but rather choices made within larger social contexts (Bronfenbrenner, 1989). During this critical developmental period of seeking autonomy from parents and exploring sexuality and identity with peers (Erikson, 1968), the peer social context becomes especially critical. According to the transformation framework (Nesi et al., 2018), the social media context affords adolescents a natural field to practice their peer relationship building skills through self-disclosure, digital communication, and providing social support. On the flipside, this social context can also be a digital space that exacerbates adolescent tendencies toward negative self-evaluations, risky peer norms, and unsympathetic online behaviors. Since early adoption of media habits is linked to later consumption (Anderson et al., 2001), parental influences that limit exposure to social technologies for younger children may be associated with a lowered risk of unhealthy amounts of consumption in older children.

The popularity of different social media sites among young teen users continues to change, therefore studies need to continually assess and adapt to these changes. A social media affordances framework is useful for understanding how the key attributes of social media platforms can influence adolescent functioning due to distinct usage features that are under constant evolution (Moreno & Uhls, 2019). Affordances are design attributes that suggest to social media users how these platforms could be used, such that how a social media site (e.g., Snapchat) is perceived to have particular affordances (e.g., time-limited posts) results in attracting certain types of users and may invite behaviors that fulfill different developmental needs (e.g., privacy when trying new identities; boyd, 2007; Karahanna et al., 2018; Moreno & Uhls, 2019). For instance, social affordances of hashtags provide users a method to influence social connections and publicness of posts.

Based on this developmental affordances framework of social media, the current study focuses on Instagram and Snapchat when determining when a participant initiates social media for the first time for the following reasons: a) the massive popularity of image-based social media sites of Instagram and Snapchat among tweens and teens; b) the affordances of Instagram and Snapchat facilitate identity formation and novel opportunities for frequent, immediate, and salient displays of friendship and feedback-seeking in a perceived safe space (boyd, 2007; Nesi et al., 2018; Throuvala et al., 2019) in more salient ways than other popular sites (e.g., Tik Tok, YouTube); c) specific affordances of Instagram and Snapchat are hypothesized to be associated with more frequent social media use. For example, the feature of “streaks” in Snapchat, highlighting the number of consecutive days online friends exchange content, has been found to increase the likelihood of more frequent and compulsive involvement with social media (Griffiths, 2018).

1.1.8. Gaps in prior literature

Despite evidence that teens under age 13 are accessing social media sites in rapidly growing numbers (Valkenburg et al., 2005), much of the existing academic research on social technology use has focused on older adolescents and young adults (Valkenburg & Peter, 2007), who are well past the critical developmental period when social technology usage begins (Davis et al., 2020). A recent review (Odgers & Jensen, 2020) confirmed that this early adolescent period has been largely neglected in prior research, despite this developmental period being highly relevant to understanding transitions into digital environments. Although there have been emerging studies that examined tweens and younger teens social technology use (e. g., Valkenburg & Peter, 2007) and parental mediation behaviors in this population (e.g., Meeus et al., 2019), the role of social media initiation among young teens and tweens in predicting later digital behaviors is sparse. Investigating positive and negative associations with early social media initiation may provide some insights into developmental benefits and tradeoffs depending on the user’s age and personal characteristics. A recent review of the benefits and costs of social media in adolescence (Uhls et al., 2017) recommended that future research examines the best age to begin using social media, under what restrictions, and what factors might moderate the relations between social media and outcomes. Additionally, the vast majority of studies have focused primarily on negative impacts of digital technologies during adolescence (Odgers & Jensen, 2020). This is the first study of its kind to cross-sectionally examine both positive and negative associations of social media initiation focusing specifically on the 2 most popular U.S. sites (Instagram and Snapchat) in three age subgroups (childhood, tween, and teen), and the moderating role of parental restrictive mediation of phones and internet use.

The aims of the present study were to a) investigate the relationship between the age period of social media initiation(e.g., child, tween, or teen) and social technology behavior outcomes and b) whether parent restrictions on technology use or frequency of checking social media moderated the relationships between age period of initiation and outcomes. We explored the following hypotheses: H1: The younger the age of initiation into social media, the greater likelihood of negative and less likelihood of positive social technology outcomes; H2: Parent restrictions on social technology use will buffer the negative associations of early initiation into social media; H3: The greater the frequency of checking social media, the more negative associations will be found from early initiation of social media.

