Parent personality, child temperament, and digital media: Pathways to language development in early childhood

Abstract

Over the past decade, children’s digital media exposure has increased rapidly, raising concerns about its potential negative impact on language development. While broader family ecology (especially parents) and the context of how media is used, are known to matter in shaping children’s language development, the roles of parent personality and child temperament in shaping media consumption remain understudied. To address this gap, this study introduces a holistic model examining parent personality, child temperament, and media use as predictors of language development in two-year-old children, both concurrently and longitudinally. In Study 1, 464 caregivers (n_female=222) of 17–30-month-old monolingual English-speaking children (n_female=323) were recruited via Cloud Research and Prolific. Caregivers were 76.94% White and had a median education of a 4-year-college degree. The study assessed parent personality, child temperament, media use duration, motivations and contexts of media use, and child language outcomes. Study 2 measured a subset of these caregivers (n=90) one year later. Results suggest parent conscientiousness directly impacts children’s media use and indirectly impacts vocabulary. The impact of children’s negative affect on vocabulary is mediated through their technology use. Educational motivations for using media and presence of joint social context did not moderate media’s impact on vocabulary. However, duration of digital media use at 2-years-old predicted vocabulary one year later, regardless of parent personality or child temperament. This study emphasizes the importance of the broader family ecologies, including parent personality and child temperament, as well as longitudinal impacts of media use, calling for more nuanced, context-sensitive guidance over one-size-fits-all recommendations.

In the last few decades, there has been an exponential rise in digital media use such that the age at which children begin being directly exposed to media has shifted significantly, from an average of four years old in 1970 to as early as four months old today (Reid Chassiakos et al., 2016). There have also been recent shifts in how much media children under 3-years consume – exposure rates have doubled in just the last few years, from under one-hour daily in 2020 (Rideout & Robb, 2020) to more than two hours of direct screen time by 2023 (Kucker et al., 2024). This increase in digital media use raises concerns about the impact of media on children’s language development (Paulus et al., 2019).

However, a large body of evidence shows that children’s language outcomes are influenced by more than just the amount of digital media exposure (e.g. Tamis-LeMonda et al., 2019). Language development during the first few years of life is also shaped by factors such as why and how media is used, parent characteristics, and individual differences in children’s traits, all of which influence outcomes both concurrently and longitudinally (Barr et al., 2024; Fitzpatrick et al., 2022; Lee et al., 2022; Sundqvist et al., 2021, 2024). Although media use differs considerably across households, the interplay between parent and child individual differences and their influence on 2-year-old’s digital media consumption remains understudied Current recommendations primarily focus on limiting screen time, overlooking critical variations in who is using digital media, why digital media is being used, and the context in which it is used.

In response to this gap, this paper introduces a holistic model that integrates parent personality and child temperament with the motivations (why) and contexts of digital media use (how) to predict language development outcomes both at age two and one year later. This approach contributes to a more nuanced understanding of how family dynamics may contribute to early language growth amid growing digital media exposure.

Digital media use prevalence and impact on language

Even during infancy, digital media is a ubiquitous part of everyday life. This exposure has significant consequences for language development during the early years. As children spend more time with digital media, the amount of language input they receive from parents and other social partners decreases (Brushe et al., 2024; Christakis et al., 2009). This leads to a reduction in their vocalizations (Christakis et al., 2009) and vocabulary growth (Sundqvist et al., 2021), both concurrently and over time (Madigan et al., 2020; Sundqvist et al., 2024). Moreover, the quality of language input from both caregivers as well as the digital media quality itself diminishes as children’s exposure to media rises (Kolak et al., 2023; Pempek et al., 2014). Such findings have led the American Academy of Pediatrics (AAP) to issue recommendations to limit the time children spend on screens (Council on Communication and Media, et al., 2016).

However, emerging research suggests that the impact of digital media on language outcomes is more nuanced than what is captured by total screen time alone (see also Kucker, 2024). Though the majority of children’s media time is still spent with videos/TV rather than interactive games (Kucker et al., 2024), there is great variability for why those videos/TV are used. Not all households have the same motivations for using digital media or use it in the same way – some parents and children use it purely for educational reasons, others for entertainment or emotional regulation strategies, and some families use digital media as a joint social activity, whereas other children use media alone (see e.g. meta-analyses by Madigan et al., 2020; Jing et al., 2023). These differences in motivations and social context of digital media use change the language learning environments for children and thus, have consequences for development.

The motivations for why individual families use digital media vary widely (see Suh et al., 2024), and emerging evidence suggests that these differences have downstream effects on language outcomes (Sundqvist et al., submitted). Some motivations for media use (e.g. for education), may be more beneficial than others (e.g. media used to entertain or calm a child). Educational use of digital media and educational programs/apps have been found to correlate with higher language skills (Jing et al., 2023), in part because educational media actively engages children in processing content and provides language input that mimics real social interactions (Linebarger & Walker, 2005). For example, Neuman et al., (2019) examined top language and literacy focused educational programs available through streaming platforms and found that attention-directing and ostensive cues helped children better focus on the target vocabulary and program in general. Such educational effects are particularly beneficial when they involve increased repetition of terms, imaginative strategies, and the use of narratives (Linebarger et al., 2017). On the other hand, digital media used purely for entertainment tend to lack these elements and have negative effects on children’s development longitudinally (Zimmerman & Christakis, 2007). This suggests that digital media used with an educational intent may have a more positive impact on language development, both concurrently and longitudinally, compared to media consumed primarily for the purposes of entertainment (see Kucker et al., 2024).

In addition to the motivations for why media is consumed, how it is used, especially socially, also varies widely across families and impacts language. For example, when digital media is co-viewed and used actively with a social partner in a joint context, it is associated with gains in children’s learning and vocabulary development (Jing et al., 2023; Madigan et al., 2020). In fact, both socially contingent virtual partners (such as via video chat) and real-life social partners that build and model virtual content have been found to enhance word learning in 2.5-year-old children (Strouse et al., 2018). Socially interactive media may be helpful to learning as it maintains contingent and responsive elements that are key for communication (Roseberry et al., 2014). In addition, co-use of digital media with a live social partner has been shown to have a more lasting positive impact on children’s language development compared to using digital media alone (Dore et al., 2020). Taken together, these findings reinforce the need to consider not just how much digital media children consume, but how (for educational reasons or not) and with whom (with a parent or alone) they engage with it, both concurrently and longitudinally.

Family factors impact on digital media use and language: Parent personality

Parents play a central role in multiple elements of children’s early development, including their digital media habits such as how much they use, why they use it, and under what circumstances. Especially in early childhood, digital media access and usage depend largely on parental control over access and supervision, including setting rules and managing device availability (Nikken & Schols, 2015). Yet despite their influence, parent’s role has been largely overlooked in research on children’s digital media use.

