Affect and behavior in preschoolers during a tablet-based assessment: insights from observational coding of BELLA use

Background. Despite widespread tablet use in early childhood education, children's affective and behavioral experiences with digital assessments remain understudied. This study addresses gaps in understanding how preschoolers emotionally and behaviorally respond to a tablet-based learning tool. Methods. We employed a mixed-method approach analyzing observational data from 161 children (ages 3-6) across three Houston preschools using the Bilingual English Language Learner Assessment (BELLA). Behavioral coding adapted from SPAFF and DPICS frameworks examined affect, engagement, digital skill, and social dynamics across sessions. Statistical analyses included descriptive analysis, McNemar tests for session comparisons, and multiple linear regression. Results. During first exposure, engaged affect was most prominent (40.6%), with 51.6% demonstrating good digital skills. Between sessions, engaged affect significantly increased (χ² = 10.42, p < .01) while detached affect decreased (χ² = 4.02, p <.05). Digital proficiency improved across all levels. Multiple linear regression (R² = .497, p < .001) revealed positive affect and digital skill as significant performance predictors, with age as the strongest covariate (β = .357, p < .001). Discussion. Digital assessments capture both cognitive competency and real-time indicators of self-regulation and engagement. Findings highlight digital literacy's critical importance in early childhood education and demonstrate that emotional engagement significantly contributes to learning outcomes in technology-mediated environments. Limitations/Future Directions. The study was limited by observational data and suburban private school sampling. Future research should incorporate structured protocols, video coding, time-stamped data collection, and more diverse samples to enhance generalizability and capture dynamic learning processes.

Introduction

This study investigates the use of tablet devices (TD) in classrooms and the TD-child interactions focusing on co-use recommendations, child-caregiver relationship, and child behavior. It relies on the Bilingual English Language Learner Assessment (BELLA) application, used as an English only kindergarten readiness assessment. Prior work with BELLA validated its psychometric properties and domain coverage (Kilani et al., 2024; Tan et al., 2023), but the affective and behavioral experience of children during digital play remains unexamined – particularly the evolution of experiences over repeated sessions and exposures. To address this gap, the study adopts a mixed method exploratory convergent approach to a secondary dataset using qualitative observations of children engaged with BELLA and their BELLA performance scores. Three research questions motivate and guide the study: (1) What affect and behavior are observed when children first interact with BELLA; (2) Do we observe significant changes in the affect and behavior of the children between the first and second sessions; and (3) How might children’s affect and behavior inform our understanding of performance scores in BELLA? In other words, to what extent do observed affect and behavior predict children’s performance on BELLA tasks, and what patterns might this reveal about the role of emotional and digital readiness in assessment settings? The following literature review contextualizes the study and addresses the rise of tablet use in childhood, learning in early childhood, affect and behavior relating to a child’s experience, identified research gaps, and a short overview of BELLA and other early assessments.

Rise of Tablet Use in Childhood

Across the world, children are increasingly surrounded by TD and other screens (e.g., televisions, computers) and mobile devices (e.g., cellphones). In fact, 98% of children aged 8 years or younger have mobile devices at home, with the majority being smartphones, followed by tablets in the home, and a child’s own personal tablet. In 2011, fewer than 1% of children in this age group owned their own tablet, compared to 42% in 2017. Tablets in the home stood at 78% in 2017 (Rideout, 2021). The same report from Common Sense – a nonprofit organization focusing on building a more healthy, equitable, and empowering future for all kids in the digital age – reveals that not only are there more devices in homes, but the time spent using mobile devices has dramatically increased going from 5 minutes a day in 2011, to 15 minutes a day in 2013, and 48 minutes a day in 2017 (Rideout, 2021). These numbers are reflected in other studies (Kabali et al., 2015), where more than 96% of children under 4 years old had used a mobile device, with many starting before the age of 1, and most having used one by the age of 2. These patterns are stronger and more salient in low-income families (Kabali et al., 2015; Rideout, 2021). For example, a seven-country study of European nations reported that tablets were children’s favorite device, thanks to the appeal of the touchscreens and portability—even within a family where a child does not own one (Chaudron et al., 2015). These numbers clearly demonstrate that screen devices are an integral part of the daily lives of many children and their families. The trend is evidently increasing, with no sign of stopping in sight. Parents are the primary driving force behind the trend, finding a use for TD in caregiving.

It is currently a routine practice to utilize these screen-based portable devices as part of caregiving (TD in particular), as pacifiers, engagers, or distractors. This is despite the mixed attitudes towards screen time for children communicated by child-focused professional organizations, public figures, and parent advocates. A recent study (Wartella et al., 2013) demonstrated that parents, while concerned over their children’s media use, often rely on these tools to soothe their child, calm them, or aid them for bedtime. In other words, portable devices are used by parents as a way to manage everyday routines with children, even though these same parents complain about screen ‘over-usage.’ It is natural to rely on new technologies to help with daily life; however, this over-usage can have a direct impact on children. In short, screen-based media is an almost unavoidable component of modern parenting, and it is likely to remain and expand its presence.

The pervasiveness of screens and parents’ reliance on them in early childhood directly goes against the recommendations of the American Academy of Pediatrics (AAP). The recommendations warn against screen time for children under the age of two. This is because, according to the AAP, the early months are critical for cognitive, language, motor, and social development, requiring environmental input that children cannot gain from screens or TD as they would from a caregiver or interacting with their surroundings.

Moreover, it is argued (Hill et al., 2016) that children have difficulty transferring knowledge gained from screens into their three-dimensional experience. It is advised that parents use screens with their young children (starting at 15 months of age and older) and ensure they reteach the content. Moreover, 24-month-old children can learn new words from live-video chatting with a responsive adult. This happens as early as 15 months old; however, learning new words during live-chat only occurred when using specially designed applications that are not commercially available (Hill et al., 2016). Ultimately, co-viewing is the primary recommendation for older children (2 years old and above), with a statement that the content needs to be age-appropriate and of high quality. In particular, preschool-aged children (3 to 5 years old) can improve their cognitive, literacy, and social outcomes when presented with very well-designed television programs (e.g., Sesame Street), as per the AAP policy statement (Hill et al., 2016). Young children engage with TD independently and regularly, particularly in low-income households (Kabali et al., 2015; Rideout, 2021). Thus, the AAP guidelines do not align with reality; they are not widely known to the public, as only about 20% of parents are familiar with them (Rideout, 2021).

