Parental Attitudes and Predictors of Support for Youth-Directed Social Media Legislation in the United States
- Open Access
- 28-08-2025
- Original Paper
Gepubliceerd in:
Abstract
Delen
Parent Predictors of Support for Youth-Directed Social Media Legislation in the United States
Across the globe, people are using social media at ever increasing rates. With over 4.59 billion users worldwide (out of a total population of just over 8 billion), social media have become a ubiquitous part of daily life (Statista, 2022). Chief among these frequent social media consumers are children and adolescents. In the United States alone, 90% of adolescents aged 13-19 reported having previously used social media, and 75% of those adolescents reported having an active profile on at least one site (American Academy of Child & Adolescent Psychiatry, 2018). Indeed, 45% of adolescents in the U.S. report being online “almost constantly” (Anderson & Jiang, 2018). Similarly, across 25 European countries, 38% of adolescents aged 9-12 years and 77% of adolescents aged 13–16 years have a profile on at least one social media platform (Lee et al., 2016). Even more, 91.5% of youth in South Korea over the age of 3 years utilize the internet in some capacity (Lee et al., 2016).
While the trajectory of social media use has been quite similar across many countries around the world, this has not necessarily translated into similar governmental responses, particularly regarding regulatory efforts. For example, South Korea, with amendments to the Youth Protection Act in 2018, imposed stricter age verification practices onto social media companies (Statutes of the Republic of Korea- Youth Protection Act, 2023). Further, the act places restrictions on the type of content users under the age of 19 have access to, including that which is sexually explicit or promotes self-harm. Similarly, the United Kingdom’s recently passed Online Safety Bill aims to provide a legal framework for online safety that is to include stricter age verification as well as impose a “duty of care” on social media companies, holding them responsible for the safety of adolescent users (Department for Digital, Culture, Media, and Sport, 2022). Conversely, the United States has put notably few regulations in place over the last two decades regarding child and adolescent social media use. However, this is rapidly changing.
As of August 2023, over 20 states have proposed or enacted some version of a bill designed to regulate child and adolescent use of social media. While many of these bills share overarching goals related to protecting young people online, they cover a myriad of topics, which vary by state. For example, in states such as Louisiana, Missouri, and a joint bill from Alabama, Connecticut, and Hawaii, the focus is largely on imposing stricter age verification to prevent users under the age of 13 from accessing social media platforms, in accordance with COPPA (National Conference of State Legislatures, 2023). In other states, such as California, New Jersey, and North Carolina, the bills center on the design of social media itself and highlight growing concerns over algorithms and individual platform features and affordances as they relate to adolescent users’ safety and well-being. Mississippi, by contrast, recently voted to enact a bill aimed at preventing pornographic media exposure to children via social media. Many of these regulatory attempts (such as Utah’s Social Media Regulation Act) contain a mixture of many of the above provisions. The Utah bill, for example, aims to strengthen age verification and increase parental tools to restrict adolescent social media use (such as allowing parents to turn off social media at a certain time of night). The bill also allows parents to access their children’s social media content, meaning that parents could gain access to everything that their child posts and interacts with on social media.
Many of these bills come in response to growing academic and societal concern over the negative effects social media use may have on young people’s well-being, both in the United States and abroad. Although the extant literature is mixed regarding the effects of social media on aspects of adolescents’ mental health (e.g., Cingel et al., 2022; Valkenburg et al., 2022), and some studies show positive, negative, and null effects that differ from individual to individual (e.g., Valkenburg et al., 2021), the popular press generally reports a negative effect (e.g., Edwards & Jackson 2023). For example, social media use has been found to be associated with poor sleep, increased anxiety and depression, and low self-esteem among a sample of Scottish adolescents (Woods & Scott, 2016). Similarly, across two surveys of 506,820 adolescents in the United States between 2010 and 2015, those who spent more time on social media were more likely to report mental health issues than their peers who reported spending more time with non-screen activities (Twenge et al., 2018). Indeed, a May 2023 advisory issued by then United States Surgeon General, Dr. Vivek Murthy, highlights the intensity of national concern over the potentially harmful effects of social media use on children’s and adolescents’ well-being (Office of the Surgeon General, 2023). The advisory describes both the potential negative outcomes associated with social media use (depression, anxiety, viewing violent and sexual content), as well as the increased vulnerability of young people due to less advanced cognitive development – and thus an increased need for protection.
Given the relative novelty of this legislation in the United States, researchers have yet to explore the extent to which parents of children and adolescents are, firstly, aware of the bills, but also whether or not they are supportive and believe in the potential for their child’s well-being to be positively impacted. As such, the goals of this paper are twofold. First, we aim to describe parents’ overall levels of awareness and support for nine major provisions which cut across the over 20 state bills. Second, we aim to utilize parent and family demographic variables, as well as attitudes, parenting styles, and political affiliations in order to predict support for each of the nine provisions. The provisions, outlined below, are an amalgamation of the main tenets common to one or more state’s proposed or enacted bills.
In order to fill these important knowledge gaps, we conducted a survey among parents of children and adolescents aged 8-17-years-old (N = 1359) in the United States. This study has several important implications, particularly in the areas of parenting, media, and public policy. First, we are able to examine whether parents believe that these laws will be beneficial for their child and provide a suitable remedy to a growing area of national concern. Second, we are able to determine which types of parents, based on personal characteristics and values, are in support of these laws and how level of support varies by provision. These findings can therefore be used by policy makers to ensure that regulatory efforts align with public opinion – specifically parents of children under 18 as they, unlike their children, are able to vote. This study is an initial examination of support, and the predictors of such support, for a growing area of social media regulation in the United States and around the world, with implications for child and adolescent well-being.