A small but growing body of literature suggests that broader familial factors such as socioeconomic status (Tandon et al., 2012), parental education (Dynia et al., 2021), and the presence of siblings (Barr et al., 2010; Paulus et al., 2024) influence children’s digital media consumption. Some studies have also targeted parent’s specific characteristics and behaviors in shaping young children’s digital media use and screen time. For example, studies indicate that children aged 4–6 years with permissive or neglectful parents tend to engage more heavily with digital media, while those with authoritarian or authoritative parents generally exhibit lower levels of media use (Lee et al., 2022; see also Linebarger et al., 2014). Other findings highlight that parent stress can predict increases in digital media use by 3-year-olds over time (McDaniel & Radesky, 2020), and that parents’ own media habits during their children’s infancy predict children’s digital media use two years later (Holmgren et al., 2024). However, there has been significantly less research focusing specifically on parental traits. This is a critical gap as parents act as regulators of digital media within the household, and their thoughts, feelings, and behaviors (i.e. their personality) may directly affect how their children engage with media. In particular, parent personality might influence both the motivations for and approaches to digital media exposure for their children, factors that could be crucial in understanding digital media’s impact on language development.

The Big Five personality model (John & Srivastava, 1999)—encompassing neuroticism, extraversion, agreeableness, openness, and conscientiousness—has been shown to capture parents’ everyday behaviors and decisions. For instance, higher extraversion and openness are associated with authoritative and emotionally involved parenting, while lower levels of these traits are linked to authoritarian and emotionally detached styles (Metsäpelto & Pulkkinen, 2003). Parents who exhibit higher extraversion, agreeableness, conscientiousness, and openness, along with lower neuroticism, are more likely to show warmth and discipline (such as rule setting). Additionally, higher agreeableness and lower neuroticism are associated with greater autonomy support (Prinzie et al., 2009). These parenting styles influence children’s digital media use (Lee et al., 2022; Linebarger et al., 2014).

Parent personality may also directly influence digital media behaviors. For example, parents with high neuroticism may experience greater stress (Uliaszek et al., 2010) and greater stress is correlated with higher digital media use (McDaniel et al., 2024). Likewise, adults with lower conscientiousness are often more likely to use digital media themselves (Whaite et al., 2018). Both parental stress and media consumption can contribute to increased media use in children (Holmgren et al., 2024; McDaniel & Radesky, 2020). Despite these established links between parent personality traits, parenting styles, and media behaviors, little research has specifically examined how parent personality is directly linked to children’s digital media use, especially in children under three years old. By exploring the facets of the Big Five personality traits, this study aims to investigate potential pathways through which parent personality may influence young children’s media habits. Because digital media also plays a role in shaping children’s language, these pathways may have implications for vocabulary development.

Parents also play a key role in their child’s language development, primarily because they serve as one of the primary sources for language input and contextual support for vocabulary growth. Not only does a parent’s style and stress impact their interactions with their child and thus their language development, but parent personality contributes to their language development as well. For example, conscientiousness has been shown to be a positive predictor of education, income, and socioeconomic status, both in the short term and long term (Damian et al., 2015; Duckworth et al., 2012; Egan et al., 2017), all of which are linked to children’s language development (Hart & Risley, 1995). Additionally, Kucker and colleagues (2021) found that parent personality accounts for a significant portion of variance in children’s expressive vocabulary, even when controlling for child age. Specifically, parental traits like conscientiousness, openness, and agreeableness positively correlate with children’s expressive vocabulary size at 17–30 months. Given that these same traits of adult personality are associated with their own adult social media use and engagement (Lampropoulos et al., 2022), examination of the interconnectedness of parent personality with child digital media use and language is particularly relevant.

Family factors impact on digital media use and language: Child temperament

Though parents can play a critical role in children’s digital media use and language, the child’s own individual characteristics, particularly their temperament, also plays a role. Temperament encompasses differences in how children respond to environmental stimuli, including their emotional regulation and activity levels (Rothbart & Bates, 2007). Current temperament models identify three primary dimensions: surgency, negative affectivity, and effortful control (Rothbart, 2007), each of which may impact children’s media use and language.

Emerging research suggests a complex interplay between temperament and digital media use among young children. For example, Radesky et al., (2014) observed that children exhibiting poor self-regulation skills—a key component of effortful control—at 9 months demonstrated increased television viewing at 2 years, underscoring a longitudinal relationship between temperament and media consumption. This tendency for children with low self-regulation to engage more with media may also affect the quality of engagement in joint media use, as parents may use screens to manage behaviors rather than to participate actively with their child. Brauchli et al., (2024) also showed bidirectional longitudinal links where higher levels of negative affect (but not effortful control) were associated with higher levels of screen exposure in high socioeconomic status 1- to 3-year-olds. Additionally, another study by Coyne et al. (2021) found that higher levels of negative affectivity and surgency in 2–3-year-olds correlated with problematic digital media use media use that negatively impacts daily functioning and is characterized by high consumption and poor emotional regulation around digital media. This study indicated that parents might use digital media to regulate children’s emotions, such that motivation for use mediates the relation between temperament and media consumption. In these cases, the reliance on digital media to manage challenging temperaments could reduce opportunities for parents to model healthy emotional regulation strategies or engage in responsive joint media use. However, findings concerning surgency are inconsistent. Shin et al., (2021), for example, found no link between surgency and screen time in children aged 1.5–3 years but did report that higher negative affectivity and lower effortful control correlated with increased screen time when controlling for rates of parental stress. These findings highlight how parent stress, potentially influenced by the child’s temperament, may lead to increased media use at the expense of shared, interactive media experiences. Furthermore, Thompson et al., (2013) identified a persistent link between early TV exposure and temperament traits in infants as young as 3 months, suggesting that infants perceived as more active or fussier had higher levels of TV exposure. Collectively, these studies indicate a potential trend wherein parents may rely more heavily on media to manage the behaviors of children with challenging temperaments. This may, in turn, impact why and how media is used which, as discussed, matters for children’s language development.

Children’s language development is also associated with their temperament. For instance, Kucker et al., (2021) reported a positive correlation between surgency and effortful control, while negative affectivity was negatively associated with measures of child vocabulary and mean length of utterance. Additionally, Davison et al. (2019) found that components of surgency assessed at 8 months, positively correlated with vocabulary scores at 24 months. These findings suggest a significant relation between temperament and language development both concurrently and longitudinally. Furthermore, a recent study revealed that children with high negative affectivity and low effortful control exhibited less engagement in conversations, which in turn was linked to poorer language skills later in life (Fields-Olivieri et al., 2024). This underscores that children showing high negative affectivity and low effortful control may be less engaged with their learning environments, thereby adversely affecting their language development.

While there is a growing body of research examining associations between temperament and media use and young children’s language development, there remains a significant gap in the literature regarding a model that integrates these factors with why and how media is used. Understanding how these components shape each other can provide deeper insights into the ways both parents and children influence digital media experiences and, consequently, their language outcomes.

Current Study

The current study aims to address the existing gaps in the literature by investigating the effects of parent personality, child temperament, and child digital media use in young children, utilizing both concurrent (Study 1) and longitudinal (Study 2) approaches. This study has three primary research questions (RQ) exploring mediation (RQ 1), moderation (RQ 2) and moderated mediation (RQ 3) pathways between parents, children, aspects of media use (duration, motivations for use, and joint use), and child language (vocabulary).