Early childhood screen use is no longer exceptional; it has become the norm. Debates regarding appropriate limitations for young children continue to occur; however, there has been insufficient investigation into both their detrimental and positive impacts. The focus is on caregiving, as it should be. However, it also needs to focus on children, particularly on their experiences with screens. According to sociocultural theory (Taber, 2025; Vygotsky, 1978, 1994a, 1994b), young children development is mediated through interactions with cultural tools: objects and symbols such as language, routines, and today, digital devices. Tools like TD are not passive; they are part of the cultural environment and actively shape how children learn, communicate, experience and resolve challenges, especially when used in isolation, or without adult scaffolding. These devices can become a child’s primary mediator when caregivers are absent, which further stresses the importance of understanding a child’s emotional and behavioral experience during TD use. Self-regulation frameworks emphasize the deep ties of early learning to the development of emotional and behavioral control. Zimmerman (2002), and Kubsch and colleagues (2025) describe how self-regulated learning involves a child’s ability to plan, monitor, and adjust cognitive and emotional processes during a task. These skills are just emerging in preschool years. Expressed emotions, such as frustration, curiosity, disengagement, or delight, are central indicators of a child’s regulatory strategies and readiness to learn. Kopp (1982) offers a complementary developmental framework by tracing the ontogeny of self-regulation across early childhood. She outlines the progression from external control (e.g., compliance with caregiver directives) to self-control and eventually self-regulation. These regulatory capacities emerge in stages, beginning with basic inhibition and culminating in the child’s ability to self-monitor, delay gratification, and adjust behavior in the absence of adult guidance. Even though TD can act as a mediator, a caregiver presence is critical to mediate this progression of self-control and self-regulation.

Together, these perspectives underscore the importance of observing and interpreting a child’s engagement with screen-based technology as part of their broader developmental trajectory, particularly in an educational setting where they will spend the majority of their time. As TD are being implemented in schools, they need to be studied within an educational context across all ages, with particular attention to the often-overlooked pre-kindergarten age bracket (up to five years old).

Early Childhood and Learning

Early childhood is a critical time for cognitive, social, and emotional development. Technology and screen-based portable media are ever-present in early childhood and have created a new context for development, shaping the contemporary learning environment. As sociocultural theory (Taber, 2025) suggests, digital tools should be viewed as cultural artifacts that participate in and transform developmental processes. When introduced early, they shape and organize the structure and pacing of interaction, learning, and meaning-making.

Digital literacy – the ability to effectively use digital devices and navigate interactive content – emerges as an increasingly recognized essential skill that needs to be carefully investigated. In fact, the increasing presence of screen-based technologies at home and in classrooms creates a stronger need to nurture this skill in schools (McManis & Gunnewig, 2012; Neumann, 2016, 2018). In support of that, interactive applications and digital tools can offer a unique opportunity for multimodal learning that strengthens the emergence of literacy, numeracy, and other academic domains, as well as executive functioning. Executive functioning refers to a set of cognitive skills that include working memory, inhibitory control, and cognitive flexibility. These skills support goal-directed behavior and adaptive problem-solving (Kolloff et al., 2025). They begin developing in early childhood and are crucial for self-regulation and academic success.

Digital tools also provide scaffolding and feedback loops that, when designed developmentally, can support children’s agency and learning progressions (Plass et al., 2015; Plass & Kaplan, 2016). They enable personalized learning experiences that can be tailored to individual learning needs at their specific developmental stage (Huber et al., 2018; Zaranis, 2016). In fact, Strouse and Ganea (2017) revealed that electronic books improve vocabulary retention more than their paper counterparts by being more interactive and maintaining interest in reading.

Learning in childhood is complex and goes beyond cognitive dimensions and knowledge. There are affective and behavioral experiences that are to be considered. Cognitive and emotional engagement is foundational for meaningful learning outcomes in technology-mediated settings. Emotions such as curiosity, confusion, and enjoyment are not just temporary states. ,they are active contributors to the deep processing and integration of information children experience during learning (Halverson & Graham, 2019). Knowing this, researchers must account for the interplay of skill development, motivational states, and socioemotional functioning to fully grasp the impact of TDs on early learning.

Affect, Behavior, and Child Experience: Their Role In Children’s Learning

Emotional states, particularly at very young ages (0 to 5 years old), contribute to a child’s learning experience and learning outcomes. Emotions such as enjoyment, frustration, engagement, and distraction can reflect underlying cognitive and motivational processes that may not be immediately visible. For example, mind-wandering or distractibility, combined with concentration, permits the emergence or the incubation of creativity (Khalaf et al., 2022). The flow of emotions, which refers to the sequence and transition of affective states during learning, is grounded in dynamic systems and emotion regulation theories. It emphasizes that emotions unfold over time and influence subsequent cognitive processing and behavioral responses (Camras, 2011; Pekrun & Linnenbrink-Garcia, 2014). Emotions inform the learning process and can provide overlooked indicators of learning to teachers. While typically negative emotional states such as disengagement or frustration may adversely influence a child’s motivation and overall learning efficacy (Huber et al., 2018; Madigan et al., 2019), as part of a flow of emotions, negative emotions can be part of learning if followed by positive affect (Zhukova et al., 2020).

Additionally, as the guidelines (AAP) suggest, while supported by observational research, social dimensions are critical. Co-use is a cornerstone of TD use among young children and relies upon caregiver-child interactions. The caregiver has a role in affecting children’s emotions, learning behaviors, and educational experiences (Barr, 2019; Strouse & Ganea, 2017). Emotional exchanges during co-use may help regulate frustration or recover from it, sustain engagement and persistence, or interpret content more meaningfully. Thus, carefully designed observational approaches can offer a valuable lens for analyzing the characteristics of caregiver-child interactions and their interplay with affective and behavioral learning indicators, thereby refining educational strategies.