1.2. Method

Middle schools in the Northeastern United States were recruited based on varying school enrollment size (e.g., small, medium, or large), 100% Internet access in school, and diverse racial/ethnic composition (e.g., a minimum of 30% of students who identified as racial/ethnic minorities). After obtaining IRB from our institution and school district-level permissions, we worked closely with school liaisons to distribute parent informed consent/opt-out forms (in English, Spanish, and Portuguese) through paper flyering, parent email listservs, and school e-newsletters. The school liaisons helped coordinate proctored, pre-scheduled advisory period class times lasting up to 60 minutes to administer the online Qualtrics survey entitled, “Adolescent Social Media Use, Health, and Parental Monitoring Project.” Students used school-provided Chromebook devices at one middle school and used program-provided Chromebooks during pre-scheduled break times during afterschool programs at the three other schools. The survey took approximately 25-40 minutes for students to complete, depending on whether the student was an active social media user or online game player. There were skip patterns built into the survey to personalize the type of questions displayed according to their usage (e.g., if they didn’t initiate social media yet, they were skipped out of social media specific questions). Since survey links were emailed to students, those who were absent during the survey administration were still able to participate in the study from home. Participation rates ranged from 42% in the afterschool programs to 91% in the whole school data collection, which includes a 5.7% parental opt-out or absence on the day the survey was administered for the whole school. An honorarium was provided to schools for participating as well as incentives to teachers to help with survey administration (e.g., gift cards). Students who took the survey were given an embossed pen from [INSTITUTION], snacks, and were entered in a raffle for a chance to win a $25 gift card.

1.2.1. Participants

A total of 733 surveys were collected from students in grades 6-8 in four suburban middle schools in the Northeast region of the United States. Among youth who completed the surveys, 49.1% were male, 49.7% were female, and 1.2% identified as “Other” gender. Participants self-identified as White (53.6%), Latinx (17.3%), Black (11.1%), Asian (4.9%), Multiracial (5.0%), and Native American (2.6%). An additional 5.4% of participants declined to report their race/ethnicity. To avoid generating estimates based on non-representative samples, only racial groups that included at least 10% of the sample were modeled individually in analytic models. All other groups were combined into a single “Other” category. The average participant was 12.6 (SD=.96) years old with 32.2% of participants in 6^th grade, 35.3% in 7^th grade and 32.1% in 8^th grade. Twenty-four percent of participants received free or reduced price lunch and 68.6% lived in a two-parent household. Table 1 provides descriptive statistics for the sample.

Table 1.

Sample Description (N=773)

	Mean (SD) or Percent
Gender
Male	49.1
Female	49.7
Other	1.2
Age	12.57 (.96)
Grade
6th	32.3
7th	35.3
8th	32.1
Race/Ethnicity
White	53.6
Latinx	17.3
Black	11.1
Asian	4.9
Multiracial	5.0
Native American	2.6
Declined to provide	5.4
Free/Reduced Price Lunch	23.2
Two-parent Household	68.6

1.2.2. Measures

The online Qualtrics survey asked students to report demographic characteristics as well as questions regarding their social technology behaviors. A series of scales were used to measure the constructs of interest: Parent Influences, Problematic Digital Technology Behaviors, Problematic Digital Relationships, Online Network Influences, and Positive Online Engagement.

Demographic Characteristics

Age was calculated by subtracting students’ self-report date of birth from the date of survey administration. Students were able to identify their gender as male (1) female (0) or “Other.” Given the extremely small prevalence of individuals selected “Other,” we recoded these responses to missing to reduce the risk of reporting erroneous or biased results regarding this small group of respondents. Students reported their race ethnicity by self-selecting which racial categories best described them. Responses included: White, Hispanic, Black, Asian, and Native American. Students selecting a single category were categorized accordingly. Students selecting more than one racial category were classified as Multiracial. Students who selected Hispanic were classified as Hispanic regardless of whether they selected an additional racial/ethnic category. Students reported whether they received free or reduced price lunch (yes/no/I don’t know). Responses of “No” and “I don’t know” were combined. Lastly, a variable identifying whether the student lived in a two parent household was created using participants’ reports of who lived in their home. Those indicating they lived with two adult caregivers were coded as “1” and all other household structures were coded as “0.”

Age of Snapchat or Instagram Initiation

Students who indicated whether they joined Snapchat and/or Instagram were asked to report the age they joined. Using the earliest age students reported joining Snapchat or Instagram, a measure of Age of Snapchat or Instagram Initiation was generated. This variable assigned an age of earliest initiation for each student who used Snapchat or Instagram. Participants were then assigned into one of three age of initiation groups: initiation at 10 years or younger (Child Initiators), initiation between 11 or 12 years (Tween Initiators), or initiation at 13 years or older (Teen Initiators).

Parent Disapproval

Parent Disapproval of Online Friends was assessed using a single item that asked participants whether they had anyone in their social media network their parent(s) would disapprove of (yes/no). Parent Disapproval of Social Media Site was measured by asking participants whether they joined a social media site their parents would disapprove of (yes/no).

