2.3.1 Traditional Screen Media

Video content remains the most common form of digital media for children 5 years and under, making up approximately 73% of viewing time, although it is increasingly viewed in streaming formats on multiple devices, from traditional family television sets, to tablets and smartphones (Reference Rideout and RobbRideout & Robb, 2020). Educational television programs designed explicitly for young children (e.g., Sesame Workshop and PBS Kids programming) have been associated with greater school readiness and better long-term educational trajectories, particularly for children who live in lower-income homes and children of color (e.g., Reference Anderson, Huston, Schmitt, Linebarger and WrightAnderson et al., 2001; Reference Linebarger, Barr, Lapierre and PiotrowskiLinebarger et al., 2014; Reference Wright, Huston and MurphyWright et al., 2001). Video content may also promote positive outcomes via prosocial behaviors that media characters model. Educational and prosocial content often models healthy emotion regulation strategies (e.g., Daniel Tiger’s Neighborhood; Reference Malti and DysMalti & Dys, 2018; Reference Rasmussen, Shafer and ColwellRasmussen et al., 2016). Exposure to well-designed children’s programming is associated with long-term positive outcomes for language and executive functioning (Reference Anderson, Huston, Schmitt, Linebarger and WrightAnderson et al., 2001; Reference Barr, Lauricella, Zack and CalvertBarr et al., 2010a; Reference Linebarger and WalkerLinebarger & Walker, 2005; Reference Linebarger and VaalaLinebarger & Vaala, 2010). Even in infancy, children are shown to be attentive to media content that matches their interests and is meaningful to them (Reference Nikken and ScholsNikken & Schols, 2015). Well-designed video content can also inspire high-quality parent–child interaction when parents watch together with their children (Reference Pempek, Demers, Hanson, Kirkorian and AndersonPempek et al., 2011).

Conversely, poorer cognitive and emotion regulation profiles in children have been associated with lower-quality content and exposure to adult-directed or background television (Reference Barr, Lauricella, Zack and CalvertBarr et al., 2010a; Reference Schmidt, Pempek, Kirkorian, Frankenfield and AndersonSchmidt et al., 2008; Reference Linebarger, Barr, Lapierre and PiotrowskiLinebarger et al., 2014). Prior research indicates longitudinal associations of longer daily television viewing duration and low-quality media content with toddler externalizing behavior (Reference Tomopoulos, Dreyer and ValdezTomopolous et al., 2007), preschooler behavioral problems (Reference Verlinden, Tiemeier and HudziakVerlinden et al., 2012), and lower early childhood psychosocial well-being (Reference Pagani, Fitzpatrick, Barnett and DubowPagani et al., 2010). Exposure to high levels of background TV, much of which may be adult-directed rather than child-centered and educational, has been associated with later emotional reactivity, externalizing behavior, and aggression in toddlers (Reference Chonchaiya, Sirachairat, Vijakkhana, Wilaisakditipakorn and PruksananondaChonchaiya et al., 2015) and lower executive function in preschoolers (Reference Barr, Lauricella, Zack and CalvertBarr et al., 2010a).

2.3.2 Interactive Touchscreen Media

Touchscreen interfaces on smartphones and tablets can also include interactivity that shapes the impact of content on children. Interactive apps and platforms provide contingent responsiveness to children’s input, allowing feedback and digital scaffolding in support of curricular goals. Interactive media are likely to support children’s learning when the interactive features focus on, rather than distract from, the lesson (Reference Furenes, Kucirkova and BusFurenes et al., 2021; Reference KirkorianKirkorian, 2018). Alternatively, digital interactivity in the form of extraneous audiovisual effects or advertising can distract from learning processes (Reference Bus, Neuman and RoskosBus et al., 2020).

Reference Hirsh-Pasek, Zosh and GolinkoffHirsh-Pasek and colleagues (2015) used Science of Learning principles, which are summarized in Table 1, to describe how effective educational programming can be designed in mobile app form. Educational approaches to interactive design emphasize children’s “minds-on” rather than automatic engagement, interactive features that support learning activities, meaningful curricula, and supporting social interactions around the app. In other words, the effects of interactive technology depend on the quality of content (e.g., Is the app based on an educational curriculum?) and aspects of interactive design (e.g., Do ads disrupt the play experience? Are the tasks repetitive versus “minds-on”?) (Reference Hirsh-Pasek, Zosh and GolinkoffHirsh-Pasek, et al., 2015). However, detailed coding of the top-rated or top-downloaded “educational” apps in the Google Play Store and Apple App Store demonstrated that these features occur only infrequently in commercial products (Reference Meyer, Zosh and McLarenMeyer et al., 2021). Moreover, many apps and child platforms that are marketed as educational are frequently monetized through advertisements (Reference Myers, Crawford, Murphy, Aka-Ezoua and FelixMeyer et al., 2018), data collection (Reference Zhao, Egelman and WeeksZhao et al., 2020), and nudges to make purchases (Reference Radesky, Hiniker and McLarenRadesky et al., 2022).

As with television and video content, the quality of interactive design may also shape parent–child interaction around media. The interactive design affordances and monetization of children’s digital products have the potential to influence how children learn from digital media and how digital media could displace or interrupt other daily activities or relational processes. For example, observational research demonstrated that parents had difficulty getting their young child’s attention (Reference Hiniker, Heung, Hong and KientzHiniker et al., 2018b) and social reciprocity decreased (Reference Munzer, Miller, Weeks, Kaciroti and RadeskyMunzer et al., 2019) when apps were fast-paced and had high levels of interactive design. In a laboratory-based experiment, toddlers showed more negativity toward their parents when prompted to transition away from a nursery rhyme app on a mobile device, compared to the same content in a print book (Reference Munzer, Miller, Wang, Kaciroti and RadeskyMunzer et al., 2021).

Interactive design features have unique implications for the types of digital media content children view. Reference Radesky, Schaller, Yeo, Weeks and RobbRadesky and colleagues (2020) tracked the mobile devices of preschool-aged children and found that the most commonly used media were mobile games and YouTube, which were used for an average of 1 hour per day. Such apps use multiple strategies to increase usage time and may expose infants and young children to more consumerist, age-inappropriate, and violent content (Reference Munzer, Torres and DomoffMunzer et al., 2022). The content of these channels and apps and their impact on children’s development is poorly understood. Design features that merit further study include engagement-prolonging design (e.g., autoplay, behavioral reinforcement with frequent rewards) as engagement-prolonging design may be more likely to displace other important developmental opportunities for young children and may have associations with greater child behavioral difficulties when transitioning off devices (Reference Munzer, Torres and DomoffMunzer et al., 2022). However, the vast array of interactive apps, immersive games, and platforms now popular with young children poses measurement challenges as interactive design analysis is time- and labor-intensive. In addition, now that video, audio, and game content is personalized for child users through automated filtering and recommender systems, researchers need to understand how algorithmic decision-making shapes the media experience of young children.

2.4.1 Parental Mediation

Parental mediation strategies have been studied quite extensively. Parental mediation guides how children learn the function of media, how to operate a device, and how to interpret the content (Reference Nikken and ScholsNikken & Schols, 2015). For mediation of television, three general categories have emerged: (1) restrictive time and content limits; (2) active parental discussion of content; and (3) co-viewing for education and entertainment (Reference Nikken and ScholsNikken & Schols, 2015; Reference Piotrowski, Barr and LinebargerPiotrowski, 2017; Reference Valkenburg, Krcmar, Peeters and MarseilleValkenburg et al., 1999). Parents typically use restrictive media practices to limit inappropriate content, whereas they are more likely to actively mediate educational content (Reference Nikken and ScholsNikken & Schols, 2015). Like screen time, structural factors also predict parental mediation (Reference Connell, Lauricella and WartellaConnell et al., 2015; Reference Wang, Lwin, Cayabyab, Hou and YouWang et al., 2023). For example, in a study conducted in the United States, younger parents and fathers were more likely to co-engage with video games and mobile technology than older parents and mothers, and Latine/x parents were more likely to co-engage with tablet computers compared to White parents (Reference Connell, Lauricella and WartellaConnell et al., 2015). Device type also matters. For instance, parents of children 8 years and younger were more likely to report co-engagement with television and books than with smartphones and tablet computers (Reference Connell, Lauricella and WartellaConnell et al., 2015). Parental attitudes about media also influence mediation practices (Reference Nikken and ScholsNikkens & Schols, 2015; Reference Piotrowski, Barr and LinebargerPiotrowski, 2017; Reference Valkenburg, Krcmar, Peeters and MarseilleValkenburg et al., 1999; Reference Wang, Lwin, Cayabyab, Hou and YouWang et al., 2023). For example, a meta-analysis of predictors of parental mediation showed that parents who had more positive and more negative attitudes toward media and had higher levels of parental involvement were more likely to engage in active mediation practices compared to parents without strong positive or negative attitudes toward media (Reference Wang, Lwin, Cayabyab, Hou and YouWang et al., 2023). The biggest predictors of restrictive media practices were child age and parent media attitudes, with more restrictive practices used for younger children and parents with more negative attitudes.