Research Gaps

Given the increase in screen or TD-based exposure and learning and the developmental environment they create, significant research gaps remain. Our understanding of young children’s emotional, behavioral and social experiences using tablet-based education is incomplete. For example, two systematic reviews highlight the extreme lack of qualitative observational research that is focused on these dimensions (Griffith et al., 2020; Herodotou, 2018). Specifically, Griffith (2020) identifies uncertainties on the impact of interactive applications on social-emotional development, despite the evident cognitive and academic benefits. Social and emotional development, in fact, is often overlooked, particularly with pre-kindergartners. Older reviews, such as Herodotou’s (2018), underline the challenges in understanding the learning and developmental impacts of tablet use.Blumberg’s (2019) review of recommendations and policies on digital games as a context for children’s cognitive development emphasize the need to investigate digital games and applications, both for cognitive and socio-emotional impact in younger populations.

Additionally, despite benefits such as emergent literacy (Neumann, 2016, 2018) and executive functioning (Huber et al., 2018), research seldom considers external influences like digital proficiency and environmental disruptions, or the specific context in which digital use occurs (Chiong & Shuler, 2010; Madigan et al., 2019). McPake and colleagues (2013) highlight the lack of research into how informal digital/screen media experiences (at home, or outside the educational setting) translate into structured educational spaces. Applications such as BELLA offer an opportunity to address these gaps by providing structured, context-rich data on children’s emotional, behavioral, and social engagement during screen-based learning.

Digital assessment in early assessment: BELLA

BELLA is a tablet-based app designed for preschoolers which serves as a school readiness assessment. Specifically designed for children between the ages of 3 to 5 years old, BELLA is both an assessment and a learning tool. It addresses skill development in major domains including early numeracy, early literacy, early science, and social-emotional development. Thus, BELLA covers critical domains for early education, as well as the social-emotional element that is often overlooked, and cognitive skills. These cognitive skills, precursors to higher-order cognitive competencies, include analytical, creative, and practical. More on the cognitive skills, items, and performance in the assessment is available in the associated publications (Kilani et al., 2024; Tan et al., 2023).

BELLA has a library of more than 700 items, which arms preschools with a variety of engaging questions that can be used throughout the years. Users can select a set of items that focus on the learning materials relevant to the student at a given time. The sessions can be repeated, which allows teachers to gauge performance and learning over time. BELLA can be a supportive tool for learning, providing activities that teachers can rely on after conducting their lesson plan to reinforce the material covered.

BELLA is one of its kind as a kindergarten entry assessment (KEA). It distinguishes itself from other assessments with its unique integration of socio-emotional assessment alongside the assessment of literacy, numeracy, and science. BELLA is designed to adapt item difficulty to the individual child’s performance, and its interface is child-friendly and interactive. Its competitors - commonly used KEAs – in the United States include tools such as Teaching Strategies GOLD (TS-GOLD), Brigance Early Childhood Screens III, and the Standardized Test for Assessment of Reading-Early Literacy Enterprise (STAR-EL).

TS-GOLD is used as a formative, observation-based assessment for children from infancy to 3rd grade. It focuses on evaluating skills through ongoing teacher observations within everyday activities, requiring teacher input. It covers 11 domains but fails at capturing dynamic and spontaneous child responses. TS-GOLD is comprehensive in its assessment of socio-emotional skills, however, the main weakness remains that the teacher must observe and document behaviors during regular classroom days, which can be overwhelming and prone to human error - affecting accuracy and consistency (Teaching Strategies, 2022). Brigance is teacher-led and includes a set of developmental screeners from infancy through first grade. These are short, and performance-based tasks that target academic, cognitive, language, and physical skills. Brigance is ideal for its valuable insights into developmental milestones, however, it does lack a digital or adaptive element, which makes it less suitable for schools relying on interactive, screen-based assessments (French, 2013; Yun et al., 2021).

STAR-EL is a digital competitor of BELLA. It is adaptive and targets early literacy and numeracy. It is designed for preschool through grade 3. Similar to BELLA, it is intended to be taken independently on a computer or a tablet. The application primarily focuses on traditional academic skills and does not specifically assess broader cognitive domains or socio-emotional skills. STAR-EL fails to fully leverage DM interactivity to support deeper engagement (Shapiro & Gebhardt, 2012).

These tools are compared with BELLA by Tan and colleagues (2023), who distinguish BELLA from other KEA programs by its combination of academic content, cognitive skill assessment and social-emotional evaluation. BELLA is not teacher administered, although it is teacher informed; and directly engages children through interactive and adaptive tasks to maximize engagement and minimize observational biases. Compared to STAR-EL (DM-based KEA), BELLA specifically focuses on preschoolers and takes a holistic (a comprehensive range of domains evaluated) approach to which academic, developmental, and cognitive domains it covers. This puts BELLA ahead of currently used KEAs, and serves as a tool for further research in pre-school TD use. It is a practical format, particularly valuable for studying child-TD interactions in an educational setting.

Materials and methods

Participants

The data presented in the study were collected from students attending three private preschools in the greater Houston, Texas, in 2022. Eligible participants were children in pre-kindergarten classrooms, comprising 161 children aged 3 to 6 years old, 74 (45.96%) girls. Table 1 shows stratification by sex and age. Racial/ethnic demographics were not collected. The participating schools collected parental consent for the participating classrooms. The participants were part of a larger study, the results of which have been published in Tan and colleagues (2023) and Kilani and colleagues (2024). The analyses presented below have not been featured in any other context.

Table 1/Таблица 1

Stratification of the Sample by Sex and Age

Стратификация выборки по полу и возрасту

Procedures

Setting

The children were removed from their classrooms and taken to a quiet room.  No more than two children were tested at a time. Disruptions were recorded, and included events such as occasional traffic into the room, hallway noise, or sharing the room with a speech therapist.