Problematic Digital Technology Use

Adapting McDaniel and Radesky’s (2018) Parent Problematic Digital Technology Use Scale for our adolescent sample, Problematic Digital Technology Use (α=.68) use was measured using their full three-item scale assessing students’ perceptions of whether the frequency of their engagement with their smartphone is problematic and intrusive. For example, students were asked to respond to the statement, “When my mobile phone alerts me to indicate new messages, I cannot resist checking them.” The response scale for these items ranged from 1 (“Strongly Disagree”) to 5 (“Strongly Agree”).

Problematic Digital Relationships

Fear of Missing Out (FoMO) (α=.60) was measured using three items from Przybylski and colleagues’ (2013) Fear of Missing Out (FoMO) scale, which assesses students’ fears of being excluded from activities with friends. For example, students responded to the statement, “I get worried when I find out my friends are having fun without me.” Each item was accompanied by a four-point scale ranging from 1 (“Not at all true of me”) to 4 (“Extremely true of me”).

Online Sexual Harassment (α=.67) was measured using two items from the Online Sexual Harassment Scale (Ybarra et al., 2007), which asks how often one receives unwanted messages with sexual content via text message or social media (e.g., “Someone tried to get you to talk about sex online when you did not want to.”) Students responded to each item using a scale ranging from 1 (“Never”) to 4 (“Often”).

Online Harassment (α=.80) asked participants about how often someone else made rude or mean comments online, someone else spread rumors online, or the participant was hurt by someone excluding them online. These three items were reported on a 4-point scale ranging from 1 (“Never”) to 4 (“Often”).

Online Network Influences

Students were classified as being Friends with Known Adults if they reported being friends with parents, aunts and uncles, teachers or coaches on social media. A value of 1 was assigned to those who reported being friends with known adults and 0 if they were not friends with known adults. Similarly, Friends with Celebrities was assigned a value of 1 if they reported being friends with sports celebrities, actors/actresses, fashion or beauty bloggers or health and fitness bloggers. Students who were not friends with these groups were assigned a value of 0. Number of Friends on Social Media was measured using a single item “About how many friends do you have on that site?” The 6-point scale ranged from 1 (“Less than 50”) to 6 (“Over 1000”).

Positive Online Engagement

Positive Social Media Use (α=.62) was measured using five items about uses of social media that encourage supportive or civically-engaged online community behaviors. These items include, “Help with your homework or a class project;” “Like or comment about friends sharing good news;” “Dislike or comment about friends sharing bad or sad news;” “Organize an event with your friends or community;” and “Raise awareness about a social issue you care about.” The response scale for these items ranged from 1 (“Never”) to 5 (“Always”).

Sympathetic Online Behavior (α=.70) was measured using three items related to having empathy for others online by avoiding any intentionally exclusionary or hurtful online behaviors, such as putting embarrassing photos of friends online, being mean or rude to others, or leaving someone out. Students used a scale ranging from 1 (“Very much like me”) to 5 (“Not at all like me”) to respond to each of these items, with higher scores indicating more Sympathetic Online Behaviors.

Parent Restrictive Mediation

Students responded to a single item that asked, “On a typical school day, what is the MAXIMUM amount of time your parents allow you to go online or use your phone?” Eight response categories were provided: “30 minutes or less,” “1 hour,” “2 hours,” “3 hours,” “4 hours,” “5 hours,” 6 or more hours,” and “No limit.” This variable was coded such that more restrictive monitoring received higher scores. For example, responses of “30 minutes or less” were given a score of “7,” while “No limit” was given a score of “0.”

Frequency of Checking Social Technology

Students were asked how often they check social media on a typical school day. Responses ranged from 1 (“Never”) to 6 (“More than every hour”).

1.2.3. Analytic Approach

The links between Age of Snapchat or Instagram Initiation and Parent Influences, Problematic Digital Technology Use, Problematic Digital Relationships, Online Network Influences and Positive online Engagement were examined within a Structural Equation Modeling (SEM) framework for the nine outcomes that were measured using multiple indicators (FoMO, Sexual Harassment, Sympathetic Online Behavior, Online Harassment, Problematic Digital Technology Use, Check Social Media, Parent Limit Setting of Technology Use, Number of Friends on SM, Positive Social Media Use). In these SEM models, the outcomes were treated as latent variables, which allowed us to account for measurement error within the modeling process. Prior to conducting analyses, we reviewed the distribution of each outcome to determine the most appropriate modeling strategy. The large majority of our outcomes had skew values between −.5 and .5 and were modeled using a Gaussian distribution with Maximum Likelihood (ML) estimation. Two of our outcomes (FoMo and Sympathetic Online Behaviors) had skew values outside of the −.5 to .5 range. These two outcomes were modeled using ML estimation with robust standard errors and a Satorra-Bentler correction.