RQ 1a: Does the duration of digital media use mediate the relation between parent personality and child vocabulary?

RQ 1b: Does the duration of digital media use mediate the relation between child temperament and child vocabulary?

RQ 2a: Does using digital media for educational purposes moderate the relation between duration of digital media use and child vocabulary?

RQ 2b: Does using digital media in a joint context moderate the relation between duration of digital media use and child vocabulary?

RQ 3a: Are there moderated mediation effects such that media use duration mediates the link between parent personality and vocabulary (RQ1a) with educational motivation moderating the link between media use duration and vocabulary (RQ2a)?

RQ 3b: Are there moderated mediation effects such that media use duration mediates the link between parent personality and vocabulary (RQ1a) with joint context of use moderating the link between media use duration and vocabulary (RQ2b)?

RQ 3c: Are there moderated mediation effects such that media use duration mediates the link between child temperament and vocabulary (RQ1b) with educational motivation moderating the link between media use duration and vocabulary (RQ2a)?

RQ 3d: Are there moderated mediation effects such that media use duration mediates the link between child temperament and vocabulary (RQ1b) with joint context of use moderating the link between media use duration and vocabulary (RQ2b)?

Hypotheses and study design were all pre-registered and raw data files and analysis scripts are available on OSF (https://osf.io/tgz4a/?view_only=bd86f2a78d92458c8a0738718c9ea115).

Study 1

The first study explores the concurrent relations between parent personality, child media use and vocabulary (RQ1a) as well as child temperament, media use, and vocabulary (RQ 1b) in a large sample of caregivers of children between 17 and 30-months-old. It also measures potential moderators of digital media’s impact on vocabulary (RQ 2a and 2b), and the complex relationship with both personality/temperament and moderators (RQ 3).

Methods

Participants

A total of 464 parents of typically developing monolingual English-speaking children between 17–30 months in the United States were recruited through Cloud Research and Prolific. From the total participants, 242 participants were selected from archival data, originally collected via Cloud Research in a study of children’s media use and language development February 2022-May 2023 (Kucker et al., 2024). The other 222 completed similar surveys and were collected through Prolific in August-October 2024. All data were cleaned for validity based on Chmielewski and Kucker (2020), which included a thorough analysis of multiple attention and validity checks. This process led to an additional 243 participants who completed the surveys but failed these screening checks to be excluded from the final sample. See Supplemental Materials for details on each step of the cleaning process. All individuals provided informed consent prior to participating and all procedures complied with APA ethical standards. There were no substantial differences in demographic traits between the two waves of data collection, therefore the samples are collapsed together. Demographic information is reported in Table 1. According to a priori recommendations for conducting moderation/mediation analysis, at least 400 participants are needed to detect a small effect (Fritz & MacKinnon, 2007; Preacher et al., 2007). A similar sample size (n = 402) is needed to detect a small effect (f² = .05) in the pre-registered exploratory linear regressions (reported in the supplemental materials) with 5 tested predictors according to a priori power analysis using G*Power 3.1.9.7. Thus, the sample size is sufficient for the planned analyses.

Table 1.

Demographic Traits of the Sample

	Study 1 (concurrent sample n =464)	Study 2 (longitudinal sample n = 90)	Comparison of Study 1 and Study 2 samples at Time 1
Child Age (Mos;Days)	M = 23;26 (SD = 3;24) [Range = 16;24 – 31;29]	M = 36;10 (SD = 4;12) [Range = 27;7 – 46;3]	t(462) = .707, p = .480
Child Gender	Male = 241 Female = 222 Not Reported = 1	Male = 48 Female = 42	χ2 (1, 463) = .07, p = .786
Child Race	White = 332 Black = 62 Asian = 8 Native American/ American Indian = 3 Multiracial = 55 Not Listed = 3 Not Reported = 1	White = 72 Black = 9 Asian = 1 Native American/ American Indian = 0 Multiracial = 8 Not Listed = 0 Not Reported = 0	χ2 (5, 460) = 6.98, p = .222
Child Ethnicity	Hispanic = 64 Not Hispanic = 400	Hispanic = 7 Not Hispanic = 83	χ2 (1, 464) = 2.26, p = .133
Parent Gender	Male = 140 Female = 323 Non-binary = 1	Male = 27 Female = 63 Non-binary = 0	χ2 (1, 463) = .001, p = .982
Parent Race	White = 357 Black = 73 Asian = 9 Native American/ American Indian = 1 Multiracial = 15 Not Listed = 6 Not Reported = 2	White = 77 Black = 10 Asian = 0 Native American/ American Indian = 0 Multiracial = 3 Not Listed = 0 Not Reported = 0	χ2 (5, 464) = 5.07, p = .4074
Parent Ethnicity	Hispanic = 42 Not Hispanic = 422	Hispanic = 4 Not Hispanic = 86	χ2 (1, 464) = 2.88, p = .090
Parent Education a	M = 5.41 (SD = 1.37) [Range = 2–8]	M = 6 (SD = 1.28) [Range = 3–8]	t(462) = −1.21, p = .226
Household Income b	M = 9.22 (SD = 4.86) [Range = 1–21]	M = 10 (SD = 5.11) [Range = 1–21]	t(459) = −1.05, p = .293

Note.

^aEducation was rank ordered; less than 7^th grade reported as a 1, Doctoral degree reported as 8; 4-year college degree was equivalent to a score of 6.

^bIncome was rank ordered in $10,000 increments. “Less than $10,000” scored as 1 up to “over $200,000” as 21.

Comparison between samples used an independent samples t-test or chi-square to measure if those only in Study 1 (time 1) were statistically different from those who also completed Study 2 (both time 1 and 2). Chi-square categories with <5 were not included.

Materials

All parents completed a set of questionnaires on their own personality, their child’s temperament, their child’s media use, and their child’s vocabulary. Descriptive statistics are reported in Table 2.

Table 2.

Descriptives of Measures

		Study 1	Study 2	Comparison of Study 1 and Study 2 samples at Time 1
Child Vocabulary *	Total vocab size	188.7 (175.60)	45.51 (30.53)	t(462) = −.087, p = .930
Child Media Use	Media duration (Mean min/day)	126.82 (101.91)	112.56 (91.43)	t(445) = −.67, p = .506
	Educational motivation for media use	Educate = 362 Not to educate = 74 No video use = 28	Educate = 52 Not to educate = 37 No video use = 1	χ2 (1, 436) = .95, p = .330
	Joint context of media use	Low joint = 202 High joint = 234 No video use = 28	Low joint = 43 High joint = 46 No video use = 1	χ2 (1, 436) = .44, p = .507
Parent Personality	Neuroticism	2.84 (.94)	--	t(462) = 1.24, p = .216
	Extroversion	2.98 (.84)	--	t(462) = 1.47, p = .142
	Openness	3.59 (.63)	--	t(462) = −.59, p = .558
	Agreeableness	3.98 (.63)	--	t(462) = −.37, p = .709
	Conscientiousness	3.88 (.69)	--	t(462) = −.70, p = .484
Child Temperament	Negative Affect	3.37 (.80)	--	t(462) = 2.71, p = .007
	Surgency	4.99 (.67)	--	t(462) = −1.24, p = .217
	Effortful Control	4.69 (.60)	--	t(462) = 1.40, p = .163

Note. Mean shown with SD in parentheses.