Data Collection

The data were collected by a trained observer who administered BELLA to children in English. Each child was given a digital tablet and had a profile created in the application. When ready, children were asked to complete a pre-assigned pilot path (a set of 33 items, including six literacy, six math, 18 science, and three social/emotional items); each path took about 20 minutes to complete. The goal was to accomplish a total of 2 pilot paths. No more than one pilot path could be completed within a day. During the evaluation, the tester provided minimal guidance to the children to encourage their independent use of BELLA. Children were allowed to interrupt the session at any point. Sessions were either paused (e.g., for bathroom breaks) and resumed when the child returned or paused and saved (e.g., when the child could not or would not continue) and resumed on a different occasion. Children could not have more than three sessions. They returned to their classrooms at the end of each session. All sessions took place in the morning.

Observer/Tester Assignment and Positionality

The tester was a female psychology undergraduate student. They were in charge of the primary data collection of the study, working with the participating schools in Texas, and using her robust training in administering BELLA. They were middle-aged, and had experience with caregiving as a mother. Their personality shone through a few of the observations where they occasionally strayed from their formal note-taking style. For example, the following sentence shows subjective commentary in the use of the word ‘Cute’:

“Child needs a lot of assistance. Asks my permission before answering every question. It is very cute the way he expresses delight at some of the pictures.”

Regarding social dynamics, it was not always possible to remain consistent with full attention, such as in this example:

“People keep coming in and out of this room. It's somewhat distracting. The child seems focused, but it is distracting me from paying attention to him. Child has trouble with dragging questions. He appears to know what he is supposed to do; he seems unable to physically do it. Child mostly refuses to answer questions in which he does not know the answer. He lets the time run out and ignores my encouragement to take a guess. This room is like a weigh station! At least 20 interruptions in 30 minutes!”

Nevertheless, the tester continued with their task and tried to keep engaging while maintaining good and truthful record of the session.

The tester acted as the observer; they are referred to as the observer from this point on. They had been trained in behavioral observation and behavioral coding in a prior research project. They independently took the initiative to record written observations for each session they conducted with the children. These handwritten observations primarily focused on the child's experience with BELLA, recording their affect, social behaviors, distractibility, digital tablet skills, and performance. Although not formally structured, observations were guided by a previous two-month-long training received by the observer for a prior project involving behavioral coding of a case study of a 28-month-old using an iPad (Khalaf et al., 2022; Zhukova et al., 2020). The recorded observations are open-ended and have no systematic structure. However, a preliminary reading of all observations identified recurring themes, affect, and behavior which are discussed under Behavioral Coding.

Materials and Measures

Behavioral Coding

The behavioral coding scheme used to code the observations was designed to consider child affect, engagement, digital skill, and social dynamics. It was adapted from the Specific Affect Coding System, SPAFF (Coan & Gottman, 2007), and the Dyadic Parent-Child Interaction Coding System, DPICS (Eyberg & Robinson, 1981), both of which were used in the above-mentioned case study (Zhukova et al., 2020). The lead researcher of the present study read the qualitative observations written during the sessions and identified the behaviors that the observer commonly noticed Specifically, three broad categories of observations were identified: affect, behaviors, and environmental events.

Regarding affect, four specific affective states were coded for using binary coding, yes or no, signifying the presence or absence of: (1) detached affect where the observer notes that the child is displaying signs of distraction or boredom; (2) engaged affect where the observer notes that the child is displaying signs of concentration; (3) positive affect where the observer notes that the child is displaying signs of happiness such as smiling or cheerfulness; and (4) negative affect where the observer notes that the child is displaying signs of frustration, anger, or sadness. Frequency of occurrence could not be extracted.

Behaviors were coded in three levels where the lowest is a negative observation, the highest is a positive observation, and the midpoint is the absence of an observation. These include two categories: (1) digital tablet skill, where the observer notes a child’s level of proficiency in using the tablet, navigating it, understanding the activities and such; and (2) social dynamics where the observer notes the nature of interactions the child has with the observer. Frequency of occurrence could not be extracted.

Finally environmental events were coded on a binary, yes or no. They are designed to keep track of events outside of the observer’s control and include three types: (1) environmental distraction, which is when the observer makes note of an event occurring in the testing room or outside of it that may disrupt the child’s focus; (2) distraction type which defines the nature of said disruptor; and (3) session glitch which keeps track of the occurrence of a technological disruption such as the app freezing or shutting off. Frequency of occurrence could not be extracted.

Table 2/Таблица 2

Description of affect, behaviors, and other variables

Описание аффекта, поведения и других переменных

Performance Indicators

BELLA was used to evaluate children’s pre-academic, social-emotional, and cognitive skills. One pilot path was given per session, with the goal of each child completing a total of two. In addition to the domains (early literacy, numeracy, science, and social-emotional development) that constitute the pilot path, items varied in cognitive skill (analytical, creative, and practical thinking) and difficulty level (easy, medium, and hard). All pilot paths follow the same structure and item order. The items, while different, measure the same domains and cognitive skills. The difficulty level varies between items across pilot paths while maintaining a balanced approach. At the end of a session, the data are saved and each item is given a score between 0 and 100. The total average was calculated for each child as a representation of their overall performance.

Interrater Reliability

Three research assistants were trained in coding the observations. The supervising graduate student provided the training and participated in the coding. A total of 20 observations from the dataset were used for training. The group discussed discrepancies, refined the coding system, and proceeded with coding the full dataset. The process unfolded across multiple weeks during an academic semester. Fleiss’ kappa was used to calculate consistency across coders. The coding lead made the final decisions in cases of disagreement. The complete dataset was coded, and consistency across coders showed κ values ranging from .36 to .89. Interrater reliability indicated almost perfect agreement for positive affect (κ = .89), fair agreement for negative affect (κ = .36), nearly perfect agreement for engaged affect (κ = .86), substantial agreement for detached affect (κ = .74), substantial agreement for digital skill (κ = .63), moderate agreement for social dynamics (κ = .47), and substantial agreement for environmental distractions (κ = .70).

Statistical Approach

Descriptive Analysis. A descriptive analysis of the sample, performance scores,  observed affect, and behavior during the first session of BELLA was conducted. The analysis utilized contingency tables to calculate the frequencies and proportions for affect (binary) variables, nominal behavior variables, and environmental distractions (binary). A stratification was conducted to illustrate the distribution of variables across sex, age, and both sex and age. Fisher’s Exact Test was applied to explore initial associations between affect and behaviors, as well as between demographics and affect and behaviors. Fisher’s Exact Test is considered appropriate due to the small sample size. The results are shown in Table 11.