For very young children, active mediation frequently takes the form of joint media engagement (JME), defined as the extent to which parents actively engage with their children during media use (e.g., playing games together, talking about the content). JME has been examined during television viewing, e-book reading, and app use (Reference Dore, Hassinger-Das and BrezackDore et al., 2018; Reference Fidler, Zack and BarrFidler et al., 2010; Reference Lauricella, Barr and CalvertLauricella et al., 2009, Reference Lauricella, Calvert and Barr2014; Reference Zack and BarrZack & Barr, 2016). Learning from media during early childhood is limited by children’s cognitive and memory constraints (e.g., Reference BarrBarr, 2013). However, high-quality JME can help overcome such constraints to increase young children’s learning. For example, JME significantly increased 15-month-olds’ transfer of learning from a touchscreen to the physical world (Reference Zack and BarrZack & Barr, 2016). Similarly, Reference Heimann, Hedendahl and OttmerHeimann and colleagues (2021) showed that JME also enhanced imitation learning in 2-year-olds.

JME also enhances adult–child conversations, which create new opportunities for early word learning. Reference Lavigne, Hanson and AndersonLavigne and colleagues (2015) found that parents included more new words per utterance during co-viewing, and this extra vocabulary carried over after the television program ended, resulting in richer parental language during a subsequent free play episode. The authors speculated that high-quality educational content could provide parents with topics that might be of interest or model language that parents could use with their young children. Such opportunities may explain why JME appears to buffer the negative effects of early media use on children’s language and literacy development (Reference Dore, Logan, Lin, Purtell and JusticeDore et al., 2020). They may also explain why JME combined with the use of educational content is associated with greater vocabulary, whereas solitary media use is associated with mostly neutral or negative language and cognitive outcomes (Reference Madigan, McArthur, Anhorn, Eirich and ChristakisMadigan et al., 2020; Reference Sundqvist, Koch, Birberg Thornberg, Barr and HeimannSundqvist et al., 2021). For example, in a study of low-income Latine/x mothers and their children, Reference Mendelsohn, Brockmeyer and DreyerMendelsohn and colleagues (2010) found that the degree to which parents reported actively discussing educational media content when their children were 6 months old predicted their children’s language outcomes at age 14 months. Specifically, verbal interactions during educational content at age 6 months were associated with better language outcomes at age 14 months. Similarly, Reference Rasmussen, Shafer and ColwellRasmussen and colleagues (2016) found that the effectiveness of a prosocial program aimed at 2- to 4-year-olds was maximized by parent reports of consistent active mediation practices in the home.

The benefits of JME for language and literacy extend to other forms of digital media use, including e-books and video chat (Reference Krcmar and CingelKrcmar & Cingel, 2014; Reference Lauricella, Calvert and BarrLauricella et al., 2014; Reference Strouse and GaneaStrouse & Ganea, 2016, Reference Strouse and Ganea2017). For example, in an experimental study of JME during video chat, children learned more words when their caregiver engaged in JME with the child than when their caregiver did not engage in JME (Reference Myers, Crawford, Murphy, Aka-Ezoua and FelixMyers et al., 2018). In some cases, carefully designed digital media may deliver some of the learning enhancements provided by JME. For example, Reference Dore, Hassinger-Das and BrezackDore et al. (2018) found that children comprehend some content from e-books using audio narration, indicating that using e-books independently may be a worthwhile activity for preliterate children while caregivers are otherwise occupied. However, results also show that children recall the most information about the e-book after reading with a parent.

Other research suggests there may be individual child differences in the benefits of JME. For example, JME disproportionately enhanced learning for children with lower vocabulary and executive functioning scores, probably because adults tailored the information to meet the needs of individual children (Reference Strouse and GaneaStrouse & Ganea, 2016). Overall, JME is likely to be highly effective and protective because learning from media is cognitively challenging and requires that children learn the features of different media types and the affordances of each device. In the face of such challenges, JME is currently the most effective way to tailor learning from media during early childhood (Reference Barr and BornsteinBarr, 2019b).

The majority of JME research focuses on parent–child media co-use. One factor that is almost entirely overlooked in this research is the role of siblings. Older siblings are likely to influence screen time and content choice. Moreover, JME patterns are likely to differ for family media use with siblings of different ages versus one-on-one parent–child interactions. Similarly, the literature rarely considers the influence of more than one caregiver, such as the degree to which parents in the same family system adopt similar media rules and practices. Going forward, media effects research could be expanded to better reflect the entire family media ecology.

2.4.2 Technoference

Technology use and the presence of technology can distract, interrupt, and intrude on parenting and parent–child interactions, which is termed technoference (e.g., Reference McDaniel and CoyneMcDaniel & Coyne, 2016; Reference McDaniel and RadeskyMcDaniel & Radesky, 2018a; Reference McDanielMcDaniel, 2020). Technoference is common across a variety of child activities (e.g., feedings/mealtimes, indoor and outdoor play). Early research on technoference focused on the impact of background television, demonstrating a decrease in the quality and quantity of parent and child play and social interactions in the presence versus absence of background television (Reference Pempek, Kirkorian and van den BulckPempek & Kirkorian, 2020). For example, in experimental research with parents and their 12- to 36-month-old children, background television decreased parents’ and children’s interactions, parents’ responsiveness to their children’s bids for attention, and the degree to which parents actively engaged in their children’s toy play (Reference Kirkorian, Pempek, Murphy, Schmidt and AndersonKirkorian et al., 2009).

More recently, research has expanded to examine technoference in the context of parents’ smartphone use. During parental smartphone use, examples of associated technoference include decreases in appropriate and timely responses to children’s bids for attention (Reference Abels, Vanden Abeele, Van Telgen, Van Meijl, Luef and MarinAbels et al., 2018; Reference Domoff, Borgen and RadeskyDomoff et al., 2020; Reference Elias, Lemish, Dalyot and FloegelElias et al., 2021; Reference Ewin, Reupert, McLean and EwinEwin et al., 2021; Reference Hiniker, Sobel and SuhHiniker et al., 2015; Reference Kiefner-Burmeister, Domoff and RadeskyKiefner-Burmeister et al., 2020, Reference Vanden Abeele, Abels and HendricksonVanden Abeele et al., 2020), reduced awareness of dangerous situations or injuries (Reference Elias, Lemish, Dalyot and FloegelElias et al., 2021), reductions in joint play, conversations, or engagement (Reference Ewin, Reupert, McLean and EwinEwin et al., 2021; Reference Konrad, Hillmann and RisplerKonrad et al., 2021b; Reference Krapf-Bar, Davidovitch, Rozenblatt-Perkal and Gueron-SelaKrapf-Bar et al., 2022; Reference Ochoa, Reich and FarkasOchoa et al., 2021; Reference Linder, McDaniel and JaffeLinder et al., 2022), decreased sensitivity/warmth (Reference Konrad, Hillmann and RisplerKonrad et al., 2021b; Reference Radesky, Kistin and ZuckermanRadesky et al., 2014a; Reference Ventura, Sheeper and LevyVentura et al., 2019), increased child behavior problems (Reference Linder, McDaniel and JaffeLinder et al., 2022; Reference McDaniel and RadeskyMcDaniel & Radesky, 2018b; Reference Radesky, Kistin and ZuckermanRadesky et al., 2014a; Reference Vanden Abeele, Abels and HendricksonVanden Abeele et al., 2020), and poorer language learning (Reference Reed, Hirsh-Pasek and GolinkoffReed et al., 2017). For example, in observations of caregivers and young children in fast food restaurants, many parents showed high attentional absorption with smartphones, which co-occurred with less parent–child conversation and more child behavior escalation (Reference Radesky, Kistin and ZuckermanRadesky et al., 2014a). In another study, when mothers used smartphones around their children, the mothers exhibited less frequent verbal and nonverbal interactions with their young children during meals (Reference Radesky, Miller and RosenblumRadesky et al., 2015) and, when interviewed, displayed diminished capacity for mentalization, or perspective-taking, about their child (Reference Radesky, Leung and AppuglieseRadesky et al., 2018b). In addition, parent phone use during mealtime interactions is associated with lower use of positive feeding practices like responsiveness to child hunger and fullness cues, less frequent modeling of healthy eating behaviors, and greater use of less desirable feeding practices such as use of food to regulate children’s emotions or pressuring children to eat (Reference Robinson, Domoff, Kasper, Peterson and MillerRobinson et al., 2022; Reference Ventura and TeitelbaumVentura & Teitelbaum 2017; Reference Ventura, Sheeper and LevyVentura et al., 2019; Reference Vik, Grasaas and PolspoelVik et al., 2021). Associations between technoference and child characteristics may be bidirectional: In a longitudinal study of parent technoference, child externalizing behavior predicted parent technoference just as strongly as parent technoference predicted increases in child externalizing behavior (Reference McDaniel and RadeskyMcDaniel & Radesky, 2018a).