^*Note due to age-related change, different versions of MCDI were used for studies 1 and 2. Time 1 expressive vocabulary was measured with the MCDI:WS (total out of 680 words); Time 2 expressive vocabulary was measured with the MCDI:III (total out of 100 words). Parent personality and child temperament were not measured at Time 2.

Comparison of those in only Study 1 to those in both Study 1 and 2 show no differences in vocabulary, media time, or parent personality at Time 1; samples differed only in children’s level of negative affect with those participating in both time points having lower NA

Parent personality.

Parent personality was assessed using the parent’s self-report measure of the Big Five Inventory (BFI; John & Srivastava, 1999). The questionnaire includes 44 items that capture five personality traits: Neuroticism, Extraversion, Openness, Agreeableness, Conscientiousness. Each trait is measured on a 1–5 scale. This scale has shown evidence of construct and convergent validity, as well as test-retest reliability (John & Srivastava, 1999). Cronbach’s alphas for the current sample were in the normal range: Neuroticism α = .877, Extraversion α = .853, Openness α = .796, Agreeableness α = .810, Conscientiousness α = .837.

Child temperament.

The Early Child Behavior Questionnaire-short form (ECBQ; (Putnam & Rothbart, 2006) was used to measure child temperament. The ECBQ measures three main domains: Negative Affect, Surgency, and Effortful Control, across a set of 107 items, each on a 1–7 scale. It is psychometrically reliable and validated for use in early childhood (Putnam & Rothbart, 2006). Cronbach’s alphas for the current sample were in the normal range: Negative Affect α = .935, Surgency α = .825, Effortful Control α = .842.

Children’s media use.

Children’s media use and exposure was assessed using questions from the Media Assessment Questionnaire version 2.1 (MAQ; Barr et al., 2020), particularly the duration or amount of digital media use (average minutes using digital media each day), motivations for technology use (for educational purposes or not), and joint context of digital media use (low joint media use or high joint media use). The current study focused exclusively on video/TV media use as it represents over 80% of children’s total digital media use at this age (Kucker et al., 2024). Educational motivations were coded as a binomial if videos were used for educational purposes at least some of the time or videos were not used for educational purposes. Joint context of digital media use was measured using a 5-point scale asking how frequently parents jointly engaged with videos with their children. For the purposes of the moderation, those who indicated 50% of the time or less (score of 1–3) were coded as “low joint context” and those with more than 50% joint use were coded as “high joint context”

Child’s language.

Parents completed the MacArthur-Bates Communicative Development Inventory: Words and Sentences (MCDI-WS; Fenson et al., 1994) that assesses children’s expressive vocabulary from a list of 680 words. This inventory has strong criterion validity and correlations with direct measures of vocabulary (Heilmann et al., 2005).

Procedure

This study was approved by the Internal Review Board at [removed for review]. All participants completed the measures above online via Qualtrics which took approximately 57-minutes to complete (SD = 23.7 minutes). Demographic questions were presented first, followed by vocabulary, child temperament, and parent personality, and ending with the media survey. Participants were compensated $5–8 for participating in the study.

Analysis

First, bivariate correlations between each of the variables at Time 1 were run using SPSS. Significant correlations of child age, parent education, and/or parent income with child media duration and/or vocabulary were included as covariates in subsequent models. Second, we ran specific analyses for each of three aims for Study 1.

The first aim explored a mediation pathway to assess whether the duration of media use mediates the relationship between parent personality and child temperament on vocabulary. Two mediation models were run using the sem function from the lavaan package in R (Rosseel, 2012) in which indirect effects were assessed via non-parametric bootstrapping. Listwise deletion was used for missing data. In model 1a, each of the five facets of parent personality (centered) predicted the child’s total media duration (average minutes/day; centered), which in turn predicted the child’s total vocabulary size (centered). Parent education was included as a covariate predicting total media duration and child vocabulary; child age was included as a covariate in predicting child vocabulary. In model 1b, each of the three domains of temperament (each centered) predicted child’s media duration (again, centered), which in turn predicted the child’s vocabulary (centered). The same covariates were included.

The second RQ investigated a moderation pathway to assess how the relation between digital media use duration and vocabulary development was moderated by 1) the motivations for media use and 2) joint context of media use. Two moderation analyses were run using the lme package in R, one for each moderator (Bates et al., 2015). Model 2a included children’s total vocabulary size as the outcome, total media duration (centered) as a predictor, and motivation for use (educational or not) as a possible moderator. Parent education and child age were again included as covariates. Model 2b was identical except for joint context of use (joint use high or low) was included as the moderator instead of motivation.

The third aim explored a moderated mediation pathway by combining elements of RQ1 and RQ2, to gain a more comprehensive understanding of these interactions. Moderated mediation models using lavaan¹ examined the role of parent personality/child temperament on media use duration and vocabulary with motivation (models 3a and 3c) and joint context (models 3b and 3c) as moderators of the association between duration of media use and vocabulary.

Transparency and Openness

Hypotheses and study design were all pre-registered and raw data files and R scripts for the analysis are available on OSF (https://osf.io/tgz4a/?view_only=bd86f2a78d92458c8a0738718c9ea115). Analyses used SPSS version 29 for bivariate correlations and descriptive statistics and R Studio 2024.04.2 for mediation and moderation analyses. Pre-registered exploratory analyses are presented in the supplemental materials.

Results

First, bivariate correlations were run to examine correlations between the variables of interests. See Table 3. Children’s raw vocabulary scores were positively correlated with their age as expected, r (462) = .42, p < .001, and negatively associated with their total duration of media time, r (445) = −.17, p < .001. Parent education was also positively associated with vocabulary, r (462) = .11, p = .019 and vocabulary was positively correlated with children’s surgency, r (462) = .14, p = .002 and effortful control, r (462) = .20, p < .001. Children’s average duration of digital media use was negatively associated with parent education, r (445) = −.16, p <.001, and parent income, r (442) = −.23, p < .001, and positively associated with child negative affect, r (445) = .15, p = .001. Because of the positive association between parent education and both child vocabulary and media use, parent education was included as a covariate in the subsequent analysis. Child age was included as a covariate for child vocabulary.

Table 3.