The four remaining outcomes (Parent Disapproval of Online Friends, Parental Disapproval of Social Media Site, Friends with Known Adults and Friends with Celebrities) were dichotomous indicators and were therefore tested within a logistic regression framework. To reduce the risk of omitted variable bias, all models included a series of covariates (gender, race/ethnicity, age, free/reduced price lunch eligibility as a proxy for socioeconomic status, and two-parent household), that have previously been associated with the manner in which youth engage with social technologies.

As previously described, the indicator representing age of social technology initiation consisted of three categories representing Childhood, Tween, and Teen initiation. This variable was dummy coded and Childhood initiators were used as the omitted referent group. After interpreting the models in which the Childhood initiators were used as the omitted referent, we re-ran all models changing the omitted referent to the Tween initiator group to ensure comparisons among all age groups were considered. To examine whether restrictive mediation (e.g. parent “limit setting”) and frequency of checking social media moderates the assocations of early social media initiation, we created interaction terms between limit setting and age of social technology initiation and frequency of checking social media and age of social media initiation. In all analyses, missing data were minimal – 69.9% of participants had no missing data and an additional 18.3% were missing only one or two data points. For models tested within an SEM framework, missing data were managed using Full Information Maximum Likelihood (FIML) estimation. In the logistic regression models, listwise deletion was used to manage missing data. All analyses were conducted in the lavaan package (version 0.6-3) in R (Rosseel, 2012). Before interpreting the results of SEM models, we required the models met at least three of the following four criteria in regard to fit indices: Comparative Fit Index (CFI) >= .90, Tucker-Lewis Fit (TLI) >= .90, Root Mean Square Error of Approximate (RMSEA) <= .08, Standardized Root Mean Square Residual (SRMR) <= .05. All models met these criteria with two exceptions. While all models met our RMSEA and SRMSR requirements, our model of FoMO demonstrated a CFI = .86 and TLI = .77, our model of Sympathetic Online Behaviors demonstrated a CFI=.89 and TLI=.82. Findings from these two models are considered preliminary.

1.3. Results

Descriptive Statistics

Ninety-two percent of students reported owning a smartphone and 73.7% of the sample had joined either Instagram or Snapchat. Out of the participants who had already received their first smartphone, the age at which this occurred ranged from age 8 or younger to 14 (M = 10.40, SD = 1.35). Participants reported the reasons why they first began using a personal smartphone with the majority of the participants indicating that it was for parents to reach them wherever they were (70%) or in case of emergencies (53%). About a third of students reported that they began using a smartphone because they had started middle school (30%), or had asked for it (27%). A small portion of students reported they began using a smartphone because they received one as a hand-me-down from a family member (12%) or due to peer pressure (11%). Forty-three percent of the sample was younger than 10 years old when they joined Instagram or Snapchat (Childhood Initiators), 48.4% joined between 11 and 12 years (Tween Initiators), and 9.1% joined at 13 or older (Teen Initiators). When asked why participants first started a social media account, half of them indicated the desire to share things they enjoyed with their friends (50%) and a third of them indicated that their peers were already on it (34%), they were curious to see what everyone was posting (33%), and to stay in touch with relatives (32%). Relatively fewer responses concerned meeting new people/making new friends (17%), having access to online games (14%), and being required to sign up (3%). Participants were asked about all of the social media sites they had joined. Table 2 provides a complete description of social media sites youth joined for the full sample and by age period of social media initiation.

Table 2.

Percent of students in the full sample and Childhood, Tween, and Teen Initiator groups who have used each social media site

	Full Sample (N=773)	Childhood Initiators (N=242)	Tween Initiators (N=276)	Teen Initiators (N=52)
	Percent	Percent	Percent	Percent
YouTube	83.6	94.2	91.3	86.5
Snapchat	68.0	98.8	87.3	71.2
Instagram	66.9	94.6	85.9	84.6
House Party	39.5	63.6	43.5	34.6
Tik Tok	37.6	53.7	44.6	23.1
Pinterest	23.9	29.3	27.9	26.9
Twitter	20.8	36.8	21.0	15.4
WhatsApp	18.1	23.1	19.6	21.2
Twitch	17.7	22.7	19.2	23.1
Discord	14.7	16.9	15.6	21.2
Vsco	13.8	25.2	14.5	7.7
Facebook	12.5	19.0	12.3	26.9
ooVoo	10.1	21.1	6.5	5.8
Steam	9.2	8.3	11.2	9.6
Reddit	9.1	11.2	10.1	7.7
Tumblr	8.4	12.4	8.0	13.5
Kik	6.3	11.6	4.7	1.9
Zepeto	2.8	5.8	1.4	0.0
Viber	1.7	2.1	2.2	1.9