A number of relational processes underlie the reciprocal association between technoference and child behavior. As noted earlier, one likely mechanism is disruption in parents’ emotional availability and their appropriate and timely responses to children’s bids for attention (e.g., Reference Domoff, Borgen and RadeskyDomoff et al., 2020; Reference Hiniker, Sobel and SuhHiniker et al., 2015; Reference Kirkorian, Pempek, Murphy, Schmidt and AndersonKirkorian et al., 2009). During observational studies researchers noted that parents often portrayed a flat affect when absorbed in mobile device content. This flat expressionless face resembles a classic still-face (Reference Konrad, Berger-Hanke, Hassel and BarrKonrad et al., 2021a, b; Reference Krapf-Bar, Davidovitch, Rozenblatt-Perkal and Gueron-SelaKrapf-Bar et al., 2022; Reference Myruski, Gulyayeva and BirkMyruski et al., 2018). Experimental studies have shown that infants find such expressionless flat affect aversive and often attempt to regain an adult’s attention via increased vocalization (Reference Goldstein, Schwade and BornsteinGoldstein et al., 2009). Eventually, infants may learn that the use of the smartphone is a cue that their mothers are unavailable and wait to reengage only when the smartphone is no longer in use (Reference Konrad, Hillmann and RisplerKonrad et al., 2021b). Researchers have posited that withdrawal of positive affect during phone use might be a potential explanatory mechanism for the negative outcomes associated with technoference. To test this hypothesis, researchers designed experimental studies. For example, Reference Krapf-Bar, Davidovitch, Rozenblatt-Perkal and Gueron-SelaKrapf-Bar et al. (2022) conducted a study of the relation between mobile phone disruptions and joint activity between mothers and infants at 12 months. Mobile phone disruptions due to texting were more likely to disrupt the formation and duration of joint attention and joint activity between mothers and their 1-year-olds than a social disruption caused by an experimenter talking to the mother. Another study compared differences between smartphone text disruptions and a non-digital disruption via a paper-and-pencil survey (Reference Konrad, Hillmann and RisplerKonrad et al., 2021b). In both interruption scenarios, maternal responsiveness and infant positive affect decreased, suggesting that it is not only technology disruptions that interrupt social exchanges. Higher levels of absorption during the task were related to lower levels of responsiveness to both digital and nondigital distractions. However, while both technological and analog behaviors may disrupt interactions, smartphone use may be more frequent and disruptive due to the frequency of notifications and ubiquity of parental smartphone use, leading to larger cumulative effects over time (Reference McDanielMcDaniel, 2021).

It has not been established whether the effects of technoference on parent–child interactions differ as a function of the content and context of parental phone use. Activities such as parental social media use or checking and responding to work emails may result in higher levels of absorption, whereas brief glances at notifications may not interfere with ongoing parent–child interactions. Furthermore, individual differences in how parents split their attention between their children and the digital screen are likely to influence the degree to which the phone use and background television are experienced as technoference (Reference Kirkorian, Choi and AndersonKirkorian et al., 2019; Reference Konrad, Berger-Hanke, Hassel and BarrKonrad et al., 2021a). Beyond the amount of overall household media use, parent’s own media use patterns and their motivations and responses to media are often overlooked. Millennial and Gen Z parents have grown up with access to a wide array of media resources. This technology use may also support parents through mechanisms such as stress relief or access to support or resources (Reference Radesky, Kistin and EisenbergRadesky et al., 2016c; Reference Torres, Radesky, Levitt and McDanielTorres et al., 2021). Specifically, smartphone use during breastfeeding is common, and although parents report they felt guilty, they also report benefits such as distraction, connection with others, or staying awake at night which is associated with persistence in breastfeeding (Reference Coyne, Shawcroft and GaleCoyne et al., 2022a). Additional research is needed to track the long-term effects of technoference on children.

2.4.3 Motivations for Media Use

Family digital media use does not occur at random. Instead, it is often motivated by parents’ needs and goals for themselves and their young children. Yet, research often overlooks the reasons that parents use technology, and parents are seldom asked why they use media (Reference NikkenNikken, 2019; Reference Nikken and ScholsNikken & Schols, 2015). When asked, parents of young children report using media to meet a wide range of needs (Reference NikkenNikken, 2019). Commonly reported reasons for child media use include keeping the child busy or entertained when the parent is occupied, taking a break from childrearing, and using media to regulate children’s behavior or emotions (Reference NikkenNikken, 2019). Parents’ motivations for media use may shape children’s media exposure. For example, Reference Nikken and ScholsNikken and Schols (2015) reported that children spend significantly more time using mobile devices and have more electronic screens in their bedrooms when their parents endorse the statement that digital media provide a moment of rest.

The fact that digital media serve so many functions might explain why children’s digital media use is predicted by parenting stress (Reference Madigan, McArthur, Anhorn, Eirich and ChristakisMadigan et al., 2020), maternal depression (Reference Bank, Barr and CalvertBank et al., 2012; Reference Coyne, Shawcroft and GaleCoyne et al., 2021; Reference Pempek and McDanielPempek & McDaniel, 2016), poor maternal relational well-being (Reference Tombeau Cost, Korczak and CharachTombeau Cost et al., 2020), and low child self-regulation (Reference Coyne, Shawcroft and GaleCoyne et al., 2021; Reference McDanielMcDaniel & Radesky, 2020; Reference Nabi and KrcmarNabi & Krcmar, 2016). The many functions of media use may also partially explain why media use is higher among lower-resourced and minority families (Reference Rideout and RobbRideout & Robb, 2020), and why media use was higher during the COVID-19 pandemic than before the pandemic (Reference Hartshorne, Huang and Lucio ParedesHartshorne et al., 2021). Indeed, some scholars have called for reframing child media use as an indicator of parental distress (Reference Hartshorne, Huang and Lucio ParedesHartshorne et al., 2021). The relation between parents’ mental health and family media use is not straightforward. Researchers have found some motivations for media use to be constructive in meeting parents’ needs in supportive ways (e.g., bonding over shared media, connecting with far-away family). For instance, depressed or highly stressed mothers choose more child-directed, educational media than their non-depressed or less-stressed counterparts (Reference Bank, Barr and CalvertBank et al., 2012; Reference Pempek and McDanielPempek & McDaniel, 2016), suggesting that depressed or stressed mothers may use media to provide cognitive stimulation for their children. More information about the types of media and media use practices associated with contextual stressors is needed.

Other investigators have found that some types of media use may displace high-quality parent–child interactions (Reference Kirkorian, Pempek, Murphy, Schmidt and AndersonKirkorian et al., 2009, Reference Kirkorian, Choi and Anderson2019; Reference Radesky, Miller and RosenblumRadesky et al., 2015). Some parental motivations for media use (e.g., to enable parents’ escape from stress or parenting, to regulate children’s emotions or behavior) may be problematic for early socioemotional development (Reference Coyne, Shawcroft and GaleCoyne et al., 2021; Reference McDaniel and RadeskyMcDaniel & Radesky, 2018). Using media to calm young children may be particularly problematic for the development of self-regulation by removing opportunities for children to learn and practice self-regulatory skills (Reference Coyne, Shawcroft and GaleCoyne et al., 2021; Reference McDaniel and RadeskyMcDaniel & Radesky, 2018). Such regulatory uses of mobile devices to calm children have been associated cross-sectionally with social-emotional delays (Reference Radesky, Peacock-Chambers, Zuckerman and SilversteinRadesky et al., 2016d) and worse emotional knowledge and reactivity (Reference Coyne, Reschke and StockdaleCoyne et al., 2023b). Longitudinal findings are fewer and provide mixed results. Some studies have shown that regulatory media use predicts negative outcomes of higher emotional reactivity (Reference Gordon-Hacker and Gueron-SelaGordon-Hacker & Gueron-Sela, 2020). In contrast, other studies have demonstrated that regulatory media use predicts positive outcomes of higher levels of empathy in preschoolers (Reference Coyne, Reschke and StockdaleCoyne et al., 2023b).

The reasons parents use digital media for and around their young children may shape the content and context of media use. For example, using digital media to escape or regulate might lead to more solitary media use, less joint media engagement, and selection of streaming media or apps with exploitative features that keep children engaged for prolonged periods of time (Reference Radesky, Hiniker and McLarenRadesky et al., 2022; Reference Radesky and HinikerRadesky & Hiniker, 2022). The extent to which digital media use disrupts positive parent–child interactions may explain why children’s self-regulation has been linked to some types of media use but not others. For example, TV viewing during infancy and toddlerhood, especially viewing adult-directed or violent content, is associated with children’s worse emotion and attention regulation (Reference Barr, Lauricella, Zack and CalvertBarr et al., 2010a; Reference Gueron-Sela and Gordon-HackerGueron-Sela & Gordon-Hacker, 2020; Reference McDaniel and RadeskyMcDaniel & Radesky, 2018; Reference Zimmerman and ChristakisZimmerman & Christakis, 2007).