Comparative Analysis. A comparative analysis was conducted to evaluate whether changes in affect and behaviors were significant from session one to session two. Participants who had only one session and those who had a session interrupted were excluded. The McNemar-Bowker test was applied. Digital skill was converted into three variables: good digital skill, neutral digital skill, and poor digital skill. Social dynamics were similarly converted into good social dynamics, neutral social dynamics, and poor digital skill. Both variables were converted post-coding. The McNemar-Bowker test was applied against the following null hypothesis: there is no difference in proportions between the two paired observations.

Multiple Linear Regression Analysis. A multiple linear regression was conducted to examine whether affect and behaviors predict the changes in performance scores controlling for age. Variables were manipulated to fit into a multiple linear regression model. Positive and Negative affect were combined to create the emotion variables. Engaged and Detached affect were combined to create the focus variables. These are described in Table 3. Digital skill and social dynamics were also converted into binary variables. These are described in Table 4. Neutral levels across all variables were used as a baseline comparison. These changes are reflected in Tables 3 and 4. Only a completed session 2 is used for these analyses. The first exposure is considered a training session. To determine the appropriate sample size, an a priori power analysis was conducted using G*Power (version 3.1.9.7). We assumed a medium effect size (Cohen’s f2 = .15) with α = .05 and a desired power (1 - β) = .80. The statistical test selected was a multiple linear regression, fixed model, R2 deviation from zero, with 11 predictors. The power analysis indicated that a minimum of N = 123 participants was required to achieve adequate power (power = .84) while fitting this model. For a higher confidence (power = .95), N = 178 participants would be required. Given the study’s sample size, this is adequate to detect effects of medium strength above the .80 threshold.

Table 3/Таблица 3

Combinations of Affect to Create Emotion and Focus Variables

Комбинации аффектов для создания переменных эмоций и фокуса внимания

Table 4/Таблица 4

Digital Skill and Social Dynamics Levels

Уровни цифровых навыков и социальной динамики

Results

Sample Description

The sample included 161 children. Of these, 87 were boys and 74 were girls. Their ages ranged from 3 to 6 years old, with the majority being 4 and 5 years old (66 and 76 participants, respectively). Only 2 participants were age 3, and 17 were age 6. The mean age of the sample was 4.65 years old, with a standard deviation of SD = .76.

Each session number and its sample characteristics were reviewed, and participants that never completed a full session were removed. Session three is unique in that it is the session where children who faced difficulty during one of the first two sessions had to have a third session. The numbers of occurrences of each affect and behavior are reported in the Tables 5-10 below.

Session 1

Table 5/Таблица 5

Counts of Affect During Session 1

Количество воздействий во время сеанса 1

Table 6/Таблица 6

Counts of Behaviors During Session 1

Количество поведений во время сеанса 1

Session 2

Table 7/Таблица 7

Counts of Affect During Session 2

Количество случаев аффекта во время сеанса 2

Table 8/Таблица 8

Counts of Behaviors During Session 2

Количество поведений во время сеанса 2

Session 3

Table 9/Таблица 9

Counts of Affect During Session 3

Количество случаев аффекта во время сеанса 3

Table 10/Таблица 10

Counts of Behaviors During Session 3

Количество поведений во время сеанса 3

First exposure to BELLA

Contingency-table analyses were conducted to identify the affect and behavior profiles children exhibited during their first exposures to BELLA. A total of 155 children were included. The attrition was due to children with incomplete first sessions, who had been removed. The most frequently displayed affect was engaged affect (n = 63; 40.6%), followed by detached affect (n = 45; 29%). Positive affect was observed in 24 children (15.5%), while negative affect was reported in 25 children (16.1%). Regarding digital tablet skills, 80 observations (51.6%) reflected it to be good, 34 (21.9%)—poor, and 41 (26.5%) neutral. Social dynamics showed a majority of neutral observations (n = 90; 58.1%), while good and negative social dynamics accounted for 48 (30.9%) and 17 (11%) observations, respectively.

Notable groupings were observed within the contingency-table, with no statistical analyses attempted. When comparing the occurrence of affect with the occurrence of behavior, we see that a positive affect frequently occurred with good social dynamics (n = 21) and least often with poor social dynamics (n = 0). Engaged affect commonly occurred with good digital skill (n = 42), whereas detached affect showed mixed associations, often paired with neutral or poor digital skill. Negative affect did not show any clear pairing with behavior.

The data were also grouped by age and sex, revealing the following: 5 year-olds were the most likely to show positive affect (n = 17), followed by 4 year-olds (n = 4) and 6 year-olds (n = 3). Engaged affect was similarly highest in 4 and 5 year-olds. Detached affect similarly shows a high prevalence in 4 (n = 21) and 5 (n = 17) year-olds. Sex differences were subtle, with girls showing slightly lower rates of affect and boys showing higher rates of good digital skill while simultaneously being more likely to display poor digital skill. In terms of social dynamics, girls had slightly more good interactions, while boys had more neutral and negative interactions.

Correlations and Dependency Between Variables

To explore associations between affective states and behavioral observations further, a series of pairwise Spearman’s rank correlation were conducted, and significant correlations are reported. The analysis revealed that age is positively correlated with social dynamics (rho = .276; p < .001) and positive affect (rho = .188; p < .05). Digital skill negatively correlated with negative affect (rho = .291; p <.001) and positively correlated with positive affect (rho = .291; p < .001). Social dynamics also had a strong positive correlation with positive affect (rho = .488; p < .001). These findings indicate that older children are more likely to show good social engagement and positive affect, and that higher digital skill is linked to both greater engagement and reduced negative affect. Fisher’s Exact Test was applied to identify dependency or relatedness between affect and behaviors, with six significant relationships. All tested variables are reported in Table 11.