Age of Initiation and Digital Behaviors

A recent meta-analysis of the role of social technology on adolescent development (Odgers & Jensen, 2020) revealed published effect sizes ranging between .06-.11. Therefore, we consider any effect sizes of .05 or greater to be practically meaningful. In general, our models accounted for a small to moderate amount of variance in digital behaviors, with R² values ranging from .07 – .15. To isolate the associations of age of initiation, we observed the change in R² value that occurred when we compared models without the categorical indicators of age of social media initiation to models that contained the categorical indicators. Change in R² values suggested that the age of social media initiation had small effects on digital behaviors ranging from .01 – .05. Overall R² and change in R² values for each model can be found at the bottom of Tables 4a and 4b. Childhood initiators were more likely to have an online friend their parent would disapprove of and experience less parent limit setting compared to both tween (Parent Disapproval of Friend: B (Unstandardized Coefficient)=−.40, OR (Odds Ratio)=0.67, p-value < .01, Parent Limit Setting: B=.66, p<.01) and teen (Parent Disapproval of Friend: B=−.69, OR=0.50, p < .01, Parent Limit Setting: B=1.89, p<.001) initiators. Childhood initiators were also more likely to have joined a social media site their parent would disapprove of compared to tween initiators (B=−.32, OR=.73, p<.05).

Table 4a.

Results of SEM and logistic regression models predicting Parent Influences, Problematic Digital Technology Use and Problematic Digital Relationships by Age of Social Media Initiation

	Parent Factors			Social Technology Use		Problematic Digital Relationships
	Parent disapproves of online friend	Parent disapproves of social media site	Parent Limit Setting of Tech Use	Social Media Freq	Problematic Digital Tech Use	Fear of Missing Out	Sexual Harassment	Online Harassment
	Est. (SE)	Est. (SE)	Est. (SE)	Est. (SE)	Est. (SE)	Est. (SE)	Est. (SE)	Est. (SE)
Tween initiators	−0.40(0.13)**	−0.32(0.16)*	0.66(0.20)**	−0.49(0.13)***	−0.32(0.16)*	−0.14(0.07)*	−0.19(0.10)	−0.18(0.11)
Teen initiators	−0.69(0.25)**	−0.60(0.35)	1.89(0.36)***	−1.27(0.24)***	−0.60(0.35)	−0.18(0.10)	−0.47(0.23)*	−0.59(0.22)**
Age	0.18(0.07)*	0.08(0.08)	−0.50(0.11)***	0.34(0.07)***	0.08(0.08)	−0.02(0.04)	0.09(0.06)	0.07(0.06)
Free/Reduced Price Lunch	−0.01(0.16)	0.15(0.18)	0.09(0.24)	0.08(0.16)	0.15(0.18)	0.06(0.09)	0.17(0.11)	0.01(0.14)
Female	−0.17(0.13)	−0.03(0.16)	−0.36(0.19)	0.36(0.13)**	−0.03(0.16)	0.42(0.07)***	0.04(0.08)	0.43(0.10)***
Black	0.07(0.21)	0.33(0.23)	0.64(0.34)	−0.59(0.23)*	0.33(0.23)	−0.18(0.12)	−0.13(0.16)	−0.45(0.20)*
Latinx	0.44(0.18)*	0.32(0.21)	0.40(0.28)	−0.40(0.19)*	0.32(0.21)	−0.02(0.11)	0.12(0.12)	−0.14(0.15)
Race other	0.41(0.18)*	0.45(0.21)*	0.29(0.27)	−0.60(0.18)**	0.45(0.21)*	−0.06(0.09)	−0.08(0.12)	−0.14(0.15)
Two-parent Household	−0.22(0.14)	−0.34(0.16)*	0.62(0.22)**	0.07(0.15)	−0.34(0.16)*	−0.08(0.07)	−0.13(0.09)	−0.30(0.12)*
Comparing Tween and Teen Initiators
Teen initiators	−0.29(0.25)	0.28(0.36)	1.23(0.35)***	−0.78(0.24)**	−0.26(0.15)	−0.03(0.10)	−0.28(0.20)	−0.41(0.21)
Change in R2	.04	.03	.05	.03	.01	.01	.04	.03
R2	.11	.11	.10	.12	.07	.15	.08	.11

Est.=Unstandardized Coefficient, (SE)=Standard Error

Change in R² = Increase in R² when age of initiation was added to the model

R² = Variance accounted for in the full model

Table 4b.