Finally, the literature points to bidirectional relations between parents’ mental health, parent and child digital media use, and child behavior. That is, more challenging child temperament (i.e., negative affect, low effortful control) predicts increased parental stress, in turn predicting increased digital media use for parents and children (Reference Coyne, Shawcroft and GaleCoyne et al., 2021; Reference McDaniel and RadeskyMcDaniel & Radesky, 2018, Reference McDaniel2020; Reference Shin, Choi, Resor and SmithShin et al., 2021). The state of the research underscores the importance of the early caregiving environment for the development of self-regulation and that media may be one feasible target for intervention. There are limitations in the extant literature such that there is little research predicting positive media motivations (e.g., to bond and relax together). Researchers have also examined how motivations for child and family media use and the type of media content chosen may fluctuate day-to-day, or even hour-to-hour (Reference Coyne, Shawcroft and GaleCoyne et al., 2021). These decisions may depend on the stress and well-being of the parent and the behavior and temperament of the child. However, these moment-to-moment decisions have received little empirical attention. Commonly used parent-report measures, which ask parents about their typical behaviors over two or more weeks, may lack precision. Collectively, this research suggests a link between parents’ own needs and their use of digital media with and around children, but much remains to be learned about the short-term dynamic processes that shape media decisions for children in the moment and how such decisions may accumulate over time to affect longer-term child outcomes.

2.4.4 Structural Factors

Structural factors, such as income inequality, institutional racism (e.g., in housing or educational opportunity), and disinvestment in marginalized communities, shape multiple child development outcomes through the opportunities available to families. Parents who use more digital media often come from low-SES backgrounds (Reference Lauricella, Wartella and RideoutLauricella et al., 2015; Reference Rideout and RobbRideout & Robb, 2020), suggesting there may be functional uses of media for families with fewer resources compared with middle- or upper-SES families. Clear evidence is emerging of disparities in access to quality content and/or stable internet connections by SES (socioeconomic status) (Reference Katz, Moran and OgnyanovaKatz et al., 2019, Reference Katz, Jordan and Ognyanova2021; Reference Ramsetty and AdamsRamsetty et al., 2020; Reference Sen and TuckerSen & Tucker, 2020). For example, children of parents with a high school degree had significantly higher mobile device use (155.2 minutes/day) compared to parents with an advanced degree (82.3 minutes/day), were 8 times more likely to use YouTube (Reference Radesky, Schaller, Yeo, Weeks and RobbRadesky et al., 2020c), and had a two- to threefold higher rate of using apps that illegally collect private data (e.g., device identifiers that marketers use for behavioral advertising, Reference Zhao, Egelman and WeeksZhao et al., 2020). Communities of color are more likely to be targets of marketing of unhealthy products and financially exploitative services (Reference Radesky, Reid Chassiakos, Ameenuddin and NavsariaRadesky et al., 2020b), which serve to maintain inequities in access to quality input. SES and related psychosocial constructs are often examined as confounders or moderators in child media research, but research also needs to consider the ways that structural inequality is interwoven with family opportunity, media ecology, and the digital ecosystem itself.

2.5.1 Sleep

Systematic reviews provide consistent evidence that exposure to screen-based media in infancy, toddlerhood, and early childhood is negatively associated with sleep duration (Reference Lund, Sølvhøj, Danielsen and AndersenLund et al., 2021; Reference Moorman, Morgan and AdamsMoorman et al., 2019; Reference Zhang, Sousa-Sá and PereiraZhang et al., 2021). Reference Mallawaarachchi, Anglim, Hooley and HorwoodMallawaarachchi et al.’s (2022) meta-analysis reported that poorer sleep outcomes were associated with everyday use of a tablet or smartphone (see also Reference Cheung, Bedford, Saez De Urabain, Karmiloff-Smith and SmithCheung et al., 2017; Reference Chindamo, Buja and DeBattistiChindamo et al., 2019). Some studies have shown that higher levels of screen exposure were associated with prolonged sleep onset latency (Reference Bellagamba, Presaghi and Di MarcoBellagamba et al., 2021; Reference Cheung, Bedford, Saez De Urabain, Karmiloff-Smith and SmithCheung et al., 2017; Reference Xu, Wen, Hardy and RisselXu et al., 2016). In a longitudinal study that controlled for several demographic variables, Reference Benita, Gordon-Hacker and Gueron-SelaBenita and colleagues (2020) found that parental use of media to calm 22-month-old infants predicted longer latency to fall asleep at 26 months and that more media exposure at 22 months was associated with less nighttime sleep at 26 months. Findings on sleep fragmentation (number of times a child wakes during the night) and clinically meaningful sleep problems are mixed (Reference Bellagamba, Presaghi and Di MarcoBellagamba et al., 2021; Reference Cheung, Bedford, Saez De Urabain, Karmiloff-Smith and SmithCheung et al., 2017; Reference Xu, Wen, Hardy and RisselXu et al., 2016; Reference Zhang, Sousa-Sá and PereiraZhang et al., 2019).

Reference Bellagamba, Presaghi and Di MarcoBellagamba and colleagues (2021) provided evidence for the importance of digital media context when examining associations between digital media use and child sleep. These researchers collected data using a media questionnaire and the Brief Screening Questionnaire for Infant Sleep Problems (BISQ; Reference SadehSadeh, 2004) from parents of children under 3 years of age. The BISQ measures the estimated amount of time an infant sleeps during the night and naps during the day, the number of times an infant wakes during the night, and sleep problems. Media use was associated with less mature sleep patterns. Critically, these associations differed based on several contextual factors. For example, when there were more devices available in the household, young children went to bed later and had less nighttime sleep and more daytime sleep. Parents’ motivations for digital media use also mattered. When parents used media to occupy the child and when they reported higher overall levels of media use, children also had later bedtimes and less nighttime sleep.

Given rapid changes in sleep patterns during early childhood, findings also vary as a function of the age of the child as well as the device type. Researchers have posited that the timing of exposure is important as media before bedtime might increase behavioral arousal or displace parent–child interaction during sleep routines, both of which might disrupt sleep onset and quality (Reference Cheung, Bedford, Saez De Urabain, Karmiloff-Smith and SmithCheung et al., 2017; Reference Mallawaarachchi, Anglim, Hooley and HorwoodMallawaarachchi et al., 2022; Reference Lund, Sølvhøj, Danielsen and AndersenLund et al., 2021). Longitudinal and experimental research that considers the content and the context (e.g., timing of media exposure) is needed to disentangle these potential processes. Advanced techniques, including actigraphy and passive sensing, should be used to detect how sleep is disrupted and how sleep patterns change as a function of child age. Such methods could detect whether sleep onset is disrupted by different forms of content, or whether there are changes in nighttime wakings due to notifications or media use during the night. Interventions would then be tailored based on these findings.

2.5.2 Language

Viewing low-quality TV content is associated with language delays in young children, whereas viewing high-quality educational programs is protective and associated with higher vocabulary (Reference Çelen Yoldas and ÖzmertCelenn Yoldas & Ozmert, 2021; Reference Jing, Ye, Kirkorian and MaresJing et al., 2023; Reference Madigan, McArthur, Anhorn, Eirich and ChristakisMadigan et al., 2020). High-quality educational content can promote language development through children’s word-learning and parents’ infant-directed speech. For example, Reference Linebarger and WalkerLinebarger and Walker (2005) analyzed whether televised content was associated with vocabulary trajectories between 6 and 30 months of age, collecting data at 3-month intervals. Educational content, interactive narratives, and content with familiar characters were associated with vocabulary growth. This finding may be due to word learning from media designed to teach vocabulary, as evidenced by a meta-analysis (Reference Jing, Ye, Kirkorian and MaresJing et al., 2023). Furthermore, unlike adult-directed television (Reference Pempek, Kirkorian and AndersonPempek et al., 2014), infant-directed videos specifically designed to model positive parent–infant interactions succeed in increasing parent–child interaction quality (Reference Pempek, Demers, Hanson, Kirkorian and AndersonPempek et al., 2011) and the lexical diversity of parents’ infant-directed speech (Reference Lavigne, Hanson and AndersonLavigne et al., 2015).

Conversely, media exposure affects the language environment and often reduces child–adult interaction because it does not facilitate socially contingent conversational turns with a language partner whose responses are immediate, reliable, and accurate in content and relevance (Reference Anderson, Hanson, Barr and LinebargerAnderson & Hanson, 2017). As described earlier, family media use can create digital distraction that affects the amount and quality of family interactions. Such technoference has implications for young children’s language development. For example, brief interruptions to parent–child interactions via a phone call disrupted word learning in toddlers (Reference Reed, Hirsh-Pasek and GolinkoffReed et al., 2017). JME has the potential to foster greater parent–child interaction. However, while some parents provide descriptions and even pause videos to discuss content with some success, such language-rich interactions tend to occur less often than during face-to-face interactions (Reference Strouse, O’Doherty and TrosethStrouse et al., 2013). Without the support of an active mediating adult, media alone cannot scaffold the content to meet the specific needs of very young children.

Researchers hypothesize that solitary use of media by the parent or the child might be negatively associated with the child’s language development (Reference Dore, Logan, Lin, Purtell and JusticeDore et al., 2020). However, if parents engage in JME and discuss the media content with the child either during media use or in follow-up activities, then such experiences might enhance language outcomes. For example, parents of 2-year-olds completed a media questionnaire and a child vocabulary measure (Reference Sundqvist, Koch, Birberg Thornberg, Barr and HeimannSundqvist et al., 2021). Researchers also provided audio recording devices to parents who recorded the auditory environment of the child. The researchers found that child vocabulary decreased as a function of parent-reported screen time for the child, and child vocabulary was lower for parents who indicated they used digital media during child routines (Reference Sundqvist, Koch, Birberg Thornberg, Barr and HeimannSundqvist et al., 2021). On the other hand, positive linguistic parental strategies, such as interactional turn-taking with the child, JME, shared book reading, and use of mental state language were positively associated with children’s vocabulary development (Reference Sundqvist, Koch, Birberg Thornberg, Barr and HeimannSundqvist et al., 2021). JME via video chat provides an even more enriched context for early word learning (Reference Myers, Crawford, Murphy, Aka-Ezoua and FelixMyers et al., 2018). In sum, the quality of parent–child verbal interactions surrounding digital media use is the key predictor of child language outcomes.