Primary Provisions Across State Bills and Laws
Due to the large quantity and breadth of topics covered by the various states’ bills, we chose to extract recurring components which cut across much of the legislation into nine primary provisions by conducting a state-by-state search of social media bills aimed at minors (see Procedure and Participants). The first of which imposes stricter age verification requirements to prevent those who are under 13 from creating and maintaining accounts. The second of which requires social media companies to obtain parental consent for anyone aged 13 to 17 to sign up for an account. The third provision allows parents to set limitations on the hours of access for adolescent users’ accounts. Provision four provides full parental access to accounts held by minors, including all content that is posted and interacted with on that account. The fifth provision of focus is regarding data privacy and limits the obtaining and sharing of adolescent users’ personal information and account data by social media companies, both on the platform and elsewhere online. The sixth provision aims to prevent companies from giving targeted ads to users under the age of 18. Provision seven focuses on banning the use of algorithms and design features which may promote harmful content on adolescent users’ feeds (i.e., sales of weapons or drugs, self-harm, etc.). Provision eight requires that social media companies block all pornographic content from users under the age of 18. Finally, the ninth provision requires blocking foreign company apps entirely to protect user data from those companies. For a summary of which states have included which provisions within their proposed or enacted bills, see Supplementary Materials. As these provisions cover a wide variety of topics, below, we review literature on how parent demographics, moral foundations, parenting style, and political affiliation might relate to support for social media regulation and justify why these are key variables to consider in the context of social media policy.
Demographic Predictors of Support
Concern surrounding the negative effects of social media use on child and adolescent well-being is shared by many parents across the United Sates. However, certain facets of parents’ identities – such as gender, race/ethnicity, religiosity, education, and income likely influence how parents perceive the threat of social media to their child’s well-being as well as the role parents believe government should take in protecting their child from the effects of social media use via regulation. Indeed, prior research has consistently linked various facets of identity, such as race/ethnicity and gender, to opinions on public policy (Hayes et al., 2021). For example, parents with more resources that allow them to be engaged with their children, be it economic (income; Lee et al., 2009), relational (marital status; Vandervalk et al., 2004), or through community (e.g., religiosity), may be less likely to see a need for increased government regulation. In contrast, various identities such as gender or race/ethnicity may shape the expectations that parents have in terms of government involvement in their parenting or their attitudes towards their role as a parent. Race/ethnicity and gender identity have consistently been linked to attitudes towards public policy (Gay & Tate, 1998), and often are intersecting identities that shape perceptions of the need for, and potential effectiveness of, various policy adjustments. These attitudes are often understood to be shaped by an individual’s social privilege and experiences of systemic inequality – which in the United States are inextricably tied to both race/ethnicity and gender (Nawyn & Gjokaj, 2014). Along with the likely influence of these parental characteristics, social media use and effects often differ as a function of a child’s age and gender, and therefore may also influence the level of importance a parent puts on regulation (Rideout et al., 2022).
Moral Foundations as Predictors of Support
A key theory of moral decision-making, Moral Foundations Theory (MFT; Haidt & Joseph, 2008) has often been used to understand how exposure to certain media content can make certain moral foundations more salient, or mentally accessible, among children (Cingel et al., 2023a), adolescents (Joeckel et al., 2013), and young adults (Tamborini et al., 2010). MFT posits that humans have five salient moral foundations relating to care (toward others), fairness (or justice), loyalty (toward one’s in-group), (respect for) authority, and purity (disgust or concerns of contamination; see Graham et al., 2013), and the relative salience or importance placed on these foundations shapes decision making. Tamborini (2012) extended this through the Model of Intuitive Morality and Exemplars (MIME) to argue that, from a selective exposure perspective, the foundations most salient to an individual should shape media selections, as individuals select media content featuring morals that align with their more salient moral foundations. Recently, Cingel et al. (2023b) found that parents’ moral foundations predicted their attitudes and motivations toward using both educational and entertainment media with their young children (ages 3–7). In this paper, we extend this research to a new media context (i.e., social media) and an older child age group by examining the ways that parents’ MFT variables predict support of each of the nine provisions included in the state social media bills. The inclusion of parents of older children is important due to the increased prevalence of social media use as adolescents get older. Indeed, the average time of social media use for 17-year-olds in the United States is a reported 5.8 h, over an hour more than the average of 4.1 h/day for 13-year-olds (Rothwell, 2023). However, it is important to note that parents of younger adolescents are not necessarily excluded from the demands of monitoring their child’s social media use, given that nearly 40% of 8-12-year-olds also report being on at least one platform (Rideout et al., 2022).
Parenting Style Predictors of Support
Another potential factor influencing how parents perceive social media regulations may be general parenting style. Often in developmental research, parenting styles are understood as predictors of child outcomes (Weiss & Schwarz, 1996). In this study, however, we recognize that parenting styles also reflect parents’ philosophy on what children and adolescents need in order to thrive. Specifically, parenting styles are often conceptualized as the intersection of two broad dimensions of parenting: parental support and parental control (Kuppens & Ceulemans, 2019). Parental support refers to the affective nature of the parent-child dynamic, capturing concepts such as parent involvement, acceptance, emotional availability, responsiveness, and warmth (Cummings et al., 2000). The second dimension, parental control, refers to parents’ efforts to regulate/manage the child’s behavior via rule setting, rewards/punishment, and supervision. Parental control of children’s behaviors, especially early in development, is associated with positive outcomes for children in most contexts (Baumrind, 1991), to the point where a lack of parental control is considered maladaptive.
In the intersection of these two dimensions – parental support and control – researchers have identified four broad parenting styles: indulgent, authoritarian, authoritative, and neglectful (Baumrind, 1991). Indulgent parenting is characterized as highly supportive, but low on control. Parents who adopt an indulgent style of parenting often believe that children need support and warmth to thrive and can set their own boundaries. In contrast, authoritarian parenting is characterized as being low on parental support, but high on control. In contrast to indulgent parenting, parents who adopt an authoritarian parenting style often believe that children require firm guidance and strict boundaries to develop into well-adjusted members of society. The third parenting style, authoritative, is characterized as being high on both support and control. These parents are often described as being supportive while still maintaining appropriate boundaries – scaffolding children as they grow and encouraging autonomy. Parents who adopt an authoritative parenting style see themselves as supportive guides for their children – providing the necessary structure to allow children to thrive while encouraging autonomy. The fourth parenting style, neglectful, is characterized on being low on support and control. It is possible that parenting styles, and the general attitudes towards the parenting role they convey, is linked to broader attitudes towards macro-level interventions to parenting via policy. For example, a more authoritarian parent may be less supportive of regulation towards adolescent social media use because they feel that they should have more control over their parenting and child. In contrast a more authoritative parent may want some additional structure, but nothing they feel is too invasive of their child’s privacy. Importantly, the inclusion of parenting style as a potential predictor of support gives us insight into who the participant is in relation to their child. While the other predictor variables are either parent (i.e., demographics, moral salience, and political identification) or child (i.e., age and gender) specific, parenting style examines who the individual is specifically in their role as parent.