Results of SEM and logistic regression models predicting Online Network Influences and Positive Online Engagement by Age of Social Media Initiation

	Online Network Influences			Positive Online Engagement
	Friends with Known Adults	Friends with Celebrities	Number of Friends on SM	Positive Social Media Use	Sympathetic Online Behavior
	Est. (SE)	Est. (SE)	Est. (SE)	Est. (SE)	Est. (SE)
Tween initiators	−0.01(0.12)	−0.22(0.12)	−0.74(0.14)***	−0.12(0.06)*	0.28(0.06)***
Teen initiators	0.10(0.22)	−0.21(0.21)	−1.65(0.28)***	−0.23(0.10)*	0.45(0.10)***
Age	−0.09(0.06)	0.20(0.06)**	0.62(0.08)***	0.08(0.03)*	−0.09(0.03)**
Free/Reduced Price Lunch	−0.20(0.15)	0.00(0.15)	0.08(0.18)	0.04(0.07)	0.01(0.08)
Female	0.39(0.11)***	0.40(0.11)***	−0.30(0.13)*	0.33(0.07)***	−0.01(0.05)
Black	−0.28(0.21)	−0.10(0.21)	−0.26(0.27)	−0.21(0.10)*	0.08(0.10)
Latinx	−0.16(0.17)	−0.30(0.17)	0.07(0.20)	−0.02(0.08)	−0.06(0.09)
Race other	−0.32(0.16)*	−0.27(0.16)	−0.55(0.19)**	−0.08(0.08)	0.00(0.08)
Two-parent Household	0.16(0.13)	0.18(0.13)	−0.06(0.16)	−0.13(0.06)*	0.13(0.07)*
Comparing Tween and Teen Initiators
Teen initiators	0.12(0.22)	0.01(0.21)	−0.91(0.28)**	−0.12(0.10)	0.17(0.08)*
Change in R2	.01	.01	.09	.02	.09
R2	.09	.10	.23	.19	.12

Est.=Unstandardized Coefficient, (SE)=Standard Error

Change in R² = Increase in R² when age of initiation was added to the model

R² = Variance accounted for in the full model

Childhood initiators also reported more Problematic Digital Technology behaviors than tween and teen initiators including more frequent checking of social media than tween (B=−.49, p < .01) and teen initiators (B=−1.27, p < .001) and higher Problematic Digital Technology Use than tween initiators (B=−.32, p<.05). For Problematic Digital Relationships, childhood initiators were more likely to experience FoMO compared to tween initiators (B=−.13, p<.05) and reported more instances of online sexual harassment and online harassment compared to teen initiators (Online Sexual Harassment: B=−.47, p<.05 and Online Harassment: B=−.59, p<.01). Childhood initiators reported having more friends on their social media network compared to tween (B=-.74, p<.001) and teen (B=−1.65, p<.001) initiators.

For Positive Online Engagement, childhood initiators reported more Positive Social Media Use, but lower Sympathetic Online Behaviors compared to tween (Positive Social Media Use: B=−.12, p<.05 and Sympathetic Online Behaviors: B=.23, p<.001) and teen initiators (Positive Social Media Use: B=−.23, p<.05 and Sympathetic Online Behaviors: B=.42, p<.001). Tween initiators experienced less parent limit setting (B=1.23, p<.001), were more frequently checking social media (B=−.78, p<.01), had more friends in their social media network (B=−.91, p<.01) and had lower Sympathetic Online Behaviors (B=.20, p<.05) than teen initiators. Table 3 provides descriptive statistics for all outcomes included in the present study.

Table 3.

Descriptive results for the overall sample and Childhood, Tween, and Teen Initiator groups

	Full Sample (N=276)	Childhood Initiators (N=773)	Tween Initiators (N=242)	Teen Initiators (N=52)
	Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)
Parent limit setting of technology use	3.06(2.38)	2.42(2.12)	3.04(2.25)	3.90(2.64)
Checking social media	3.78(1.92)	4.71(1.48)	4.31(1.57)	3.77(1.65)
Problematic Digital Technology Use	2.61(1.01)	2.76(.98)	2.72(.96)	2.57(.91)
Fear of Missing Out	1.61(.62)	1.76(.63)	1.62(.61)	1.59(.58)
Sexual Harassment	1.25(.55)	1.37(.68)	1.25(.52)	1.10(.30)
Online Harassment	1.48(.66)	1.63(.73)	1.52(.65)	1.30(.56)
Number of friends on SM	2.96(1.59)	3.51(1.55)	2.79(1.46)	2.69(1.26)
Positive Social Media Use	2.19(.73)	2.43(.74)	2.25(.65)	2.14(.79)
Sympathetic Online Behavior	4.49(.75)	4.23(.85)	4.52(.69)	4.69(.50)
	Percent	Percent	Percent	Percent
Parent disapproval of online friends	20.8	28.6	17.7	14.6
Parent disapproval of social media site	10.3	15.7	9.1	6.1
Friends with known adults	42.0	48.9	49.2	49.0
Friends with celebrities	50.3	63.8	56.4	60.8

Tables 4a and 4b provide the complete results of the models examining the links between age of initiation and Parent Influences, Social Technology Use, Problematic Digital Relationships, Online Network Influences and Positive Online Engagement.