Bivariate Correlations

	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15
1. Child Age T1	1
2. Child Age T2	.93**	1
3. Total Vocabulary T1	.42**	.41**	1
4. Total Vocabulary T2	.22*	.25*	.66**	1
5. Media Duration T1	0.05	0.05	−.17**	−.34**	1
6. Media Duration T2	−0.01	−0.03	−0.06	−0.18	.51**	1
7. Parent Education	0.07	0.04	.11*	0.18	−.16**	−.25*	1
8. Income	0.04	−0.01	0.07	0.16	−.23**	−.23*	.50**	1
9. Parent Neuroticism	0.00	0.01	0.01	−0.15	−0.03	0.16	−.21**	−.20**	1
10. Parent Extraversion	0.01	0.03	0.06	0.06	0.02	−0.09	.14**	.14**	−.38**	1
11. Parent Openness	−0.04	−0.03	0.01	0.01	0.06	0.05	0.03	0.05	−.11*	.19**	1
12. Parent Agreeableness	−0.05	−0.04	0.02	0.10	−0.01	−0.13	0.05	0.07	−.43**	.21**	.13**	1
13. Parent Conscientiousness	−0.02	−0.09	0.03	0.06	−0.07	−.24*	.17**	.20**	−.48**	.24**	.18**	.41**	1
14. Child Negative Affect	0.04	−0.03	−0.02	−0.11	.15**	.22*	0.01	−0.04	.18**	0.01	0.05	−.21**	−.18**	1
15. Child Surgency	0.04	0.12	.14**	0.10	0.03	−.22*	0.06	.13**	−0.01	0.09	0.01	.21**	.15**	0.01	1
16. Child Effortful Control	.10*	0.09	.20**	.21*	−0.05	−0.05	0.04	0.01	−.25**	.18**	.18**	.25**	.31**	−.14**	.20**

Note. Time 1(T1) N = 464; Time 2 (T2) N = 90. Confidence intervals for the correlations are available in the supplemental materials, Table S1. Partial correlations controlling for child age show similar results (see supplemental materials, Table S2).

^*p<.05

^**p<.01.

Independent groups t-tests were also run to examine differences in variables of interest between educational motivation and joint context groups. These were largely non-significant (see supplemental materials, Table S3). Contrary to prior work, there were no significant correlations between any parent personality facet and child vocabulary (Kucker et al., 2021; see Table 3). No facet of parent personality was correlated to child digital media use either.

RQ 1: Does the duration of media use mediate the relation between parent personality or child temperament and child vocabulary?

RQ 1a examines duration of media use (average minutes/day) as a mediator of the relation between the five facets of parent personality (neuroticism, extraversion, openness, agreeableness, conscientiousness) and children’s total vocabulary size after controlling for parent education and child age (see Figure 1). Model fit based on the CFI was good (≥.95; Hu & Bentler, 1999), CFI = .957, and reasonable based on the RMSEA (.05–.08; Browne & Cudeck, 1992), RMSEA = .066. Contrary to prior work (Kucker et al., 2019), there were no significant direct effects of any facet of parent personality on vocabulary. However, parent conscientiousness had a significant direct link to duration of media use and a marginal indirect effect on vocabulary when mediated by duration of media use. Specifically, parents with higher conscientiousness tended to have children with lower media use, and less media use was associated with higher vocabularies. Parent neuroticism and openness also marginally predicted child media use with higher neuroticism associated with less media and higher openness associated with more media use. The indirect pathways to vocabulary from neuroticism and openness were both marginally significant. There was also a significant negative relationship between duration of media use and children’s vocabulary regardless of parent personality, consistent with prior work (Madigan et al., 2020; Jing et al., 2023; Kucker et al., 2024). See Figure 1 and Table 4.

Figure 1. Mediation Model of Parent Personality.

Indirect effects: Neuroticism - Duration of Media Use -Vocabulary: β=.018, p=.063^; Extraversion – Duration of Media Use - Vocabulary: β =−.002, ns; Openness – Duration of Media Use - Vocabulary: β =−.012, p=.071^; Agreeableness – Duration of Media Use - Vocabulary: β =.003, ns; Conscientiousness – Duration of Media Use - Vocabulary: β =.016, p=.056^

Note: Covariates represented in light gray boxes. Standardized betas shown. Unstandardized betas and all p-values are listed in Table 4.

^p<.10, *p<.05, ***p<.001.

Table 4.

Parent Personality Mediation Model, Predicting Vocabulary Outcome

Path	β	B	SE	z	95% CI	P
Neuroticism (N)
N-Duration of media use (a)	−.10	−.10	.05	−1.90	[−.21, −.00]	.057^
N-vocab (c’ direct)	.07	.07	.06	1.21	[−.04, .19]	.228
N-vocab (c indirect)	.02	.02	.01	1.86	[.00, .04]	.063^
Extraversion (E)
E-Duration of media use (a)	.01	.01	.05	.25	[−.08, .10]	.805
E-vocab (c’ direct)	.06	.06	.05	1.15	[−.04, .17]	.249
E-vocab (c indirect)	−.00	−.00	.01	−.25	[−.02, .01]	.806
Openness (O)
O-Duration of media use (a)	.07	.07	.04	1.83	[−.01, .14]	.068^
O-vocab (c’ direct)	.01	.01	.04	.17	[−.08, .09]	.867
O-vocab (c indirect)	−.01	−.01	.01	−1.81	[−.03, .00]	.071^
Agreeableness (A)
A-Duration of media use (a)	−.02	−.02	.04	−.44	[−.10, .07]	.659
A-vocab (c’ direct)	.04	.04	.05	.73	[−.06, .14]	.467
A-vocab (c indirect)	.00	.00	.01	.44	[−.01, .02]	.661
Conscientiousness (C)
C-Duration of media use (a)	−.09	−.09	.05	−1.99	[−.19, −.00]	.046*
C-vocab (c’ direct)	.03	.03	.05	.64	[−.07, .13]	.521
C-vocab (c indirect)	.02	.02	.01	1.91	[.00, .04]	.056^
Duration of media use-vocab (b)	−.18	−.18	.02	−9.55	[−.22, −.14]	<.001***
Parent Ed-Duration of media use (cov)	−.17	−.17	.05	−3.71	[−.26, −.08]	<.001***
Parent Ed-Vocab (cov)	.06	.06	.05	1.25	[−.03, .14]	.211
Child Age-Vocab (cov)	.43	.44	.04	10.45	[.35, .52]	<.001***

Note. R² (Mediator: Duration of media use) = 4.2%; R² (Dependent variable: Vocabulary) = 23.5%

^{^}p<.10

^*p<.05

^***p<.001.

RQ 1b examined a nearly identical mediation model but with the three domains of child temperament (negative affect, surgency, effortful control) as the initial predictors (see Figure 2). Model fit was good (Hu & Bentler, 1999); CFI = .99, RMSEA = .03. In this model, negative affect was directly associated with higher duration of media use but had no direct association with vocabulary. Surgency and effortful control were not predictive of duration of media use, but each positively predicted total vocabulary. There was a significant indirect effect of negative affect on vocabulary as mediated by duration of media use, suggesting duration of media use fully mediates the relation of children’s negative affect on their vocabulary: Children’s negative affect predicts higher media use, which in turn lowers vocabulary. The indirect pathways for surgency and effortful control were not significant. See Figure 2 and Table 5. Taken together, both mediation models 1a and 1b suggest that some elements of both parent personality and child temperament may play role in predicting children’s media use and vocabulary, with particular effects of parent conscientiousness and child negative affect.

Figure 2. Mediation Model of Child Temperament.