Political Identification as Predictors of Support
Finally, we identified political identity as another potentially influential factor toward support for each of the nine provisions. While the current literature does not address how political identity relates specifically to the content of each provision, we do know that political party affiliation is commonly used to construct policy preferences more generally (Goren, 2005). For many decades, the partisan influence hypothesis has been used to explain the metaphorical “screen” based on political party affiliation through which individuals’ view any given issue (Campbell et al., 1960; Bisgaard & Slothuus, 2018). Indeed, Goren (2005) found that party identification significantly affected participants’ attitudes toward the issue of limited government (among others). Oftentimes, this is a result of policy being framed in ways that contain signals to alert the public of which political values are at stake. Because the values which are deemed important differ by political party (Capara et al., 1999; Luna et al., 2020), we would expect to see partisan differences in levels of support for each of the nine provisions, based on which values are being signaled as at stake within them. Similarly, these core political values are thought to develop early in adulthood and be continuously reinforced (Feldman, 1988; Brandt et al., 2019), thus remaining relatively stable throughout the lifespan. We would then, for example, not expect to see as high of levels of support for regulation of any kind from individuals’ belonging to a political party which does not typically value government involvement.
The Present Study
Despite extensive research examining social media’s effects on adolescent well-being (Woods & Scott, 2016; Twenge et al., 2018) and growing legislative efforts to regulate adolescent social media use across the United States (National Conference of State Legislatures, 2023), there remains a significant gap in our understanding of how parents view these regulatory measures. While previous work has examined parental attitudes toward media regulation more broadly (Cookingham & Ryan, 2015), no studies to date have systematically investigated parental awareness of and support for specific social media legislation provisions, nor identified the demographic, moral, parenting style, and ideological predictors of such support. We therefore pose the following research questions:
RQ1: Which parent demographic variables (gender, race/ethnicity, religiosity, and income) and child demographic variables (age and gender) relate to parent support for each of the nine provisions?
RQ2: How do parents’ (a) care, (b) fairness, (c) loyalty, (d) authority, and (e) purity salience relate to support of each of the nine provisions?
RQ3: How do (a) indulgent (b) authoritarian (c) authoritative and (d) neglectful parenting styles relate to support of each of the nine provisions?
RQ4: How does parent (a) political affiliation and (b) strength of affiliation relate to support of each of the nine provisions?
These research questions allow us to examine how different aspects of parents’ identities, values, and approaches to childrearing may relate to their support for social media regulation. By investigating demographic characteristics, moral foundations, parenting styles, and political identity, we can better understand the factors from different, important areas of parents’ lives that may shape attitudes toward these legislative efforts.
Method
Participants and Procedure
Parents of children and adolescents aged 8–17 were recruited across the United States. Following ethical approval at University of California, Davis [ID 2074523-1] the Qualtrics survey was fielded August - September 2023. Parents were contacted through the survey company Dynata, via their online panel portal or via email, and invited to take part in a research study. Dynata sources survey respondents through three channels: Loyalty (brand partnerships), Open (web/mobile recruitment), and Integrated (publishers/social networks). Prior to answering any survey questions, we procured consent to participate. The survey took approximately 10–15 min to complete. After completing the survey, we thanked participants for their time and participants received compensation commensurate with Dynata policies. In total, 1701 participants completed the survey, but participants who completed fewer than 95% of questions (thus having large amounts of missing data), or who missed one or more attention checks, were excluded (N = 342). The final sample included 1359 participants.
The sample was comprised of the following groups: 71.44% White (N = 972), 3.61% Asian (N = 49), 13.54% Black (N = 184), 10.22% Hispanic (N = 139), 0.81% American Indian or Alaskan Native (N = 11), 0.15% Native Hawaiian or Pacific Islander (N = 2), and 0.15% chose not to respond (N = 2). The sample included 31.63% fathers (N = 430) and 67.9% mothers (N = 925). Additionally, 0.15% participants (N = 2) selected ‘prefer not to say’ regarding gender, while 0.15% chose not to respond (N = 2). Nearly two-thirds of participants (63.8%; N = 867) held a college degree or higher. The median household income bracket was $75,000–$99,999. Parents specifically reported on one child in their household between the ages of 8 and 17 years (Mchildage = 12.75 years; 53.42% boys, N = 726; 45.33% girls, N = 616; 0.29% transgender, N = 4; 0.59% non-binary or gender queer, N = 8). Based on 2020 Census data (U.S. Census Bureau, 2021), the sample closely matched the US population for White (71.44% vs 71.1%) and Black (13.54% vs 12.4%) representation but underrepresented Hispanic (10.22% vs 18.9%) and Asian (3.61% vs 6.1%) parents. Notable demographic skews include overrepresentation of mothers (67.9% vs ~50%; U.S. Census Bureau, 2022). Participants who had more than one child in this age range were asked to answer based on their child whose birthday came next.
The nine included provisions emerged from a search of every state’s legislation for key words related to social media bills for minors (e.g., “adolescent social media use”, “social media minors”, etc.). From this compiled list, two of the authors read through the bills, highlighting the main provisions included in each and then cross-referencing bills for commonalities.
Measures
Social media bills
General awareness
In order to measure participants’ general awareness of the current legislation (both proposed and recently passed), we began with the prompt “A growing number of States have introduced legislation regarding adolescent social media use. We are going to ask you questions about specific parts of these bills”. Following this, participants were asked to “Please indicate your overall level of awareness of the social media bills”, with response options ranging from 1 = Fully not aware to 5 = Fully aware.