1.3.1. Exploring moderators: Parent limit setting and frequency of checking social media

Results of models examining whether parental limit setting can moderate the potentially negative associations of early social media initiation presented mixed findings. In some instances, parent limit setting appears to attenuate the link between social media initiation and digital behaviors. For example, childhood initiators with high limit setting reported more connections with known adults than childhood initiators with low parental limit setting. However, among tween initiators, this relationship was reversed, with higher parental limit setting being associated with fewer connections with known adults (B=.13, p < .05). Similarly, parental limit setting was associated with lower rates of FoMO among childhood initiators (B=.10, p < .05), but the relationship was reversed among tween and teen initiators – parental limit setting was actually associated with higher rates of FoMO among tween and teen initiators (see Figure 2).

Figure 2.

Age of social media initiation, parent limit setting, and feelings of FoMO

In the models examining whether the frequency of checking social media moderates the link between childhood social media initiation and digital behaviors, results suggested lower rates of checking social media were associated with less online harassment for all youth (see Figure 3). In addition, childhood initiators who frequently checked social media were the most likely to experience online harassment (B=.16, p < .05).

1.4. Discussion

The main objective of the study was to explore if the age at initiating Instagram and Snapchat is associated with positive or negative early adolescent digital behaviors, particularly since these social media platforms are most popular at this age. The overall pattern of findings suggested that the oldest initiator group (age 13+) was associated with the most frequent parent limit setting, the fewest online friends in networks, fewer problematic digital technology behaviors, less problematic digital relationships, and higher sympathetic online behaviors. This suggests that the industry-based age of 13 as a “cut-off’ minimum in allowing the youngest users access to the dominant social media sites in our society through COPPA may potentially be a good minimum standard given the preliminary state of research in this area, considering that childhood (10 and younger) and tween initiators (aged 11 and 12) did demonstrate significantly more problematic digital behaviors and relationships and less sympathetic behaviors than their 13 year old and older initiator counterparts. More longitudinal research that can delineate which age is most developmentally advantageous to initiate social media use is needed to further understand these age cut-offs.

In the case of problematic digital behaviors (e.g., problematic digital technology use, frequency of checking social media and problematic digital relationships), childhood initiators more frequently reported this negative outcome than tweens, teens, or both, suggesting more vulnerability during this transitional period between late elementary school and early middle school. In some cases, tween and teen initiators were more alike than the childhood initiators and in other analyses, tweens were more akin to the childhood initiators, demonstrating that there are no clear “cut-off” ages when problematic behaviors suddenly arise. This could partially be explained by the variation in when a personal mobile device has been introduced to the child/tween. A 2016 Digital Trends Study revealed that the average age when children get their first mobile device is 10.3 years. In addition, prior research demonstrates that problematic digital technology use during early adolescence predicts conduct problems over time and eventually issues with self-regulation (George, et al., 2018). Our cross-sectional findings corroborate other research that there could be a relationship between early digital technology use and susceptibility to risky and addictive online behaviors.

Contributing to scholarship that examines both the risks and opportunities of online behaviors (e.g., Charmaraman et al., 2018; Erreygers, et al., 2017; Ito et al., 2020), our descriptive results illustrate that overall, young adolescents were engaging in more positive and sympathetic online behaviors compared to problematic online behaviors, regardless of age of initiation. Our results revealed that childhood initiators were more likely than their older counterparts to engage in positive online behaviors, such as socially supportive social media posts, fostering awareness of social issues, or organizing events through social media. It could be that beginning social media at a younger age increases the likelihood that users will become more familiar with potentially prosocial uses of social media platforms. The longer someone has to create and build their online networks, the more likely they are to capitalize on its powers of connection and spreading information expediently. As found in previous literature (e.g., Jones & Mitchell, 2016), positive online engagement is associated with less harassment perpetration and increased helpful bystander actions, which can have implications for incorporating more online civic behaviors within digital citizenship education.