Stratification by age or sex is not reported in Table 11. Findings reveal that amongst 4 year- olds (n = 62), positive affect was associated with social dynamics (p < .05), negative affect was associated with digital skill (p < .05) and social dynamics (p < .05) as well. Engaged affect was associated with digital skill (p < .05), and detached affect was associated with social dynamics (p < .05). Amongst 5 year-olds (n = 74), the association between positive affect and social dynamics persisted (p < .001) as well as the association between negative affect with digital skill (p < .05). Detached affect was associated with both social dynamics and environmental distraction. Digital skill was also associated with social dynamics. Finally, the 6 year-old group (n = 17) revealed no significant associations. This may be due to the smaller number of participants in that group. Sex-related findings aren’t significant, but we observed a few trends within the sample where being a boy had a modest association with negative affect and digital tablet skills. Being a girl was associated with both positive affect and good social dynamics, as well as engaged affect and good digital tablet skills.  

Table 11/Таблица 11

Fisher’s Exact Test: Relatedness Between Categories

Точный тест Фишера: взаимосвязь между категориями

Change in Affect and Behavior from Session 1 to Session 2

To assess whether affect and/or behaviors changed between the first and second exposure to BELLA, a McNemar test was applied to paired data from 133 participants. These participants completed both sessions without glitches or interruptions. Findings revealed that engaged affect (χ² = 10.42, p < .005) significantly increased, and detached affect significantly decreased (χ² = 4.02, p < .05). Positive and negative affect exhibited no significant changes and remained relatively stable across sessions.

Regarding behavior, results show that all three levels of digital skill significantly shifted between sessions. Specifically, poor digital tablet skills decreased (χ² = 6.50, p < .05) while neutral (χ² = 28.90, p < .0001) and good digital tablet skills increased (χ² = 13.43, p < .001). Overall, this indicates that children’s use of BELLA and the tablet generally improved across just two sessions. Social dynamics show no significant changes across its three levels, suggesting that the social engagement of children remained relatively stable across sessions.

Linear Regression: Performance

Sample and Scores

The sample used for the linear regression included 148 participants. The smaller sample size is due to children missing performance scores, children that only had one session, and children with incomplete sessions. Only a complete second or third session for each child was included in this analysis. The first session is considered a training session. The sample completed a total of 33-items, similar across pilot paths. Performance scores were calculated as percentages to reflect the number of correct answers. Performance scores ranged from 20.24 to 93.39. The median is 66.12 and the mean is 63.96. The standard deviation is 15.63.

Dummy-Coding

New variables were created in line with Tables 3 and 4. Positive affect and Negative affect were combined to create an Emotion variable. Emotion has four levels: positive affect only, negative affect only, mixed positive affect and negative affect, and neutral affect. Neutral affect served as the comparative baseline for the regression. Engaged affect and detached affect were combined to create a Focus variable. Focus has four levels, similar to Emotion. Neutral Focus served as the baseline. Digital tablet skills and Social dynamics each have three levels. Each level has been decomposed into Good, Poor, and Neutral, giving three levels for each variable. Neutral is used as the baseline.

Regression

Overall, the regression model was statistically significant, F(11, 136) = 12.23, p < .001, and explained approximately 49.7% of the variance in BELLA performance scores (R2 = .497). Specifically, two of the predictors were statistically significant. Children who displayed positive affect without negative affect scored higher on BELLA (β = .168, p < .05). Children observed to have negative digital tablet skills (e.g., incorrect tablet manipulations) scored substantially lower (β = –.403, p < .001). Age was a strong covariate (β = .358, p < .001). No other emotion, focus, or social dynamic combinations were statistically significant. See Table 12 (below) for the full set of regression coefficients. Given the number of predictors, multicollinearity was examined using pairwise correlations (Table 13) and variance inflation factors (Table 14). The strongest pairwise correlation was between positive affect and positive social dynamics with r = .55. All other predictor pairs were below .40. These do not meet the threshold of .7 for potential multicollinearity concerns. Variance inflation factors ranged from 1.04 to 2.29, with all tolerance values above .43, confirming no problematic collinearity. The adjusted R2 for the full model was .457. A reduced model retaining only the three significant predictors (positive affect, poor digital skill, and age) yielded an adjusted R2 of .414. Model comparison indicated that the full model provided a significantly better fit, F(8,136) = 2.41, p < .05.

The results show that only negative digital skill and positive affect significantly contributed to the performance scores. The non-significant results, however, still show trends that can inform further investigation. Specifically, the presence of a negative affect and detached affect seemed to be associated with lower scores, while engaged affect was associated with higher scores. Good digital tablet skills trended negatively, which may be due to noise in the data. Positive social dynamics were associated with lower performance but not as strongly as negative social dynamics. The nature of observed social interactions needs to be more specific in order to differentiate between disruptive social interactions that take away from the task at hand, or interactions that are task oriented. Age remains the most critical factor influencing performance. Given the age range (3 to 6 year-olds) and the fact that it is a critical developmental period, this is expected.

Table 12/Таблица 12

Multiple linear regression results: performance change

Результаты множественной линейной регрессии: изменение производительности

Note: «*» -  Significant p-values at the 0.05 level.

Table 13/Таблица 13

Pairwise Pearson Correlations Among Regression Predictors

Парные корреляции Пирсона между предикторами регрессии

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

Table 14/Таблица 14

Variance Inflation Factors and Tolerance Values for Regression Predictors

Коэффициенты инфляции дисперсии и допустимые значения для предикторов регрессии

Note. VIF > 5 indicates concern; VIF > 10 indicates serious multicollinearity. Tolerance < .20 is concerning. All values are within acceptable ranges.

Discussion

The study examined affective and behavioral patterns exhibited by children when using BELLA in an isolated setting external to their regular classroom. The goal was to explore the behavioral and emotional user experience of the TD based assessment, and the learning process as they unfold during the sessions, which has been long overlooked in preschoolers. There were three steps to evaluating affect and behaviors: (1) an initial descriptive analysis to grasp what characterizes children’s first time experience when exposed to a tablet-based assessment; (2) a comparative analysis to identify the change in children’s experience with BELLA from the initial session to the second session; and (3) a regression analysis to identify which affect and behaviors contributed to children’s performance on BELLA items.