Support for provisions
Participants’ level of support for each of the nine provisions was ascertained by providing a brief description of each provision, such as “Social media companies are required to verify that anyone seeking to maintain or open a social media account is at least 13 years of age”. This was followed with the prompt “Please indicate to what extent you support or oppose this specific part of the bill”. Response options ranged from 1 = Strongly oppose to 5 = Strongly support.
Attitudes toward provision effectiveness
The extent to which participants believe that each of the nine provisions will improve children’s and adolescents’ well-being was measured by providing a brief description of each of the nine provisions, such as “Social media companies are required to obtain consent of a parent or guardian before someone between the age of 13 and 17 may open or maintain an account”. Following this, participants responded to the prompt “Please indicate to what extent you agree that this part of the bill has the potential to improve child and adolescent well-being”, with response options ranging from 1 = It will have a strong negative impact on child and adolescent well-being to 5 = It will have a strong positive impact on child and adolescent well-being.
Independent variables
Demographics
In order to assess for a variety of demographic variables, participants responded to the following questions. Participant gender was collected by asking “How do you identify?”, with response options including 1 = Female, 2 = Male, 3 = Transgender, 4 = Non-binary/gender queer, 5 = Prefer to self-describe, and 6 = Prefer not to say. Religiosity was measured by asking participants “Are you religious?”, with response options of 1 = Yes and 2 = No. In addition, we assessed socioeconomic status by asking participants “In which of the following ranges does your total household income (before taxes) fall?”, with response categories including, 1 = Less than $25,000, 2 = $25,000 to $49,999, 3 = $50,000 to $74,999, 4 = $75,000 to $99,999, 5 = $100,000 to 124,999, 5 = $125,000 to $149,999, and 6 = $150,000 or more. Participants indicated racial background by answer “How do you identify?” with response options of 1 = Hispanic, 2 = White, 3 = Black, 4 = Asian, 5 = AIAN (American Indian/Alaskan Native), and NHPI (Native Hawaiian/Pacific Islander). Finally, participants who had more than one child between 8 and 17 years old were asked to answer the survey with one child in mind. This child’s age was assessed by asking participants “How old is [insert child’s name as parent previously entered it]?” with response options ranging from 1 = 8 years and 10 = 17 years.
Moral foundations
Moral foundation salience was assessed through the Moral Foundations Questionnaire (Graham et al., 2011). This measure includes 30 items, however, for our purposes, participants answered half due to the length of the survey. Participants responded on a 7-point scale ranging from not at all relevant to extremely relevant. Participants were asked, “When you decide whether something is right or wrong, to what extent are the following considerations relevant to your thinking?” Items included phrases such as “whether or not someone suffered emotionally” and “whether or not some people were treated differently than others”. Reliability for the five subscales were as follows: Care α = 0.75, Fairness α = 0.78, Loyalty α = 0.78, Authority α = 0.69, and Purity α = 0.71.
Parenting style
Parenting style was assessed using the Parenting Style Scale (Saunders et al., 2012). This measure consists of 19 total items (5 items each for Indulgent, Authoritative, and Authoritarian; 4 items for Neglectful). Participants responded on a 5-point scale ranging from never to always. Participants were asked to “Please indicate how often you do the following…” on items such as “I let my child express feelings about being punished or restricted” (Indulgent) and “I am consistent with my discipline techniques” (Authoritative). Reliability for the four parenting subscales were as follows: Indulgent α = 0.79, Authoritarian α = 0.81, Authoritative α = 0.78, and Neglectful α = 0.83. The sample was comprised of 47.8% (n = 650) Indulgent, 34.9% (n = 474) Authoritative, 14.8% (n = 201) Authoritarian, and 2.5% (n = 34) Neglectful.
Political identification
In order to ascertain political identification, participants were asked to “Please select the option that best describes your political party affiliation”. Response options included 1 = A strong Democrat, 2 = A not very strong Democrat, 3 = Independent, lean toward Democrat, 4 = Independent (close to neither party), 5 = Independent, lean toward Republican, 6 = A not very strong Republican, and 7 = A strong Republican. Our sample was comprised of: Strong Democrat (n = 282, 20.75%), Not Very Strong Democrat (n = 180, 13.25%), Independent, lean Democrat (n = 128, 9.42%), Independent (Neither; n = 262, 19.28%), Independent, lean Republican (n – 156, 11.48%), Not Very Strong Republican (n = 139, 10.23%), Strong Republican (n = 211, 15.53%). One participant (0.1%) chose not to report political identification.
Strength of political affiliation
In addition, the strength of political affiliation was measured by asking participants “How would you rate yourself on this scale?”, with response options including 1 = Very liberal, 2 = Somewhat liberal, 3 = Middle of the road, 4 = Somewhat conservative, and 5 = Very conservative. Our sample was comprised of: Very liberal (n = 171, 12.58%), Somewhat liberal (n = 211, 15.53%), Middle (n = 547, 40.25%), Somewhat conservative (n = 265, 19.50%), and Very conservative (n = 163, 11.99%). Two participants (0.1%) chose not to report strength of political affiliation.
Analytic Approach
To address the first aim of the paper of describing parents’ overall levels of awareness and support for the legislation, we calculated the mean and standard deviation of both general awareness surrounding the bills as well as level of support and belief surrounding the potential for impacting children and adolescents’ well-being for each of the nine provisions. To address the second aim of predicting support for each of the nine provisions data were analyzed using nine hierarchical linear regressions. First, to examine links between demographic variables and parental support for the first provision regarding age verification, we ran a linear regression with support for provision one as the dependent variable. Demographic variables included in this block of the model were parent gender, religion, income, race, child gender, and child age. The second block of this regression included the five moral foundations (care, fairness, loyalty, authority, and purity). The third block of the regression included the four parenting styles (indulgent, authoritative, authoritarian, and neglectful). For the fourth and final block of the regression, political identification and strength of political affiliation were added in. This process was repeated, substituting only the dependent variable for perceived level of support for each provision.