Our second goal focused on the moderating role of parental restriction on mobile phone and internet use and whether it can affect the negative consequences of early Instagram and Snapchat initiation. It could be possible that the socializing influence of parents restricting use may be associated with a hesitancy to actively engage and connect with others on social media platform for older social media initiators, due to past fears or dismissive attitudes, therefore these later initiators may not readily see benefits from its use. Results of the role of parental restriction were mixed, and suggested that parental limit setting may act differently for childhood, tween, and teen initiators. For example, although childhood initiators reported the highest rates of FoMO, parental limit setting was associated with lower rates of FoMO in this group. However, among teen initiators, parental monitoring was actually linked with higher rates of FoMO. Perhaps teens who are allowed to use their phones more often experience less FoMO simply because they are not “missing out” as they are keeping in more frequent contact with their peers. While limit setting may act differently for child and teen initiators, later initiation in our sample is linked with lower reports of FoMO, regardless of the level of parental limit setting reported. While high rates of checking social media were associated with more frequent experiences of online harassment for all groups, childhood initiators who frequently checked social media reported the highest rates of online harassment of any group. We cannot ascertain the directionality of these findings, given our correlational data -- it could be that youth who are being harassed online are checking social media more frequently in an effort to monitor or push back against online harassment. Regardless of the direction of this relationship, these preliminary findings raise cause for concern, especially in light of results suggesting childhood initiators were significantly more likely to engage in secretive online behaviors that they kept hidden from their parents. Overall, these findings underscore the role of parental knowledge of children’s social technology behaviors that is critical at this developmental stage.

In all, these emerging findings suggest that a potential strategy to support families with children, tweens, and teens is to a) keep track of social media sites joined and online friend networks, b) set even one rule about screen use (i.e., limiting duration of use on school nights), and c) monitor children’s frequency of checking, particularly if using social media at age 10 and younger. Additionally, restrictions on use are but one way to socialize healthy use -- parents could also model their own self-regulated use of social media in order to reduce any potential for mixed messaging. Finally, although some monitoring is advisable to prevent negative impacts, overly restrictive policies may be ill advised if social technologies are being used as a valuable source of social or emotional support (especially at a young age) or if these digital skills are critical for new online, civic, and interpersonal skills of the future (Odgers & Jensen, 2020).

1.4.1. Limitations and Future Directions

Given that this study was a convenience sample from the Northeast, we acknowledge that the findings may not generalize to a national adolescent population. It is important to note that the racial/ethnic composition of our school-based sample (53.6% White) is more multi-ethnic compared to the population of the state where data was collected (69.6% White) and in the U.S. population (61.6% White) (U.S. Census, 2021). As with most studies relying on adolescents to report on their own social technology use, self-reports of digital behaviors are often not well correlated with objective measures of usage, such as phone logs (George, et al., 2018). Asking an adolescent when he or she first signed up for a particular social media site can be subject to potential errors in recall, especially the longer time period it has been since initiation. However, collecting data during the middle school years when the majority of adolescents are beginning to join these sites helps reduce this possibility.

Since this is the first cross-section study of this kind to examine the behavioral associations of early social media initiation, future longitudinal research is needed to a) replicate these findings in a different sample and b) understand bidirectional influences of why some children join social media sites earlier than others and whether they are at-risk for problematic digital behaviors due to individual factors that are present at the outset or whether the online behaviors play a role in negative outcomes. Another future direction would be to explore whether the adolescent’s personality or ability to self-regulate media use or whether the socializing influence from the family context came before the initiation of technology behaviors. Future research might consider analyzing multiple reports of phone- or internet-limiting rules (i.e., from child’s and parent’s point of view) since in our prior qualitative research on family media use, parents and children often did not match in their reports of what rules are actually understood and enforced in the household (AUTHORS, 2018).

Although a few of the scales were shortened versions of the original scale, our exploratory study focused on measuring as many potentially critical variables as possible for greater possibility of testing alternative hypotheses (Widaman et al., 2011). Other measures of parental restriction beyond phone or internet limit setting, such as parental mediation styles (e.g., active, surveillant, etc.) could explain the strength and direction of the moderating role. Prior research demonstrates that rule-setting alone may not be as effective in preventing problematic digital relationships as cultivating supportive relationships at home and at school (Davis & Koepke, 2016). Given that caveat, our goal was to isolate a specific digital access rule within a household rather than a parental media restriction “style” or “preferences,” in order to tease out a practical finding that could be of interest to parents, educators, practitioners, and youth development professionals.

Figure 1.

Conceptual Model

Figure 3.

Age of social media initiation, frequency of checking, and rates of online harassment

Highlights.

• Early adolescents more frequently engage in positive digital behaviors compared to negative ones.

• Using Instagram or Snapchat before age 11 is related to more problematic digital behaviors.

• Youngest social media initiators are more likely to engage in supportive online behaviors.

• Limiting access to social media lessened some negative effects of early social media use.