2.5.3 Executive Function and Attention

Longitudinal studies have revealed complex patterns among digital media use, attention, and executive function outcomes. In a longitudinal study, higher levels of household media usage including maternal mobile usage and background televisions at 18 months predicted worse infant attention at 22 months (Reference Gueron-Sela and Gordon-HackerGueron-Sela & Gordon-Hacker, 2020). In this study, in addition to a traditional measure of child screen time, the investigators collected measures of the family media ecology, including parental use of digital media to regulate child behavior, use of background television, and maternal mobile usage. They analyzed their data as a function of each media type and created a cumulative measure which was a proportion of the maximum value of each media type. Concurrent measures of focused attention were not associated with cumulative family media use. However, higher cumulative family media use at 18 months was associated with poorer focused attention at 22 months and at 26 months. Similarly, in a short-term longitudinal study of 2- to 3-year-olds, which controlled for a range of covariates including verbal ability, Reference McHarg, Ribner, Devine and HughesMcHarg and colleagues (2020) found no concurrent association between screen time and executive function; however, screen time at age 2 negatively predicted child executive function at age 3.

There is some evidence that media content may moderate associations between digital media use and cognitive development. For instance, exposure to adult-directed (but not child-directed) television at age 1 year negatively predicted executive function at age 4 years (Reference Barr, Lauricella, Zack and CalvertBarr et al., 2010a). Similarly, viewing violent or non-violent entertainment television (but not educational television) before age 3 years predicted attention problems at age 7 years (Reference Zimmerman and ChristakisZimmerman & Christakis, 2007). Together, these findings suggest that negative associations between early television viewing and later attention and executive function depend on digital media content, with child-directed and educational programs buffering the negative associations seen with adult-directed or non-educational content. However, these studies did not include other contextual and parenting factors that might mediate these associations. In addition, it is hard to interpret findings based on global parent reports of screen time alone because, as mentioned earlier, parents have wide-ranging motivations for using digital media with and around their young children. For example, associations between children’s digital media use and their attention development may be bidirectional, with parents using digital media as a tool to regulate children with worse attention and inhibitory control skills (Reference Cliff, Howard, Radesky, McNeill and VellaCliff et al., 2018).

To the extent that digital media use disrupts executive function and attention, digital media use may have broader implications for cognitive development in general. For examples, Reference Lin, Cherng, Chen, Chen and YangLin et al. (2015) found a higher frequency of cognitive delay in a high-television exposure group compared to a matched low-television exposure group in 2-year-olds; Reference Tomopoulos, Dreyer and BerkuleTomopoulos et al. (2010) found a longitudinal negative relation between television exposure at 6 months and cognitive development 1 year later; and Reference Zhang, Adamo and OgdenZhang et al. (2022) found a difference in working memory capacity, with higher performance among 3- to 5-year-olds who adhered to screen time recommendations performing better than 3- to 5-year-olds who did not.

2.5.4 Social Competence

Socioemotional skills, including a child’s ability to read and respond to social cues, recognize and regulate emotional states, and adapt to external challenges, have also been linked to early digital media use. Longitudinal studies have demonstrated that early television viewing, and especially viewing adult-directed or violent content, is associated with worse emotion and attention regulation (Reference Gueron-Sela and Gordon-HackerGueron-Sela & Gordon-Hacker, 2020; Reference Hinkley, Verbestel and AhrensHinkley et al., 2014; Reference McDaniel and RadeskyMcDaniel & Radesky, 2018; Reference Zimmerman and ChristakisZimmerman & Christakis, 2007). For example, in a longitudinal study of 2- to 6-year-olds that controlled baseline levels of emotional and social behaviors, early digital media use was associated with later emotional problems but not social problems (Reference Hinkley, Verbestel and AhrensHinkley et al., 2014). As with other outcomes, disruptions to routines due to digital media use were associated with poorer social emotional outcomes (Reference Raman, Guerrero-Duby and McCulloughRaman et al., 2017). Processes are not well understood but have been attributed to a range of factors, including displacement of high-quality social interactions as described elsewhere.

Parents’ use of media to escape from stress and parenting or to regulate children’s emotions and behavior may be particularly disruptive to infant socioemotional development (Reference Coyne, Shawcroft and GaleCoyne et al., 2021; Reference McDaniel and RadeskyMcDaniel & Radesky, 2018). Moreover, there is evidence for bidirectional effects, with negative child affect and poor self-regulation predicting media use several months later for both children (Reference Coyne, Shawcroft and GaleCoyne et al., 2021; Reference McDanielMcDaniel & Radesky, 2020; Reference Neville, McArthur, Eirich, Lakes and MadiganNeville et al., 2021) and parents (Reference McDaniel and RadeskyMcDaniel & Radesky, 2018). Reference Neville, McArthur, Eirich, Lakes and MadiganNeville and colleagues (2021) collected data at ages 3, 5, 7, and 9 years and recorded digital media quality and behavioral problems at each time point. They found bidirectional associations between screen time and internalizing problems. However, children’s initial levels of externalizing behaviors were associated with later increases in screen time, but not vice versa. Reference McDanielMcDaniel and Radesky (2020) also observed a pattern of bidirectional associations between media quantity and externalizing behaviors in 1- to 5-year-olds. Cascading effects are shown in longitudinal associations between digital media use, child behavior, and parenting stress (Reference McDaniel and RadeskyMcDaniel & Radesky, 2018, Reference McDaniel2020). That is, more media use may increase behavioral problems, but media may also be used to reduce parenting stress or calm a child’s difficult behavior resulting in a cascade of increasing screen use and negative child behaviors.

In summary, multiple facets of family media ecology have been studied in isolation. Proposed processes underlying these patterns of results include displacement of uninterrupted sleep and play and changes in parent–child interactional quality, among others. Research in some areas has clearly illustrated the importance of digital media content as a moderator of associations between digital media use and child outcomes. Other areas have begun to focus on contextual factors that could moderate effects. Yet few studies have examined both content and context, and many studies are limited to simple measures of child screen time that do not capture variations in content and context or indirect effects resulting from digital media use by others in the family system. Despite near-universal access and adoption of digital media use during a child’s early years (Reference Rideout and RobbRideout & Robb, 2020), current theories do not often consider how these facets are interrelated and contribute to the developing child growing up in a digital world. Thus, a new conceptual framework is needed to advance the field of family media ecology.

3.1.1 Differential Susceptibility through Child Factors

Children may be differentially susceptible to digital media effects due to factors such as differences in temperament, which can change the probability that media will be used as an emotion regulation technique. Infants with difficult temperament, excessive fussing, or regulatory problems are exposed to more media (Reference Radesky, Silverstein, Zuckerman and ChristakisRadesky et al., 2014b; Reference Thompson, Adair and BentleyThompson et al., 2013). Parents who report that their children have more difficult temperaments also report that they are more likely to use media to regulate children’s emotions (Reference Coyne, Shawcroft and GaleCoyne et al., 2021; Reference Gordon-Hacker and Gueron-SelaGordon-Hacker & Gueron-Sela, 2020; Reference McDaniel and RadeskyMcDaniel & Radesky, 2018), which may in turn disrupt children’s ability to self-regulate as well as the parents’ ability to support children as they learn to identify and manage their mood and physiological states. The use of digital media to regulate contributes to a cascade of negative effects as children find it more difficult to regulate, more difficult to transition away from media (Reference Gordon-Hacker and Gueron-SelaGordon-Hacker & Gueron-Sela, 2020), and more difficult to sleep. Children then experience behavioral problems and parents in turn use more media to regulate children’s behavior (Reference Coyne, Shawcroft and GaleCoyne et al., 2021), perpetuating the negative cascade. Conversely, children with easygoing temperaments may not exhibit negative behaviors when asked to transition away from media, even in the context of regulatory use or engagement-prolonging design and may be less likely to develop as many problematic media use habits. Research is needed to assess these possibilities.

3.1.2 Differential Susceptibility through Sociocontextual Factors

According to the DSMM and multiple studies, distal contextual factors moderate digital media effects in several ways. For example, digital media exposure rates and practices differ by ethnicity/race (Reference Calvert, Rideout, Woolard, Barr and StrouseCalvert et al., 2005). Black and Latine/x parents are more likely than White parents to report discussing media content with their children (Reference Lauricella, Cingel and Beaudoin-RyanLauricella et al., 2016). Latine/x parents also report greater concern than White or Black parents about the risks of digital media exposure, and they are more likely to report setting rules about media usage (Reference Lauricella, Cingel and Beaudoin-RyanLauricella et al., 2016; Reference RideoutRideout, 2015). Some digital advertising is disproportionately targeted at families of color, which may impact usage practices (Reference Radesky, Reid Chassiakos, Ameenuddin and NavsariaRadesky et al., 2020b).