In addition, to further consider child age, we ran a MANOVA with parents split into two groups based on whether their child is legally allowed to be on social media (aged 13–17) or not (aged 8–12). Results showed a non-significant effect of child age on parental support across provisions (Wilks’ Λ = 0.0091, F(9, 1328) = 1.35, p = 0.206). Mean levels of parental support for each provision, based on child age, are reported in supplementary files.
Results
We first calculated the mean and standard deviation for parental awareness of the current efforts to implement social media legislation generally. Overall, participants scored at about the midpoint of the scale for awareness of the legislation (M = 3.14, SD = 1.20). Parents were found to be somewhat to strongly supportive of each of the nine provisions. The provisions regarding blocking pornographic content from underage users and placing restrictions on algorithms and platform design received the highest levels of support. Conversely, provisions regarding age verification to ensure users are at least 13 years old, requiring parental consent for users 13 -17 to create and maintain accounts, and blocking certain apps in order to protect data from foreign companies received the lowest levels of support. For a comprehensive summary of these results, please see Table 1. Similarly, we also calculated the mean and standard deviation to describe parents’ overall beliefs regarding the potential impact each of the nine provisions may have on children and adolescents’ well-being. These are reported in supplementary files. Generally, we found that parents believed that most of the provisions would have a positive impact on children’s and adolescents’ well-being. Finally, we report correlations between the main study variables in supplementary files. Additionally, the correlations between the dependent variables can be found in the supplementary files.
Table 1
Mean and standard deviations for general awareness of legislation and overall support for each provision
M | SD | |
|---|---|---|
General Awareness | 3.14 | 1.20 |
Age Verification | 4.29 | 1.10 |
Parental Consent | 4.39 | 0.97 |
Hours of Access | 4.43 | 0.93 |
Full Parental Account Access | 4.42 | 0.97 |
Data Privacy | 4.47 | 0.88 |
Targeted Ads | 4.46 | 0.91 |
Algorithms and Platform Design | 4.56 | 0.84 |
Pornographic Content | 4.68 | 0.81 |
Blocking Foreign Company Apps | 4.39 | 0.97 |
Age Verification
We ran the first regression to predict parents’ level of support toward age verification. The overall model was significant, F (8,1337) = 7.29, p < 0.01, R2 = 0.103. For a full summary of these results, please see Table 2. The first block containing demographic variables was significant. Parent gender, religion, income, race, and child age positively and significantly related to support for age verification such that mothers, those higher in religiosity and income, White parents (compared to non-White parents), and parents with older children showed higher levels of support. We then added in moral foundation saliences to the next step of the regression series. The overall model was significant, with care, fairness, and authority all positively and significantly relating to support for age verification, and loyalty negatively relating. In the third step of the regression series, we added parenting style, and while the overall model was significant, none of the four parenting styles were individually significant within the model. Finally, we added in variables related to political identity. The overall model was not significant, showing that neither political party identification nor strength of political affiliation significantly relate to support for age verification.
Table 2
Hierarchical Linear Regression Results for Age Verification, Parental Consent, and Data Privacy Provisions
Age Verification | Parental Consent | Data Privacy | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
Independent Variables | Model 1 | Model 2 | Model 3 | Model 4 | Model 1 | Model 2 | Model 3 | Model 4 | Model 1 | Model 2 | Model 3 | Model 4 |
Demographics | ||||||||||||
Parent Gender | 0.22*** | 0.20*** | 0.26*** | |||||||||
Religion | 0.09*** | 0.09*** | 0.03 | |||||||||
Income | 0.07*** | 0.04*** | 0.04*** | |||||||||
Hispanic | 0.1 | 0.37** | 0.20 | |||||||||
White | 0.30** | 0.40*** | 0.24** | |||||||||
Black | −0.01 | 0.39*** | 0.10 | |||||||||
Child Gender | 0.02 | 0.02 | 0.02 | |||||||||
Child Age | 0.02** | −0.01 | −0.00 | |||||||||
MFT | ||||||||||||
Care | 0.12** | 0.03 | 0.19*** | |||||||||
Fairness | 0.11* | 0.12** | 0.14*** | |||||||||
Loyalty | −0.09 | −0.04 | −0.10** | |||||||||
Authority | 0.14*** | 0.18*** | 0.13*** | |||||||||
Purity | −0.06 | −0.05 | −0.12*** | |||||||||
Parenting Style | ||||||||||||
Indulgent | 0.09 | 0.08 | 0.19*** | |||||||||
Authoritative | 0.07 | 0.13** | 0.14** | |||||||||
Authoritarian | 0.06 | 0.09* | 0.01 | |||||||||
Neglectful | −0.04 | −0.09 | −0.11*** | |||||||||
Political Identity | ||||||||||||
Political Identification | 0.01 | 0.001 | 0.001 | |||||||||
Strength of Affiliation | 0.01 | 0.06** | 0.04 | |||||||||
R2 | 0.042 | 0.029 | 0.023 | |||||||||
∆R2 | 0.046 | 0.014 | 0.001 | 0.063 | 0.037 | 0.005 | 0.094 | 0.060 | 0.005 | |||
Parent Consent
The next regression predicted parents’ level of support toward requiring parent consent to create and maintain an account. The overall model was significant, F (8,1337) = 5.01, p < 0.01, R2 = 0.134. For a full summary of these results, see Table 2. The first block containing demographic variables was significant. Parent gender, religion, income, and race positively and significantly related to support for parental consent such that mothers, those higher in religiosity and income, and White, Black, and Hispanic parents (compared to non-White, non-Black, and non-Hispanic parents) showed higher levels of support. We then added in moral foundation saliences to the next step of the regression series. The overall model was significant, with fairness and authority positively and significantly relating to support for requiring parental consent. In the third step of the regression series, we added parenting style. The overall model was significant, with authoritative and authoritarian parenting styles positively and significantly relating to support, and neglectful parenting style negatively relating. Finally, we added in variables related to political identity. The overall model was significant, with strength of political affiliation positively and significantly relating to support, such that more strongly conservative parents showed higher levels of support for requiring parental consent.