Indirect effects: Negative affect – Duration of Media Use - Vocabulary: β=−.027, p=.016*; Surgency – Duration of Media Use - Vocabulary: β =−.007, ns; Effortful control – Duration of Media Use - Vocabulary: β =.007, ns

Note. Covariates in light gray boxes. Standardized betas shown. Unstandardized betas and p-values are seen in Table 5.

*p<.05, **p<.10, ***p<.001.

Table 5.

Child Temperament Mediation Model, Predicting Vocabulary Outcome

Path	Β	B	SE	z	95% CI	p
Negative Affect (NA)
NA-Duration of Digital media use (a)	.15	.15	.05	3.00	[.05, .26]	.003**
NA-vocab (c’ direct)	.01	.01	.04	.14	[−.07, .09]	.890
NA-vocab (c indirect)	−.03	−.03	.01	−2.78	[−.05, −.01]	.005*
Surgency (S)
S-Duration of Digital media use (a)	.04	.04	.05	.76	[−.06, .14]	.447
S-vocab (c’ direct)	.10	.10	.04	2.57	[.02, .17]	.010**
S-vocab (c indirect)	−.01	−.01	.01	−.75	[−.03, .01]	.454
Effortful Control (EC)
EC-Duration of Digital media use (a)	−.04	−.04	.06	−.67	[−.14, .07]	.503
EC-vocab (c’ direct)	.14	.14	.04	3.30	[.06, .23]	.001**
EC-vocab (c indirect)	.01	.01	.01	.68	[−.01, .03]	.500
Duration of Digital media use-vocab (b)	−.18	−.18	.02	−9.61	[−.21, −.14]	<.001***
Parent Ed – Duration of Digital media use (cov)	−.16	−.16	.04	−3.78	[−.25, −.08]	<.001***
Parent Ed – Vocab (cov)	.05	.05	.04	1.12	[−.04, .14]	.265
Child Age – Vocab (cov)	.41	.42	.04	10.31	[.34, .50]	<.001***

Note. R² (Mediator, Digital media use) = 5.3%; R² (Dependent variable, Vocabulary) = 24.8%

^*p<.05

^**p<.01

^***p<.001.

RQ 2: Does using digital media for educational purposes or in a joint context moderate the relation between duration of digital media use and child vocabulary?

RQ 2 focused on possible moderators of the link between duration of digital media exposure and total vocabulary. The first model examined educational motivation (media used for educational purposes or not) as a moderator whereas the second model examined joint context (low vs. high levels of jointly using media). Parent education and child age were again included as covariates. Though duration of media use and child age significantly predicted children’s total vocabulary, there were no significant moderating effects – media use duration has a negative association with child vocabulary regardless of motivation or context of digital media use. See Table 6.

Table 6.

Moderation of Educational Motivation and Joint Context

Model	Coefficient	β	SE	t-value	p-value	Adjusted R2
Motivation					<.001	.210
	Duration of Digital media use	−39.43	10.11	−3.90	<.001***
	To Educate	−.49	10.01	−.05	.961
	Digital media use*Educate moderator	−12.30	10.07	−1.22	.223
	Child age (cov)	75.01	7.60	9.88	<.001
	Parent Ed (cov)	9.13	7.70	1.19	.236
Joint Context					<.001	.208
	Duration of Digital media use	−31.70	7.75	−4.09	<.001***
	Joint Context	−4.06	7.59	−.54	.593
	Digital media use*Context moderator	−5.00	7.62	−.66	.512
	Child age (cov)	76.64	7.67	9.99	<.001***
	Parent Ed (cov)	8.89	7.67	1.16	.247

Note. (cov) indicated covariate.

^***p<.001.

RQ 3: Are there moderated mediation effects between parent personality, child temperament, digital media use, and vocabulary?

The final aim combined the prior two RQs into a single moderated mediation model in which either parent personality or child temperament predicted duration of digital media use, and duration of digital media use predicted total vocabulary. The link between duration digital media use and vocabulary was predicted to be moderated by educational motivation (models 3a and 4a) or joint context (models 3b and 4b), leading to a total of four models – parent personality facets as predictors with educational motivation as the moderator between media use and vocabulary (model 3a), parent personality facets as the predictors with joint context as the moderator, (model 3b), child temperament domains as the predictors with educational motivation moderator (model 3c, and child temperament domains as predictors with joint context as moderator (model 3d). Each of the analyses were pre-registered so were all completed as planned, but each of the four models were a poor fit for the data so results should be taken with some caution; CFI’s range = .11– .29; RMSEA’s range = .33 – .48 (see supplemental materials).

In the first two models exploring parent personality we find nearly identical results as were found in RQ 1. There was a significant direct effect of duration of media use on vocabulary regardless of parent personality, as well as a significant direct effect of conscientiousness and neuroticism on duration of media use. Openness had a marginal direct effect on media use. There were no significant indirect effects. There was also no significant moderated mediation for educational motivation use (Figure 3, panel A) nor joint context (Figure 3, panel B).

Figure 3. Mediated Moderation for Parent Personality.

Indirect effects: Neuroticism – Duration of Media Use - Vocabulary: β=−.028, ns; Extraversion – Duration of Media Use - Vocabulary: β =.002, ns; Openness – Duration of Media Use - Vocabulary: β =.014, ns; Agreeableness – Duration of Media Use - Vocabulary: β =−.004, ns; Conscientiousness – Duration of Media Use - Vocabulary: β =−.024, ns

Indirect effects: Neuroticism – Duration of Media Use - Vocabulary: β=−.013, ns; Extraversion – Duration of Media Use - Vocabulary: β =.001, ns; Openness – Duration of Media Use - Vocabulary: β =.006, ns; Agreeableness – Duration of Media Use - Vocabulary: β =−.002, ns; Conscientiousness – Duration of Media Use - Vocabulary: β =−.011, ns

Note. Moderator of educational motivation (Panel A) or joint context (Panel B) in models with parent personality as the predictor. Covariates in light gray boxes. Standardized betas reported. See supplemental materials, Table S5 for full results.

^p<.10, *p<.05, **p<.01, ***p<.001.

A third and fourth models explored moderated mediation for child temperament domains (Figure 4, Panel A and B). Again, the results here mirror that of RQ 1 with no significant moderated mediation, but a direct significant negative association between duration of media use and total vocabulary, a positive association between negative affect and media use, a positive association between surgency and vocabulary, and a positive association of effortful control and vocabulary. There were no significant indirect effects or moderations.

Figure 4. Mediated Moderation for Child Temperament.

Indirect effects: Negative affect – Duration of Media Use - Vocabulary: β=.038, ns; Surgency – Duration of Media Use - Vocabulary: β =.005, ns; Effortful control – Duration of Media Use -Vocabulary: β =−.008, ns.

Indirect effects: Negative affect – Duration of Media Use - Vocabulary: β=.014, ns; Surgency – Duration of Media Use - Vocabulary: β =.002, ns; Effortful control – Duration of Media Use - Vocabulary: β =−.003, ns.