Digital media access also differs by socioeconomic status (SES: Reference Barr and BornsteinBarr, 2019b). There are disparities in access to quality content and stable internet access by SES (Reference Katz, Moran and OgnyanovaKatz et al., 2019). Most families now own mobile devices, but there are issues of ongoing inequality in the form of inconsistent access to the Internet due to poor bandwidth and reliance on older devices (Reference Katz, Moran and OgnyanovaKatz et al., 2019), suggesting there may be functional differences in media use for families with fewer resources compared with middle- or upper-SES families. Reference Linebarger, Barr, Lapierre and PiotrowskiLinebarger and colleagues (2014) demonstrated that children from lower-income homes who viewed more educational television content had better attentional and socioemotional regulation than those from lower-income homes who did not view high levels of educational media content, but this effect was not observed for children from higher-income homes. Relatedly, children from low-SES homes who viewed violent content and who viewed media in the evening were more likely to have a poor quality sleep than children from high-SES homes who had similar media use patterns (Reference Garrison, Liekweg and ChristakisGarrison et al., 2011).

The DSMM has been a useful model for examining variation in digital media effects, particularly across individuals. While the DSMM acknowledges digital media use is likely to vary as a function of developmental factors, such as attention skills, much of the research based on the DSMM has focused on school-age children and adolescents, with relatively less emphasis on infancy and early childhood. Moreover, the DSMM places little emphasis on the relational context of digital media use, which is especially important during early development. In addition, the theory acknowledges media content as it relates to individual differences (e.g., selection of media that matches one’s interests) but pays less attention to novel design affordances of newer media. The model also focuses on distal contextual factors (e.g., SES) with less attention to proximal contexts (e.g., parent motivations for using digital media, indirect effects of parents’ own media use). Filling some of these gaps, the Interactional Theory of Childhood Problematic Media Use has overlapping dimensions with the DSMM while additionally focusing on proximal contexts and the ways in which digital design maintains child engagement with media.

3.3.1 Parent Motivations for Digital Media Use

Some theories focus on the drivers of parent and child digital media use. For example, the Uses and Gratifications framework (Reference Rubin, Bryant and ZillmannRubin, 1986) posits that decisions to use media at all, and to select certain types of media, vary based on momentary needs and wants. While initially created to characterize adults’ own digital media use, it has been extended (albeit sparingly) to characterize parents’ decisions about their child’s digital media use. The IT-CPU focuses specifically on parents’ use of media to regulate their child’s emotional and behavioral states, but there are many other reasons that parents may use (or not use) media with and around their young children. For example, Reference NikkenNikken (2019) observed that parents report allowing child media use to meet a wide range of parents’ needs, such as occupying children so the parent can get things done. Other reasons for media use include (but are not limited to) escaping from stress and conflict, educating children, and bonding through shared media use. The causes and consequences are parents’ motivations to use (or not use) digital media with and around their young children are best understood through a family systems perspective, to which we turn next.

3.3.2 Family Systems Perspective

Family systems theory posits that interactions between different family members represent an open and ongoing system that may be influenced by the environment (in this case, digital media; Reference BroderickBrockerick, 1993; Reference Galvin, Dickson, Marrow, Braithwaite and BaxterGalvin et al., 2006). When individuals within the family system interact with one another, change and growth occur over time. Additionally, these patterns of interactions are shaped by external forces, such as digital media. This might play out in multiple ways (e.g., Reference Padilla‐Walker, Coyne and FraserPadilla-Walker et al., 2012). For example, a positive and warm relationship between parent and child might impact the child’s relationship with digital media, perhaps posting family events, choosing to interact with family members, and so on. Alternatively, the child’s relationship with digital media might impact the relationship between their parents. Perhaps the child is overusing media and shows problematic use and parents are arguing over best parenting practices. In other words, different systems within the family unit and around digital media shape the way those in the family interact over the course of development. In this way, a family systems approach can encompass relational factors (e.g., technoference, joint media engagement) that have been largely overlooked by other theoretical models.

Given that parents are often gatekeepers of young children’s media use, a complete understanding of the causes and consequences of media use requires one to consider family processes that lead to digital media use. Such processes are an integral part of the family media ecology and can be understood through a family systems lens. For example, some scholars have posited that strict prohibitions on children’s digital media use can be a risk factor for parental burnout (Reference Mikolajczak and RoskamMikolajczak & Roskam, 2018). Such a framing implies that parents use digital media as resources to promote parents’ own well-being, which could have positive downstream effects for children. Such a family systems lens would consider not only the direct effects of digital media use on the child but also indirect effects via parents’ mental health and wellbeing. We posit that the degree to which media use represents a risk versus a resource depends on several factors, including the reasons parents choose media (e.g., to take a break for themselves), the way media are used (e.g., solitary versus joint media engagement), and the way media use is experienced (e.g., as a source of parenting support/advice versus shame/doubt; as a constant distraction versus a much-needed respite). In this way, research on the causes and consequences of media use can consider the impact on multiple family members; what might be an opportunity cost for one family member (e.g., displacing opportunities for high-quality parent–child interaction) may be a resource for another family member (e.g., providing a moment to destress instead of lashing out; Reference Torres, Radesky, Levitt and McDanielTorres et al., 2021).

3.3.3 Relational Dynamics and Developmental Cascades

Child development is nested within early relationships with caregivers and siblings that shape each other over time (Reference SameroffSameroff, 2010). In addition, development across different domains is dynamic, and development in one domain affects development in another in the form of developmental cascades. These cascades result in both immediate changes in responses and cumulative long-term changes in child outcomes (Reference Tamis-LeMondaTamis-LeMonda, 2021). For example, when infants’ gradually transition from crawling to walking, they are able to explore objects in different ways which also elicits different verbal responses from others around them (Reference Karasik, Tamis-LeMonda and AdolphKarasik et al., 2016). A relational dynamics perspective recognizes that interactions among individuals and the environment are variable across short timescales. For example, in addition to large-scale developmental changes over time, one’s emotional and behavioral states vary moment to moment and day to day. Family members may feel more exhausted or more connected on some days than others.

Many theoretical approaches aimed at understanding digital media effects have an implicit assumption that effects are static, manifesting as between-person differences, although there are some exceptions (e.g., arousal response states characterized by the DSMM). Although the DSMM recognizes the transactional nature of children’s media use and preferences, it does not explicitly examine the relational processes that occur around media (e.g., parent–child interaction around tablets; Munzer et al., 2019), or are disrupted or supported by media (e.g., technoference; Reference McDaniel and CoyneMcDaniel & Coyne, 2016). Furthermore, the interplay between caregiver media use, child media use, and relational concepts such as attachment, co-regulation, and mentalization (i.e., caregiver mental models of their child; Reference Rosenblum, McDonough, Muzik, Miller and SameroffRosenblum et al., 2002) are not addressed by DSMM. For example, child regulatory uses of media are inherently a dyadic process in which a caregiver interprets a child’s emotional state, experiences their own emotional reaction that drives a co-regulation or media-regulation behavior, and experiences relief when the child’s affect is subsequently regulated. Although IT-CPU includes parenting style, parent–child relationship quality, and proximal parent–child interaction factors as drivers and maintainers of PMU, it does not consider how these factors shape each other over time, through micro-cascades in daily interactions or macro-cascades over different developmental windows. Furthermore, while IT-CPU’s narrow focus on PMU allows for examination of a clinically related outcome, a broader lens on social, emotional, language, cognitive, sleep, parent well-being, and other health and developmental domains is needed. A comprehensive theoretical framework would also allow for testing digital media effects both between and within individuals and across multiple timescales.

3.3.4 Cognitive Constraints and Developmental Considerations

Several researchers have focused on cognitive constraints and how they directly influence learning from digital media (e.g., Reference Barr and KirkorianBarr & Kirkorian, 2023; Reference FischFisch, 2000, Reference Fisch2017; Reference Hipp, Gerhardstein, Zimmermann, Barr and LinebargerHipp et al., 2017). Learning from digital media is cognitively challenging for children under 4 years of age because the images are presented in two dimensions and have to be related to objects in the three-dimensional world. Additionally, digital media often violate the physical constraints of the real world (e.g., abrupt jumps in time and space) and may incorporate fantastical or unfamiliar content. Learning from media therefore places a high degree of cognitive load on the child.

During early childhood, at least until the age of 4 years, children typically experience a transfer deficit, learning approximately 50% less from media presentations than from face-to-face interactions (Reference BarrBarr, 2013; Strouse & Samson, 2021). Several processes have been proposed to explain the transfer deficit. Media are perceptually impoverished and contain fewer cues compared to real-life presentations (e.g., Barr & Hayn, 1999). There are also fewer social and interactional cues available in media (e.g., Reference Hipp, Gerhardstein, Zimmermann, Barr and LinebargerHipp et al., 2017). Additionally, there are information processing constraints on learning. Until about 3 years old, children are less likely to remember information when cues at encoding do not match cues at retrieval, for example, when the color of the object is different in one situation than another. The ability to remember information in the face of such perceptual changes has been termed representational flexibility (Reference BarrBarr, 2013, 2019). Additionally, these young children do not fully understand that one object can represent or stand in for another object, which is termed a lack of symbolic understanding (Reference TrosethTroseth, 2010; Reference Troseth, Flores and StuckelmanTroseth et al., 2019). For example, children under 3 years of age may not understand that the screen is an object in itself and that actions shown on a screen can stand for or symbolize actions out in the real world. All of these cognitive constraints likely contribute to a transfer deficit in young children.