Data Privacy
We ran a third regression to predict parents’ level of support toward the provision regarding adolescent user data privacy. The overall model was significant, F (8,1337) = 3.87, p < 0.01, R2 = 0.182. For a full summary of these results, see Table 2. The first block containing demographic variables was significant. Parent gender, income, and race positively and significantly related to support for age verification such that mothers, those with higher income, and White parents (compared to non-White parents) showed higher levels of support. We then added in moral foundation saliences to the next step of the regression series. The overall model was significant, with care, fairness, and authority all positively and significantly relating to support for age verification, and loyalty and purity negatively and significantly relating. In the third step of the regression series, we added parenting style. The overall model was significant, with indulgent and authoritative parenting styles positively and significantly relating to support, and neglectful negatively and significantly relating. Finally, we added in variables related to political identity. The overall model was not significant, showing that neither political party identification nor strength of political affiliation significantly relate to support for increased measures regarding adolescent user data privacy.
Blocking Targeted Advertisements
We ran the fourth regression to predict parents’ level of support toward blocking targeted advertising for users under 18. The overall model was significant F (8,1337) = 5.42, p < 0.01, R2 = 0.140. For a full summary of these results, see Table 3. The first block containing demographic variables was significant. Parent gender, income, and race positively and significantly related to support such that mothers, those with higher income, and White parents (compared to non-White parents) showed higher levels of support. We then added in moral foundation saliences to the next step of the regression series. The overall model was significant, with care, fairness, and authority all positively and significantly relating to support for blocking targeted ads, and loyalty negatively and significantly relating. In the third step of the regression series, we added parenting style. The overall model was significant, with indulgent parenting style positively and significantly relating to support, and neglectful parenting style negatively and significantly relating. Finally, we added in variables related to political identity. The overall model was not significant, showing that neither political party identification nor strength of political affiliation significantly relate to support for blocking targeted advertisements for users under 18.
Table 3
Hierarchical Linear Regression Results for Blocking Targeted Ads, Hours of Access, and Complete Parent Account Access Provisions
Targeted Ads | Hours of Access | Parent Account Access | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
Independent Variables | Model 1 | Model 2 | Model 3 | Model 4 | Model 1 | Model 2 | Model 3 | Model 4 | Model 1 | Model 2 | Model 3 | Model 4 |
Demographics | ||||||||||||
Parent Gender | 0.30*** | 0.14** | 0.25*** | |||||||||
Religion | 0.02 | 0.10*** | 0.11*** | |||||||||
Income | 0.05*** | 0.04*** | 0.04*** | |||||||||
Hispanic | 0.30 | 0.11 | 0.10 | |||||||||
White | 0.23* | 0.10 | 0.20 | |||||||||
Black | 0.10 | 0.04 | 0.15 | |||||||||
Child Gender | 0.04 | 0.03 | 0.04 | |||||||||
Child Age | −0.00 | −0.01 | −0.01 | |||||||||
MFT | ||||||||||||
Care | 0.14*** | 0.10 | 0.04 | |||||||||
Fairness | 0.13*** | 0.14*** | 0.10 | |||||||||
Loyalty | −0.06 | −0.10* | −0.01 | |||||||||
Authority | 0.12** | 0.20*** | 0.22*** | |||||||||
Purity | −0.12*** | −0.10* | −0.10** | |||||||||
Parenting Style | ||||||||||||
Indulgent | 0.21*** | 0.16*** | 0.10* | |||||||||
Authoritative | 0.03 | 0.05 | 0.10* | |||||||||
Authoritarian | 0.03 | 0.20*** | 0.20*** | |||||||||
Neglectful | −0.09*** | −0.110** | −0.11*** | |||||||||
Political Identity | ||||||||||||
Political Identification | 0.02 | 0.01 | 0.00 | |||||||||
Strength of Affiliation | −0.02 | 0.03 | 0.10*** | |||||||||
R2 | 0.031 | 0.023 | 0.035 | |||||||||
∆R2 | 0.070 | 0.038 | 0.001 | 0.094 | 0.060 | 0.005 | 0.075 | 0.061 | 0.013 | |||
Parental Ability to Limit Hours of Access
We ran the fifth regression to predict parents’ level of support toward the provision allowing parents to limit hours of access for adolescent users. The overall model was significant F (8,1338) = 2.66, p < 0.01, R2 = 0.182. For a full summary of these results, please see Table 3. The first block containing demographic variables was significant. Parent gender, religion, and income were positively and significantly related to support such that mothers and those with higher levels of religiosity and income showed higher levels of support. We then added in moral foundation saliences to the next step of the regression series. The overall model was significant, with care, fairness, and authority all positively and significantly relating to support for allowing parents to limit hours of access, and loyalty and purity negatively and significantly relating. In the third step of the regression series, we added parenting style. The overall model was significant, with indulgent and authoritarian parenting styles positively and significantly relating to support, and neglectful parenting style negatively and significantly relating. Finally, we added in variables related to political identity. The overall model was not significant, showing that neither political party identification nor strength of political affiliation significantly relate to support for allowing parents to limit hours of access for adolescent users.
Parent Full Access to Adolescents’ Accounts
We ran the sixth regression to predict parents’ level of support toward allowing for full parental access to adolescent users’ accounts. The overall model was significant, F (8,1336) = 6.03, p < 0.01, R2 = 0.184. For a full summary of these results, see Table 3. The first block containing demographic variables was significant. Parent gender, religion, and income positively and significantly related to support for age verification such that mothers and those higher in religiosity and income showed higher levels of support. We then added in moral foundation saliences to the next step of the regression series. The overall model was significant, with authority positively and significantly relating to support for allowing full parental access to adolescents’ accounts and purity negatively and significantly relating. In the third step of the regression series, we added parenting style. The overall model was significant, with indulgent, authoritative, and authoritarian parenting styles positively and significantly relating to support, and neglectful negatively and significantly relating. Finally, we added in variables related to political identity. The overall model was significant. Strength of affiliation was positively and significantly related to support, such that more conservative parents showed higher levels of support for allowing parents full access to their child’s account.