Note. Moderation by educational motivation (Panel A) and social context (Panel B). Covariates in light gray boxes. Standardized betas shown. See supplemental materials Table S6 and S7 for full results.

*p<.05, **p<.01, ***p<.001.

Discussion

The overall aims of Study 1 were to examine concurrent relations between parent personality, child temperament, media use (duration, motivations, and contexts), and children’s vocabulary. We found that parent conscientiousness, (and to some extent, neuroticism and openness) predicted children’s media time and had weak, but marginally significant, indirect links with child vocabulary even after controlling for parent education and child age. That is, digital media use duration was a significant mediator of the effect of some aspects of parent personality on children’s vocabulary (Figure 1), most notably parent conscientiousness. These results ought to be taken in context, however, given that, contrary to prior work (Kucker et al., 2021), there were no significant direct correlations between parent personality and vocabulary and the effects, while significant, are relatively small.

There was stronger evidence for the role of the child temperament on children’s digital media use and vocabulary. Consistent with prior work, we find direct influences of both surgency and effortful control on child vocabulary (Balázs et al., 2024; Kucker et al., 2021) as well as demonstrating that higher negative affect was associated with higher media use (Baruchli et al., 2024; Shin et al., 2021). The current work builds on this by also finding a significant mediation of digital media duration on the link between negative affect and vocabulary (Figure 2). While the findings indicate that parent personality may play a role in children’s media use and vocabulary, and a child’s own temperament may be particularly relevant to consider when assessing children’s media use and impacts on vocabulary during the first few years of life. These results, however, are limited in that they assess only cross-sectional comparisons – do current traits and current digital media use correlate with current vocabulary outcomes? Though Study 1 examined a range of different ages, a crucial question for understanding development is change over time. Thus, Study 2 examines vocabulary in a small subset of these children one-year later as predicted by their earlier traits and media use.

Study 2

Study 1 investigated concurrent associations between parent personality, child temperament, duration of digital media use, and total vocabulary. It is likely, however, that the effects of media may be more cumulative over time, leading to bigger impacts as children age. In Study 2, a series of exploratory analyses examined possible longitudinal effects of parent personality, child temperament, and media use around 2-years-old on children’s vocabulary one year later around 3-years-old. This analysis was pre-registered, but analyses were considered exploratory given the small sample.

Methods

Participants

All participants from the first wave of Study 1 were contacted via Cloud Research to complete a follow-up study approximately 1 year later. Of those, 90 parents (19% of Study 1) completed the follow-up around 12-months after their initial visit (M = 12;24, SD = 1;19, range = 9;12 – 17;6). Data here were again cleaned according to Chmielewski and Kucker (2020) as well as checked for consistency in reporting from time 1 to time 2. 11 additional participants were dropped for failing these validity checks. The time 2 sample was nearly identical to the time 1 sample in demographics (see Tables 1 and 2). Change in media time from Time 1 to Time 2 was only marginally significant t(87) = 1.78, p = .078.

Procedure

A nearly identical set of assessments as used in Study 1 were given during the Study 2 follow-up, also via Qualtrics and participants were paid $5. The primary difference was that children at time 2 were older and thus were given the MCDI-III to capture expressive vocabulary (Fenson et al., 2007) which is validated for children 30–37-months-old. The MCDI-III is a checklist of 100 more difficult words appropriate for this age. The ECBQ (child temperament) and BFI (parent personality) were not examined again to minimize length of the survey and because child temperament and parent personality are considered relatively stable and thus, not predicted to have changed.

Analysis

Study 2 includes similar aims to Study 1 but predicts children’s vocabulary one-year later instead of concurrently. Thus, the outcome variable is total vocabulary size (centered) based on the MCDI-III and the covariate of age was time 2 age, but other predictors, mediators, and moderators were the same. The same set of RQs and corresponding analyses were run. Similar findings were expected at the one-year follow-up.

Results

Identical models to Study 1 were run here but with total vocabulary at time 2 (as reported on the MCDI-III) as the outcome variable. Parent education at time 1 and child age at time 2 (centered) were again included as covariates. The mediation analyses for Aim 1 examined the relation between parent personality and child temperament on vocabulary as mediated by digital media time. The effects of personality on language on year later as mediated by media use was largely non-significant (see supplemental materials). A nearly identical model testing the mediation of media use on children’s temperament and vocabulary one year later found marginally significant indirect effects of negative affect and effortful control on children’s time 2 vocabulary as mediated by their time 1 duration of media use. That is, children who are higher in negative affect or lower in effortful control used more media during the first 1–2 years of life, and this in turn lowered vocabulary one-year later. Thus, digital media use has both cross-sectional and longitudinal impacts on the link between some aspects of temperament and children’s vocabulary. Moreover, the duration of media time at time 1 significant predicted vocabulary at time 2 regardless of parent personality or child temperament, replicating the finding that there are more general direct longitudinal effects of media use on children’s language.

The models of RQ 2 (time 2) tested if there was a moderation of the association between media use and vocabulary one year later by educational motivation and/or joint context. These were non-significant (see supplemental materials) with only duration of media exposure at time 1 and child age at time 2 predicting time 2 total vocabulary.

RQ 3 (time 2) tested the moderated mediation models predicting vocabulary one-year-later. There were no significant effects for parent personality or child temperament in RQ 3 except that negative affect continued to predict duration of media time. Again, digital media use duration at time 1 significantly predicted vocabulary at time 2 regardless of parent personality or child temperament.

Discussion

Study 2 examined longitudinal effects of individual traits and digital media use on child vocabulary. While there was less support for parent personality’s longitudinal relation between media and vocabulary, and only marginally significant effects of child temperament on vocabulary at time 2, there were consistent negative associations of earlier digital media use with later vocabulary. More specifically, links between both negative affect and effortful control and children’s vocabulary one year later were marginally mediated by the duration of children’s digital media use, and children who used more media when they were 17–30-months were significantly more likely to have lower vocabulary scores one-year later even after controlling for parent education and child age. Although exploratory due to small sample size, the results confirm the correlation between digital media use duration and vocabulary, even one year later, and reinforce that children’s temperament may play a key role shifting the pathways by which digital media impacts children’s vocabulary.

General Discussion

Overall, we find some evidence for both parent personality and child temperament in shaping children’s digital media use and vocabulary acquisition as well as longitudinal effects of media use on vocabulary a year later. First, there were significant association between some elements of parent personality and children’s duration of media use concurrently. Children of low conscientiousness parents (and potentially low neuroticism and/or high openness) are more likely to use digital media. Moreover, there were marginally significant mediations of digital media use on vocabulary such that these children were also more likely to have a lower vocabulary as a result. Second, we find that a child’s level of negative affect is associated with children’s digital media use which in turn predicts vocabulary size. Consistent with prior work that found higher negative affect was associated with higher screen time (Brauchli et al., 2024), the current study also found that this is associated with lower vocabulary, even predicting vocabulary outcomes up to one year later. The results confirm the important role both the parent and the child play in children’s media use and how such use may impact vocabulary.