By 4 years of age, the transfer deficit typically declines, likely due to increased experience with media as well as the acquisition of age-related cognitive skills such as working memory and inhibitory control (Reference Barr and KirkorianBarr & Kirkorian, 2023; Reference Hipp, Gerhardstein, Zimmermann, Barr and LinebargerHipp et al., 2017; Reference KirkorianKirkorian, 2018). Once children overcome the transfer deficit, the extent to which they learn from media continues to depend on both children’s own characteristics (e.g., prior knowledge, working memory capacity) and the degree to which media content is designed in ways that increase or decrease cognitive load (Reference Barr and KirkorianBarr & Kirkorian, 2023; Reference FischFisch, 2000, Reference Fisch2017). The Science of Learning pillars of learning (Pillar 1: Active Learning, Pillar 2: Engagement in the Learning Process, Pillar 3: Meaningful Learning, Pillar 4: Social Interaction, see Table 1) have been applied to media content as quality indicators (Reference Hirsh-Pasek, Zosh and GolinkoffHirsh-Pasek et al., 2015). As described earlier, content analyses applying these learning pillars have found the quality of educational apps is typically poor (Meyer et al., 2021). However, a well-designed educational content will increase the likelihood of transfer of learning (Reference FischFisch, 2017; Reference Wright, Huston and MurphyWright et al., 2001).

Taken together, cognitive constraints on information processing, combined with media design, will result in different levels of cognitive load for individual children at different points in development (Reference Barr and KirkorianBarr & Kikorian, 2023; Reference Hipp, Gerhardstein, Zimmermann, Barr and LinebargerHipp et al., 2017). These constraints on processing need to be considered when determining media effects, including the degree to which educational media content may produce positive outcomes or buffer against negative outcomes of media use.

3.3.5 The Role of Human–Computer Interactions

Modern digital media (e.g., mobile devices, touchscreen apps, immersive video games, personalized video-sharing sites, video streaming services) have novel design affordances that require updated conceptual frameworks and measurement approaches compared to traditional media. With the introduction of smartphones in 2007 and tablets in 2010, children and parents could take a range of different media types (e.g., video, games, email, messaging) everywhere and use them at all times of day. Mobile devices can be taken into bedrooms, used during mealtimes and bedtime routines, and brought to other settings such as parks and outdoor play spaces. Handheld size and faster Internet speeds and computing power have led to family media ecologies in which digital devices are interwoven into daily life, used on-demand, and are gathering data about users to shape advertising and content delivery. Interactive touchscreen interfaces have also revolutionized how young children engage with and learn from media content, with potential for interactive scaffolds overcoming cognitive constraints or audiovisual enhancements causing distraction and cognitive overload.

Research on how design affordances interact with user psychology – the field of Human–Computer Interaction (HCI) – demonstrates that subtle changes in interactive design can shape human behavior in meaningful ways. For example, designs that constrain user choice, create a false sense of urgency, or prompt the user to return to technology frequently to meet the technology developer’s business interests of data collection or advertising impressions – at the expense of the user’s interests (i.e., time, money, sleep) – have been termed “dark patterns.” These manipulative design patterns have been documented as widespread in video games (Reference Gray, Kou, Battles, Hoggatt and ToombsGray et al., 2018), e-commerce websites (Reference Mathur, Acar and FriedmanMathur et al., 2019), and children’s apps (Reference Radesky, Hiniker and McLarenRadesky et al., 2022), but no studies have examined how they influence family well-being. Other relevant design affordances include engagement-prolonging designs that use variable rewards, low-friction (e.g., autoplay next episode), and high-pleasure content (e.g., infinite scroll with satisfying content) that extend children’s and caregivers’ time online at the expense of higher-friction play or relational activities.

HCI-related concepts relevant to parenting have been explored in interviews with parents, who described feeling “hooked” by their device’s rewards and notifications, overloaded by the amount of information available, and conflicted about whether to respond to child needs versus “escape” into a satisfying experience of social media (Reference Radesky, Kistin and EisenbergRadesky et al., 2016c). Similarly, analysis of children’s YouTube viewing habits suggests that the majority of children’s time spent on this platform includes low-quality user-generated content featuring commercial products attractive to children (e.g., candy, toys; Reference Radesky, Schaller, Yeo, Weeks and RobbRadesky et al., 2020c). In contrast, human-centered or child-centered design has interactive features that support the goals and agency of the user, allowing discovery of positive content as well as disengagement (Reference Radesky and HinikerRadesky & Hiniker, 2022).

Once design affordances and HCI concepts are emphasized in child development research, the importance of individual differences becomes more evident. Engagement-promoting features interact persuasively with users’ unique attentional, emotional, and social characteristics (Reference FoggFogg, 2009). First, children and parents have differential responses to interactive design nudges based on characteristics such as executive functioning, visual processing, or emotional reactivity. Children with weaker impulse inhibition may be more susceptible to following engagement-prolonging designs such as autoplay or in-game rewards. Parents with higher emotional reactivity might find the frictionless digital environment of social media more reinforcing when using media as a “virtual escape” in the setting of parent–child conflict. Furthermore, parent and child characteristics are inferred and profiled through their use of digital platforms such as social media, mobile games, and video-sharing services, which shapes what content they are recommended. Therefore, to understand how HCI processes drive or maintain media use behaviors it is important to characterize both the user’s strengths and weaknesses as well as the design affordances of media that they use most.

4.1.1 Individual, Relational, and Family Characteristics

Our framework includes individual child characteristics, such as child temperament and working memory, that have been associated with media effects. However, we consider that individual child characteristics such as temperament may also moderate how children respond to media. Other individual differences, including neurodiversity and gender development, also need to be considered. Similarly, parent characteristics (e.g., parent executive function), relational characteristics (e.g., child–parent attachment style), and family characteristics (e.g., presence of siblings, family structure, cultural practice) can influence whether and how media are used and how children and parents respond to media.

4.1.2 Structural Factors

The DREAMER framework also considers structural factors that can be associated with media use patterns and effects. For example, there are differences in media use as a function of household income and ethnicity/race. These findings are often descriptive and have not received enough empirical investigation. For example, African American children have higher rates of media exposure than children from other racial and ethnic groups (e.g., Reference Calvert, Rideout, Woolard, Barr and StrouseCalvert et al., 2005; Reference Hedderson, Bekelman and LiHedderson et al., 2023; Reference Kabali, Irigoyen and Nunez-DavisKabali et al., 2015). The reasons for this use within families of color with children under 5 are poorly understood, but could include factors such as lower access to child care or recreational centers in historically disinvested communities. In addition, the representation of different races and ethnicities within digital media can be considered both a structural factor (e.g., due to low diversity among writers, animators, actors) and a design affordance factor (e.g., due to platform recommendation of content containing racial/ethnic stereotypes or insufficient removal of white supremacist content that violates platform policies). In each case, existing research using global screen time measures is not explanatory. Structural factors can also moderate media effects. For example, because well-designed educational content has the potential to enhance learning and serve as a resource, educational content has stronger effect sizes for children from low-resourced homes (e.g., Reference Wright, Huston and MurphyWright et al., 2001). Conversely, low-quality content that includes engagement prolonging strategies is likely to exacerbate negative outcomes, particularly for children who do not have access to other activities (e.g., Reference Linebarger, Barr, Lapierre and PiotrowskiLinebarger et al., 2014). Other structural factors within low-income homes impact media use. For example, housing density is higher in low-income families, and siblings may share bedrooms and devices which changes media use patterns (e.g., younger siblings observing media content and activities of older siblings). As one example, infants with older siblings watch more child-directed content (Barr et al., 2010). Such dynamic, relational factors within the family have largely been ignored in the literature but can be tested within the DREAMER framework.

4.1.3 Media Motivations

We argue that during early childhood, parents act as gatekeepers of media use in the household and media may be used (or avoided) by both parents and children for multiple purposes. Parents’ (including mothers’ and fathers’) reasons for using and not using media predict media use patterns, such as whether, when, how, and with whom media are used. Such motivations may differ between parents in the same family system. For example, parents may use media to regulate their own emotions (e.g., to calm down when upset) or to regulate their children’s emotions and behavior (e.g., to calm them down, to help them fall asleep). If digital media are used to help a child fall asleep, then this family media use practice will determine the timing of media use in the household. Similarly, parents may have reasons for refraining from media use, such as wanting to protect children from harmful media content or preferring to engage in other activities (e.g., reading picture books, playing outdoors). Furthermore, parents use media to calm themselves down, to mentally escape, to relax or to take a break from their children, or to seek social support (Reference Suh, Kirkorian, Barr, Kucker, Torres and RadeskySuh et al., 2024; Reference Torres, Radesky, Levitt and McDanielTorres et al., 2021). These parental motivations for media use may result in technoference in the moment, but they may also improve parent–child interactions later on by providing a means for parents to receive support from friends or family. Additional research is needed to investigate how parents’ motivations for media impact child behaviors and parent–child relationships.