Algorithms
We ran the seventh regression to predict parents’ level of support toward regulating the use of algorithms and harmful design features. The overall model was significant, F (8,1330) = 5.52, p < 0.01, R2 = 0.215. For a full summary of these results, please see Table 4. The first block containing demographic variables was significant. Parent gender, religion, income, and race positively and significantly related to support for age verification such that mothers, those higher in religiosity and income, and Hispanic and White parents (compared to non-Hispanic and non-White parents) showed higher levels of support. We then added in moral foundation saliences to the next step of the regression series. The overall model was significant, with care, fairness, and authority positively and significantly relating to support for allowing full parental access to adolescents’ accounts and loyalty and purity negatively and significantly relating. In the third step of the regression series, we added parenting style. The overall model was significant, with indulgent, authoritative, and authoritarian parenting styles positively and significantly relating to support, and neglectful negatively and significantly relating. Finally, we added in variables related to political identity. The overall model was not significant, showing that neither political party identification nor strength of political affiliation significantly relate to support for regulating the use of algorithms and platform design in order to limit harmful content.
Table 4
Hierarchical Linear Regression Results for Algorithms and Platform Design, Blocking Pornographic Content, and Blocking Foreign Company Apps Provisions
Algorithm and Platform Design | Pornographic Content | Foreign Company Apps | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
Independent Variables | Model 1 | Model 2 | Model 3 | Model 4 | Model 1 | Model 2 | Model 3 | Model 4 | Model 1 | Model 2 | Model 3 | Model 4 |
Demographics | ||||||||||||
Parent Gender | 0.21*** | 0.20*** | 0.23*** | |||||||||
Religion | 0.10** | 0.03 | 0.10*** | |||||||||
Income | 0.04*** | 0.02* | 0.04*** | |||||||||
Hispanic | 0.30** | 0.39*** | 0.27* | |||||||||
White | 0.30** | 0.33*** | 0.18 | |||||||||
Black | 0.10 | 0.20 | 0.22 | |||||||||
Child Gender | 0.01 | 0.05 | −0.02 | |||||||||
Child Age | 0.01 | 0.01 | 0.01 | |||||||||
MFT | ||||||||||||
Care | 0.11*** | 0.14*** | −0.04 | |||||||||
Fairness | 0.20*** | 0.11*** | 0.10** | |||||||||
Loyalty | −0.10** | −0.07** | −0.01 | |||||||||
Authority | 0.18*** | 0.20*** | 0.24*** | |||||||||
Purity | −0.15*** | −0.15*** | −0.01 | |||||||||
Parenting Style | ||||||||||||
Indulgent | 0.14*** | 0.18*** | 0.02 | |||||||||
Authoritative | 0.11** | 0.07 | 0.17*** | |||||||||
Authoritarian | 0.20** | 0.06 | 0.17*** | |||||||||
Neglectful | −0.11*** | −0.11*** | −0.04 | |||||||||
Political Identity | ||||||||||||
Political Identification | 0.02 | 0.01 | 0.01 | |||||||||
Strength of Affiliation | 0.00 | 0.06** | 0.08*** | |||||||||
R2 | 0.032 | 0.025 | 0.026 | |||||||||
∆R2 | 0.111 | 0.069 | 0.003 | 0.096 | 0.069 | 0.010 | 0.084 | 0.050 | 0.012 | |||
Limits to Pornography on Social Media
We ran the eighth regression to predict parents’ level of support toward blocking pornographic content from users under 18. The overall model was significant, F (8,1337) = 4.32, p < 0.01, R2 = 0.200. For a full summary of these results, please see Table 4. The first block containing demographic variables was significant. Parent gender, income, and race positively and significantly related to support such that mothers, those with higher income, and Hispanic and White parents showed higher levels of support. We then added in moral foundation saliences to the next step of the regression series. The overall model was significant, with care, fairness, and authority positively and significantly relating to support for blocking pornographic content and loyalty and purity negatively and significantly relating. In the third step of the regression series, we added parenting style. The overall model was significant, with indulgent parenting style positively and significantly relating to support, and neglectful negatively and significantly relating. Finally, we added in variables related to political identity. The overall model was significant. Strength of affiliation was positively and significantly related to support, such that more conservative parents showed higher levels of support for blocking pornographic content from users under 18.
Block Foreign Company Apps
We ran the final regression to predict parents’ level of support toward blocking apps owned by foreign companies in order to protect adolescent user data. The overall model was significant, F (8,1336) = 4.38, p < 0.01, R2 = 0.172. For a full summary of these results, please see Table 4. The first block containing demographic variables was significant. Parent gender, religion, income, and race positively and significantly related to such that mothers, those higher in religion and income, and Hispanic parents (compared to non-Hispanic parents) showed higher levels of support. We then added in moral foundation saliences to the next step of the regression series. The overall model was significant, with fairness and authority positively and significantly relating to support for blocking apps owned by foreign companies. In the third step of the regression series, we added parenting style. The overall model was significant, with authoritative and authoritarian parenting styles positively and significantly relating to support. Finally, we added in variables related to political identity. The overall model was significant. Strength of affiliation was positively and significantly related to support, such that more conservative parents showed higher levels of support for blocking apps owned by foreign companies.
Summary of Results
In response to RQ1, parent gender and income were the most consistent predictors of support across all provisions, such that mothers and parents with higher incomes showed greater levels of support across all nine provisions. Conversely, race (for Black when compared to Non-Black parents) only related to increased support for requiring parental consent to register for an account. Further, child gender was not significantly related to support for any provision. In response to RQ2, each of the five moral foundation saliences differentially related to level of support across the nine provisions. Care was positively related to support for five out of the nine total provisions, while fairness was positively related to support for eight of the nine total provisions. Conversely, loyalty was negatively related to five provisions. Authority salience was the most predictive of the moral foundations, relating positively to all nine of the provisions. Finally, purity salience was negatively related to six total provisions.
In response to RQ3, each of the four parenting styles differentially influenced support across the nine provisions. Neglectful parenting was the most predictive, being negatively related to support for seven of the nine total provisions. This was followed by indulgent parenting, which was positively related to six provisions total. Authoritative and authoritarian parenting styles were each positively related to support for five of the nine provisions. Interestingly, parenting style was not related to support for the provision regarding age verification. In response to RQ4, there was very little evidence that either aspect of political identity was related to support for the provisions. Political affiliation was not found to be significantly related to support for any of the provisions. However, strength of affiliation (i.e., how liberal or conservative an individual is) was positively and significantly related to four of the nine provisions.