Contrary to prior work, there were no significant moderators of educational motivations for use or joint use on vocabulary outcomes. In prior work, when educational motivation or joint social contexts were considered, the potentially harmful effect of digital media on high negative affect children’s vocabulary were buffered (Sanders et al., 2019; Madigan et al., 2020; Jing et al., 2023). In the study here, however, both educational motivation and joint context were measured dichotomously as simply present/not or low/high, and, as further discussed below, it is likely these variables are much more complex constructs which possibly contributed to this null effect.

Taken together, these findings point to the importance of a holistic approach to children’s digital media consumption, considering not just the quantity of media use but also who is using the media. In particular, the results stress the importance of considering how the broader digital ecology affects child development beyond just limiting screen time. A recent framework, the DREAMER model (Barr et al., 2024), provides a useful lens for understanding this complexity. The framework emphasizes the importance of individual child and parent factors, including child temperament and parent personality. It also emphasizes the dynamic and relational use of digital media within the family ecology. In particular, during early childhood parents make decisions throughout the day regarding whether or not to use digital media to regulate the child’s emotions, to occupy them while they get other things done, to educate them, or to connect with their child (Barr et al., 2024; Nikken, 2019; Suh et al., 2024). In the present study, individual parent personality and child temperament factors played a role in both child screen time and child vocabulary both concurrently and to some extent longitudinally. Understanding these broader dynamics can guide interventions that go beyond a one size fits all approach of limiting screen time to creating more personalized approaches to build healthy media habits and positive media environments within individual families.

Implications

This study highlights the importance of considering individual differences in both parent personality and child temperament when developing recommendations for digital media use. Caregivers who are less conscientious may want to be more mindful of media time with their child. Children with higher negative affect may need more guidance and supervision in how they interact with digital media to ensure that it benefits their development. For example, digital media exposure may need to be more structured, or of consistently higher quality or complemented by other activities to mitigate potential negative effects on language learning.

It is also important to keep in mind that not all digital media exposure is inherently harmful or may not be harmful for all children. Some children (those with caregivers who are higher in conscientiousness or children with lower negative affect for instance) may also have less opportunities to be exposed to media and thus, there might be fewer concerns about media’s impact on their development. But, for those who are more likely to get exposed to media, however, it may be pertinent to find ways to offset or buffer the effects on their language. One potential avenue may be in educational media which can act as a buffer, supporting vocabulary development. Programs like Sesame Street, which are designed to be both educational and engaging, may offer considerable benefits over less intentionally designed content that does systematically and age-appropriately include language-promoting strategies (see Linebarger et al., 2017). When digital media use is motivated by educational goals and supported by joint, active engagement (e.g., asking questions, interacting with the content), prior work suggests it can have a more positive impact on children’s vocabulary development (Jing et al., 2023). However, little prior work has incorporated both how and why media is used and individual differences across children and parents.

Limitations

Contrary to the hypothesis, the current study found no significant moderation of educational motivations on children’s language outcomes. Though the current study used a similar approach to prior work to measure educational use – simply asking caregivers if they use media for educational purposes or not (e.g. Madigan et al., 2020) - this measure is also likely very coarse and unable to capture the nuances of digital media today. What a caregiver thinks might be educational or what they intend to use the media for may not match what learning effect (if any) the media itself actually has. The idea here is that caregivers who are using media for educational purposes may also be using specific types of media or more likely to engage with the child during its use. Nonetheless, dissecting intended use from actual use of media is difficult in the current study. Moreover, there is a recurrent challenge of defining “educational” media experiences as there are very few national guidelines for doing so (Meyers et al., 2021). In the present study, a significant limitation is that we did not independently code the quality of the media content. Researchers have identified content that promotes language outcomes for children in the age range (e.g., Vaala et al., 2010). However, the current digital media environment is now vast and so content is more difficult to quantify. Future studies do need to consider systematic ways to capture media content quality. Distinguishing between media content and motivations for use are important in understanding the true impact of digital media on children’s vocabulary development. Further research is needed to better understand these mechanisms.

Moreover, further work is needed to expand the measurement and sample. Even though the results are consistent with prior research, the current study is limited to online parent-reported data coupled with a relatively small sample of longitudinal data. Prior work has found parent-reported personality, temperament, and vocabulary to be reliable, but work on measuring media use via parent report is still in development (see e.g. Barr et al., 2020, 2024; Suh et al., 2024) and is one avenue for future work. This data was also collected shortly after the height of the COVID-19 pandemic which could have led to changes in children’s vocabulary (Pejovic et al., 2024; Zuniga-Montanez et al., 2024) and media use (Hedderson et al., 2023), both of which may have been differentially impacted by parent personality during the pandemic. In addition, the current examination focused on children’s exposure to videos/TV as it represents over 80% of children’s screen time at 2-years-old (Kucker et al., 2024). This fails to capture the nuances in media use by children, however, especially in time 2 as children get older, more independent, and increase their use of other media sources such as apps and games, which may account for the marginal decrease in total video time at time 2. Thus, future work and recommendations ought to continue to expand the applications with the entire media ecology (Barr & Kirkorian, 2023).

Future Directions

Though the current work begins to integrate individual differences in our models of development in a digital ecology, many more pathways are left open for future work – e.g. socioeconomic status factors, media content, media practices and beliefs as well as cultural differences in media use. Consistent with prior research (Dynia et al., 2021), we also find that parent education and income predict media use. Likewise, parents with positive views of digital media report higher rates of educational and joint use of media with their children (Griffith, 2023), and parent beliefs may be tied to their personality.

Little prior work has examined parent personality traits as they relate to children’s digital media use and language development and thus, the current study takes a critical first step toward filling this gap and providing a more comprehensive picture of the digital media ecology. One step will be to expand measurement and tools to further capture children’s real-time digital media ecology, such as integrating home recording data such as LENA (Language ENvironment Analysis) software and passive sensing apps. Both allow for tracking parents’ and children’s digital media use and child language outcomes. For instance, Sundqvist and colleagues reported that conversational turns between parents and their 2- and 5-year-old children were also a significant buffer in moderating the effects of digital media time on language outcomes (Sundqvist et al., 2021, 2022, submitted). Expanding the developmental outcomes beyond vocabulary will also be relevant to examine how digital media ecology intersects with other aspects of language and cognitive development. Most importantly, the present findings suggest the need for rigorous testing of interventions that go beyond simple screentime recommendations.

Conclusion

As this special issue makes clear, the broader digital media ecology in which children are developing is a critical, new, yet understudied frontier. Digital media use is complex both in how and when it is used as well as by whom, but few current recommendations take such nuances into account. The current study is just a first step toward understanding the pathways by which individual differences and motivations for media use also play a role. As digital media use becomes more prevalent and children are exposed at younger ages, more research is needed to define the parameters of the family digital media ecology and how it intersects with children’s developmental pathways.

Public Significance Statement.

As digital media use by young children rises, it becomes increasingly critical to consider the broader family and digital ecology in which media is used. This study examines how parent personality and child temperament, as well as why and how media is used, influences children’s digital media use and vocabulary. The findings have potential to tailor guidelines to create individualized developmental outcomes for children in an increasingly digital world.