4.1.4 Media Use Patterns

We argue that media use patterns must be comprehensively measured to capture the quantity, content, design affordances, and proximal context in which digital media are used (e.g., when, with whom). These media use patterns differentially interact with child characteristics and parental factors. For example, if media are used at bedtime, then for children who find it more difficult to fall asleep, engagement-prolonging media may increase arousal and exacerbate sleep onset difficulty. We also argue that different media use patterns may predict different individual (e.g., feeling calm versus aroused) and relational responses (e.g., feeling connected versus distant). Parent use has typically only focused on mothers and not other caregivers in the family media ecology such as fathers, grandparents, and childcare providers. Similarly, most research examines a single focal child, overlooking the interdependence among siblings in households with more than one child.

4.1.5 Responses to Media

Immediate responses to media (e.g., calming down after a media break versus having difficulty transitioning away from media) are analogous to the cognitive, affective, and physiological responses states conceptualized in the DSMM. However, the DREAMER framework expands this conceptualization to add relational and behavioral responses and to represent all individuals in the family system. There may be individual differences in responses to media. For example, some individuals may respond with high arousal to media, whereas others have little to no response to the same content. Such responses can also differ in individuals over time based on different media use patterns. For example, a given child may have more difficulty transitioning away from media rife with engagement-prolonging design features than media without these features. As another example, solitary media use may lead parents to feel less connected with their child, whereas joint media engagement may increase their sense of connection.

4.1.6 Longer-Term Individual and Relational Outcomes

The DREAMER framework illustrates how immediate responses to media use can accumulate over time to produce longer-term outcomes. Such outcomes can also become mediators of future outcomes. For example, an immediate response of heightened arousal may disrupt the amount and quality of sleep. Over time, repeated sleep disruption can worsen cognitive and behavioral outcomes. Similarly, the degree to which media use increases versus decreases parents’ stress can have long-term implications for parental burnout. For example, remote work could be a double-edged sword, increasing time for parents but also potentially increasing cognitive load due to balancing work and home demands (Radesky et al., 2016). Finally, the combination of media use patterns by different family members could impact relationships over time. For example, due to decreasing costs of devices, the parent and child often view their own media on separate devices at the same time changing how media are shared (Reference Barr, Kirkorian and RadeskyBarr et al., 2020). Like other media effects frameworks, the DREAMER framework can be used to conceptualize potential impacts on a range of child (e.g., language, academic achievement, self-regulation, social competence), parent (e.g., burnout, self-efficacy), and relational outcomes (e.g., parent–child interaction quality) while taking into account the dynamic and relational use of media within the family media ecology.

5.1.1 Media Attitudes Questionnaire

The Media Attitudes Questionnaire (MAQ) represents global measures of media attitudes, media practices, and individual and structural factors that may affect both parents’ and children’s media use and responses. It includes questions about structural factors and parent attitudes and behaviors regarding media, which maps on to the structural factors and individual child and parent characteristics sections of the DREAMER framework. There are also questions about the amount, content, and context of media use. Context questions capture a wide range of motivations for media use, including instrumental purposes (e.g., occupying the child so the parent can get things done or take a break) and regulatory purposes (e.g., calming down) for both parent and child media use. Other context questions ask about joint media engagement and technoference. This questionnaire is supplemented with other widely used measures of both parent and child factors, including parental stress, parental burnout, child language, child sleep, and parent and child problematic media use.

5.1.2 Time Use Diary

The time-use diary is an online survey that logs the child’s activities (e.g., sleep, digital media, outdoor play) throughout the entire day as well as overnight sleep. It can be completed first thing in the morning (asking about yesterday’s activities) or at night (asking about today’s activities). Digital media content and context (media use in the model) is captured using specific questions on what the child was watching and with whom. Follow-up questions capture the context of media use including technoference (e.g., presence of background TV or parent mobile device use during play and mealtimes), media content (e.g., target age group, title of program), and joint media engagement (i.e., other people engaging in child media use). A proxy for physical activity can also be calculated from the reported time playing outdoors. Because the time use diary logs the onset and offset of all activities throughout the day, including overnight sleep, it can be synchronized with other time-stamped data streams (e.g., Chronicle passive mobile sensing data capturing parents’ frequency of checking their mobile device during each child activity throughout the day).

5.1.3 Passive Mobile Sensing with the Chronicle App

Mobile device use is assessed with Chronicle, a passive mobile sensing app for iOS and Android devices to provide continuous data on the timing and content of mobile device use. Chronicle is installed on mobile devices, including devices owned by the child when available. This method is particularly robust for capturing the harder-to-remember, short bursts of activity associated with mobile device use (Reference Barr, Kirkorian and RadeskyBarr et al., 2020; Reference Radesky, Weeks and BallRadesky et al., 2020a). Passive sensing is more accurate than parents’ reporting of their own mobile device use (Reference Yuan, Weeks and BallYuan et al., 2019) and their preschool-age children’s smartphone and tablet use (Reference Radesky, Weeks and BallRadesky et al., 2020a). Reference Yuan, Weeks and BallYuan and colleagues (2019) conducted a study to test the feasibility of passive sensing on the smartphones of parents with young children. They also collected self-reported data from the same parents on their estimation of both the total duration of use and the pickup frequency. Parent self-reported screen use was inaccurate, with most parents underreporting rather than overreporting both duration and pickup frequency. This imprecision is likely due to the handheld nature of mobile device use (i.e., parents may have more difficulty monitoring mobile device use when children are in different rooms) and engagement that is both intermittent and immersive. In addition, data collection with Chronicle has provided insight into children’s overnight device usage, access to age-inappropriate mobile games, and inequities in access to quality content (Reference Radesky, Weeks and BallRadesky et al., 2020a). Output from the Chronicle Android dashboard provides app lists for each user, including the title, time, and duration of each app used during the data collection period and for categories of use (e.g., social media, productivity) for iOS devices. This method provides an objective measure of capturing media use for both parents and children. The time-stamped app usage logs can be integrated with other time-stamped data streams (e.g., EMA, time use diaries) to understand parents’ and children’s device use in naturalistic contexts.

5.1.4 Ecological Momentary and Daily Assessment (EMA and EDA)

Most research relies on parents’ global estimates of “typical” media use, disregarding the real-time family dynamics that underlie parent and child emotions and behavior that lead to, and result from, media use. The EMA and EDA approaches permit examination of multiple parts of the model including media use, motivations, responses, and outcomes, and so represent a particularly useful tool in examining dynamic and relational components of the family media ecology. This method involves an intensive longitudinal design where participants receive multiple EMA surveys each day (usually via text message) about real-time media use, motivations for such use, and responses to it (e.g., family connectedness, child behavior, and parent emotional states). Additionally, participants receive an EDA survey that allows them to reflect on the overall day (for example how typical their child’s media use was or how confident they felt in managing their child’s media use for the day). The EDA surveys ensure capturing shorter or less frequent behaviors (e.g., video chat, e-books) that may not be captured using EMA alone. Both surveys can also be expanded to capture data relevant to individual studies (e.g., collecting URLs for YouTube videos watched by the child each day to assess content quality). Most existing research on media use during early childhood has used a between-subjects approach, where individuals with high screen use are compared to those with low screen use, for example. Between-person effects reveal differences at the group level but do not allow for individual heterogeneity to exist and likely mask the true development and impact of media within families. Individual heterogeneity and dyadic processes are marked by moment-to-moment and day-to-day fluctuations around media use, which are characterized as within-person effects. The use of EMA and EDA allows for both between and within comparisons, and in so doing emphasizes dynamic and relational aspects of the DREAMER framework across multiple timescales. These surveys can also be synchronized with other time-stamped data streams (e.g., parents’ mobile device use in the hour preceding the EMA survey message).

5.1.5 Future Directions: Assessment of Design Affordances

Because the CAFE Toolkit allows assessment of the vast array of mobile, television, and video content that children and caregivers use, through survey, time use diary, and passive sensing methods, an important next step is to classify the quality of this content. The approach to content and design affordance analysis will vary based on the hypotheses of the particular study – for example, the impact of YouTube educational content on language outcomes in a toddler being cared for by Spanish-speaking grandparents; the impact of engagement-prolonging designs on sleep quantity in a child with weaker impulse control in the context of lax parenting; or the impact of negative media ethnic representations on children’s social and identity development. Platform algorithms bin children by characteristics to recommend individualized content, that may or may not support child development. While many factors shape identity development, future research on the effects of such algorithms on identity development and other long-term outcomes is necessary. In addition, content and design affordance coding will need to evolve as digital product design evolves (e.g., evaluating trending content; the impact of different algorithmic recommendations). For our current studies, design affordance coding has largely focused on aspects of interactivity that introduce a heightened cognitive load (e.g., disruptive advertising such as pop-up ads; irrelevant hotspots), contribute to extended time on or difficulty transitioning away from media, recommend negative content, or constrain user agency (see Table 2).