Descriptive Analysis

Studies have scarcely investigated the emotional and behavioral experience of preschool age children during digital media use, focusing on cognitive outcomes (Huber et al., 2018) or narrow learning domains such as literacy only (Neumann, 2018). The descriptive analysis offered insight into how preschoolers experience digital media within the educational context, extending current knowledge. Additionally, the observations are made during the first experience with BELLA, offering a chance to observe how the child adapts to a new task, with multiple knowledge and cognitive domains administered on a digital tablet. The observed responses reflect both developmental readiness and the degree to which children are familiar with navigating digital environments independently. Good digital skill is important, and was linked with increased positive affect and decreased negative affect, suggesting that procedural fluency may buffer emotional dysregulation and promote learning enjoyment. The evidence demonstrates that digital skill matters for more than task performance; it matters for children’s affective experiences, which are associated with their overall performance.

Positive affect was closely associated with good social dynamics and engaged focus, suggesting that emotional, social, and behavioral components jointly contribute to how children approach digital tasks. Previous research supports the role of emotionally supportive engagement in learning. (Denham & Liverette, 2019), and Strouse & Ganea (2017) demonstrated that positive affect and attention mediated vocabulary learning in toddlers interacting with electronic books. However, the findings were obtained in controlled lab settings with parent-child dyads, rather than in naturalistic contexts where children may interact with digital tools independently. This study extends the affect and learning link to a school-based setting in which the preschoolers navigated a tablet assessment without direct adult scaffolding in their first time exposure. Given the interdependence patterns identified, it is evident that an engaged (focused) child is more likely to have good digital skill, while a detached (distracted) child is less likely to. This holds for both a lab and a familiar school setting. Teachers should promote engagement and make sure that children have  adequate digital skills to be able to maintain their focus.  Such a pattern may also indicate that both emotional and behavioral readiness shape how children approach digital tasks. The associations identified were more robust for the 4-5 years old range, although that may be explained by most of the sample being in that age range, while 3- and 6-year-olds were a minority. Sex on the other hand did not have significant relationships and only suggests that boys have more behavioral extremes (good vs poor digital skill), and girls showed slightly more positive social dynamics. 

Overall, the analysis suggests that children were able to concentrate and fully pay attention to the application, and that half of them had good proficiency with the tablet interface. Specifically, the first experience with BELLA evokes positive affect for some, and difficulty for others when they do not display good digital skill. The associations identified strengthen the idea that age 4 to 5 is a key developmental period for emotion-behavior integration in digital contexts (Kolloff et al., 2025; Zimmerman, 2002). Children in this range appear to be sensitive to both internal states and external task challenges. Age is a critical driving force that affects every child’s experience.

Comparative Analysis

When comparing children’s affective and behavioral responses between their first and second sessions, clear patterns emerged. Engaged affect significantly increased, while detached affect significantly decreased, suggesting that upon the second exposure, children were more focused and less bored or distracted. These shifts align with research on learning engagement and self-regulatory development in early childhood, which shows that familiarity with a task can reduce cognitive load and improve sustained attention (Kubsch et al., 2025; Zimmerman, 2002). In parallel, digital skill improved. The total number of children showing poor digital navigation decreased, while neutral and good digital skill increased. Even a brief repeated interaction with the tablet was sufficient to support perceptible digital skill learning. BELLA’s interface could be a contributing factor to the ease of navigation and control of the application. The familiarity with the tool could have helped children struggle less while navigating the screen, showing less errors related to BELLA use, as opposed to errors with manipulating a tablet. 

In contrast, social dynamics remained unchanged across sessions. This can be interpreted in different ways. Children’s social interactions with the observer may be more resistant to change because children need more time to become familiar with new individuals, or could be due to inconsistencies in the collection of behavioral data. Nevertheless, while social engagement in early childhood is often context-dependent and is typically particularly affected by digital media (Blumberg et al., 2019; Herodotou, 2018), interactions with the caregiver were relatively brief and confined by the nature of the activity.

Predictive Analysis

The multiple linear regression analysis examined whether affective states and behavioral observations during the session were associated with children’s performance scores on BELLA. The overall model was statistically significant and explained nearly half of the variance in performance, suggesting that observable emotional and behavioral cues meaningfully contribute to academic outcome in digital assessments. Specifically, three predictors stood out. Children who displayed positive affect, without concurrent negative affect, had significantly higher scores. Those with poor digital skills scored substantially lower. Age was the strongest predictor in the model, with older children performing better overall. This echoes findings in Tan and colleagues (2023) where item performance increased with age. Age, however, is not the sole metric that needs to be included in future studies. Specifically, we propose that emotional positivity and digital skill contribute independently to positive performance, and highlight the importance of a child’s positive experiences with regards to enjoyment and skill in early childhood assessment. The emotional and behavioral attributes of an experience are as important to account for, particularly in this age group that is still learning self-regulation.

Other emotion and focus combinations did not reach statistical significance, but revealed meaningful trends. Negative affect and detached affect were generally associated with lower performance, while engaged focus was associated with higher performance. This pattern is consistent with emotion-cognition interaction models, which suggest that sustained attention and affective engagement support deeper processing and task completion (Halverson & Graham, 2019; Pekrun & Linnenbrink-Garcia, 2014). Negative emotion or disengagement may disrupt cognitive flow or increase off-task behavior (D’Mello & Graesser, 2014; Tan et al., 2021).

Interestingly, social dynamics were not predictive of performance at all. This could be due to the fact that the observer tried not to intervene, or observation inconsistencies. It could also be that social interactions are simply not a contributor to performance when using a tablet-based assessment that was designed for independent use with minimal caregiver intervention.

Overall, findings point to the importance of emotion and digital skill in predicting children’s success with digital assessment tools. Age remains the dominant contributor, however, it is not operating on its own. Age influences these behavioral and affective dimensions, particularly given the early developmental period examined.

Implications

The findings carry several implications for early childhood education. First, the observed relationships between affect, behavior, and performance underscore that digital assessments like BELLA do not only capture competency and learning; they elicit real-time indicators of readiness, self-regulation, and engagement. Emotional cues such as engaged affect or negative affect (e.g., frustration), and behavioral observations such as digital skill and fluency provide essential context to the child learning process. The findings align with sociocultural perspectives that emphasize digital tools as influential cultural mediators in development (Taber, 2025), and support the idea that children’s responses to digital environments are dynamically shaped by both internal and contextual variables.