Discussion
In a relatively short amount of time, the United States has gone from leaving child and adolescent social media use largely unregulated, to proposing or enacting related bills in over 20 states. However, for as varied and diverse as the country is, so too are the regulatory efforts. In contrast to more sweeping legislation, such as the Child Online Safety Act in the UK, the US has taken a more piecemeal approach. As such, the primary goal of this study was to parse through the legislation and begin to understand which approaches parents are in support of, and which they are not. To this end, we investigated four primary research questions: (1) RQ1: Which parent demographic variables (gender, race/ethnicity, religiosity, and income) and child demographic variables (age and gender) relate to parent support for each provision? (2) How do parents’ moral foundation saliences relate to support of the provisions? (3) How do different parenting styles relate to support of the provisions? and (4) How does parent political affiliation and strength of affiliation relate to support? Our findings reveal several key insights about the factors influencing parental support for social media regulation.
Theoretical and Practical Implications
Research question 1
Our analysis revealed that demographic factors played a significant but nuanced role in predicting support for social media regulation. Parent gender and income were the most consistent predictors of support across all provisions, such that mothers and those with higher incomes showed greater levels of support across all nine provisions. This finding may be explained by female parents typically carrying a heavier burden of childcare and wealthier parents having increased scheduling flexibility, both of which could lead to greater awareness of and investment in their child’s social media use (Zamarro & Prados, 2021; Zady & Portes, 2001). Notably, child demographics (gender and age) were rarely significant predictors of support, suggesting that support and concern are more parent driven rather than child specific.
Research question 2
The examination of moral foundation saliences revealed complex relationships with support for regulation. Authority salience was the most predictive of the moral foundations, relating positively to all nine of the provisions. Care salience was positively related to support for five provisions, while fairness salience showed positive relationships with eight provisions. Conversely, loyalty was negatively related to five provisions, and purity salience was negatively related to six provisions. These findings extend our understanding of Moral Foundations Theory (MFT) in a new context, demonstrating that parents’ moral foundations may shape their attitudes toward media regulation for older children and adolescents, similar to how they shape parents’ attitudes about younger children’s media use (Cingel et al., 2023b).
Research question 3
Analysis of parenting styles revealed distinct patterns in their relationships with support for regulation. Neglectful parenting was the most predictive, being negatively related to support for seven of the nine total provisions. Indulgent parenting was positively related to six provisions, while authoritative and authoritarian styles each showed positive relationships with five provisions. Interestingly, parenting style was not related to support for the provision regarding age verification. These patterns suggest that both the support and control dimensions of parenting styles influence attitudes toward regulation, as parents with either high support (indulgent) or high control (authoritarian/authoritative) showed similar levels of support for the provisions, while the absence of both dimensions (neglectful parenting) consistently predicted lower support. These findings align with previous research on parental media monitoring behaviors. For example, Padilla-Walker & Coyne (2010) found that authoritative parenting predicted higher levels of proactive media monitoring, with parental regulation specifically associated with increased restrictive mediation. This parallel between authoritative parenting and support for regulatory measures suggests a consistency between parents’ personal approaches to media monitoring and their attitudes toward broader regulation.
Research question 4
Perhaps most striking was our finding was the relative lack of significance of political identity in predicting support for social media regulation. Indeed, political party affiliation was not found to be significantly related to support for any of the provisions. Strength of political affiliation was significantly related to only four of the nine provisions. This relative absence of partisan division is particularly noteworthy given that partisanship has previously been deemed the single most dividing line in American public policy (PEW Research Center, 2019). The bipartisan nature of support for social media regulation suggests a broader societal consensus about the need to protect youth online.
Strengths, Limitations, and Future Directions
These findings hold several important theoretical and practical implications. In terms of theory, by identifying moral foundation salience as a consistent predictor of legislative support, we extend our understanding of both MFT and the extent to which moral sensibilities influence individuals’ attitudes and opinions. The consistent influence of authority-related variables (both in MFT and parenting styles) suggests that attitudes toward control and authority may be particularly important in shaping views on social media regulation. For policymakers, these findings indicate that legislation aimed at protecting youth on social media is not only desired by constituents but is also likely to receive widespread support. The high levels of support across provisions suggest policymakers could pursue more comprehensive regulation, rather than continuing the piecemeal approach. The bipartisan nature of support, combined with particularly strong backing for provisions related to algorithms, platform design, and content restrictions, suggests that more sweeping federal legislation might be both feasible and welcomed by parents. Additionally, social media companies should consider these findings as a signal to implement things such as stronger parental controls and age verification systems to address parent concerns. The high levels of support for parental controls, age verification systems, and algorithmic restrictions suggest that platforms could increase levels of public trust by voluntarily implementing these more desired features. This could include more robust age verification processes, enhanced parental oversight capabilities, algorithmic adjustments to limit harmful content exposure, and improved data privacy protections for young users.
While this study fills an important gap in the current literature, it is not without limitation. Our sample was primarily composed of White, upper-middle-class, educated parents. Particularly when examining the attitudes of Americans regarding current legislation, it is desirable to access as diverse a subset of the population as possible to more closely represent the voting public. Thus, an important next step would be to examine parental attitudes on a nationally representative scale. While the present study has illuminated parental attitudes toward the various provisions, there still exists a need to test the efficacy of the legislation as it is enacted. Future studies should therefore examine the level to which each of the aspects of regulation are actually effective in supporting child and adolescent mental health and well-being, specifically by collecting longitudinal data to examine directionality. Particularly in response to concerned parents and public, the ability to provide evidentiary support will be crucial.
Supplementary information
The online version contains supplementary material available at https://doi.org/10.1007/s10826-025-03071-6.
Compliance with Ethical Standards
Conflict of Interest
D. P. C. has been retained as an expert witness in U.S. social media litigation. The authors declare no other competing interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.