Second, the predictive value of digital skill for performance reinforces the arguments that digital literacy is now a common developmental skill. Digital readiness shapes how children access, navigate, and benefit from digital learning. Teachers should not assume that preschoolers already know how to use TD. Digital literacy should be supported through practice and scaffolding, especially when children have limited access at home (McManis & Gunnewig, 2012; Neumann, 2016).

The link between positive affect and higher performance supports the evidence that emotional engagement, or focus, facilitates cognitive processing and deeper learning. These affective experiences are part of what makes digital learning effective, when well-designed. Educational technologies for children shouldn’t only measure knowledge. They should be designed to facilitate emotional regulation, and to allow for a level of co-use with the caregiver. The latter component is critical with digital screen use in education, although the results in this study do not support it.

From a research standpoint, the results affirm the value of a mixed-method approach, capitalizing on observational information. There is already much that relies on parental reports, performance data, and other measures. Children’s inherent affective and behavioral experience is overlooked although it is part of critical dimensions of early development.

Limitations and Future Directions

The study is limited by its use of secondary observational data. While the data are rich in naturalistic insights, they were relatively free-form, and subject to the observer’s bias. A systematic naturalistic strategy is needed, with strict control and consistency in the types and scope of observations to be collected. To our knowledge, such research is limited, and this work presents an example of the information that might be obtained from such observational studies. Additionally, the observations were not able to capture the dynamic nature of the BELLA sessions and the flow of emotions and behavior. A careful, time-documented data collection plan needs to be utilized in future research. This would also allow researchers to demonstrate the influence the caregiver has on emotions, learning behaviors, and overall educational experiences, given the emphasis on social dynamics. In the current study, the data are limited and cannot fully support this claim. In addition, while the regression model accounted for a relatively high proportion of variance in performance, collinearity diagnostics indicated no problematic shared variance among predictors. However, a reduced model with only the three significant predictors explained nearly as much variance, suggesting that the non-significant predictors contributed modestly beyond the core three. Future studies may benefit from a more parsimonious model or a larger sample to improve the predictor-to-participant ratio.. Another limiting factor is that the study did not evaluate the pre-existing digital skill of children or their current educational competencies. The sample was also from a private suburban pool of preschools, which limits the generalizability of the results. Future studies should incorporate structured observational protocols, video coding, and perhaps screen-capture of the tablet manipulations in order to maximize the information recorded. Time-stamping behaviors and affect are critical to reveal the dynamic nature of learning-related behaviors and emotions. A more diverse sample needs to be studied, including public and lower-income schools from across states to increase generalizability. Finally, a clear breakdown of the affect and behaviors of interest will add a level of detail that can differentiate between specific emotional states such as happiness and excitement, or sadness and anger.

1.  Barr, R. (2019). Growing up in the digital age: Early learning and family media ecology. Current directions in psychological science, 28(4), 341-346.
2. Blumberg, F. C., Deater‐Deckard, K., Calvert, S. L., Flynn, R. M., Green, C. S., Arnold, D., & Brooks, P. J. (2019). Digital games as a context for children's cognitive development: Research recommendations and policy considerations. Social Policy Report, 32(1), 1-33.
3. Camras, L. A. (2011). Differentiation, dynamical integration and functional emotional development. Emotion Review, 3(2), 138-146.
4. Chaudron, S., Di Gioia, R., & Gemo, M. (2015). Young children (0-8) and digital technology. A qualitative exploratory study across seven countries. Joint Research Centre. European Commission.
5. Chiong, C., & Shuler, C. (2010). Learning: Is there an app for that. Investigations of young children’s usage and learning with mobile devices and apps. New York: The Joan Ganz Cooney Center at Sesame Workshop,
6. Coan, J. A., & Gottman, J. M. (2007). The specific affect coding system (SPAFF). Handbook of emotion elicitation and assessment, 267, 285.
7. D’Mello, S. K., & Graesser, A. C. (2014). Confusion. In International handbook of emotions in education (pp. 289-310). Routledge.
8. Denham, S. A., & Liverette, K. H. (2019). The emotional basis of learning and development in early childhood education. In Handbook of research on the education of young children (pp. 43-64). Routledge.
9. Eyberg, S. M., & Robinson, E. A. (1981). Dyadic parent-child interaction coding system. Seattle, WA: Parenting Clinic, University of Washington.
10. French, B. (2013). Brigance screens III technical manual. North Billerica: Curriculum Associates LLC.
11. Griffith, S. F., Hagan, M. B., Heymann, P., Heflin, B. H., & Bagner, D. M. (2020). Apps as learning tools: A systematic review. Pediatrics, 145(1).
12. Halverson, L. R., & Graham, C. R. (2019). Learner engagement in blended learning environments: A conceptual framework. Online learning, 23(2), 145-178.
13. Herodotou, C. (2018). Young children and tablets: A systematic review of effects on learning and development. Journal of Computer Assisted Learning, 34(1), 1-9.
14. Hill, D., Ameenuddin, N., Reid Chassiakos, Y., Cross, C., Hutchinson, J., Levine, A., Boyd, R., Mendelson, R., Moreno, M., & Swanson, W. S. (2016). Media and Young Minds. Pediatrics, 138(5). https://doi.org/10.1542/peds.2016-2591
15. Huber, B., Yeates, M., Meyer, D., Fleckhammer, L., & Kaufman, J. (2018). The effects of screen media content on young children’s executive functioning. Journal of experimental child psychology, 170, 72-85.
16. Kabali, H. K., Irigoyen, M. M., Nunez-Davis, R., Budacki, J. G., Mohanty, S. H., Leister, K. P., & Bonner Jr, R. L. (2015). Exposure and use of mobile media devices by young children. Pediatrics, 136(6), 1044-1050.
17. Khalaf, S., Kilani, H., Razo, M. B., & Grigorenko, E. L. (2022). Bored, distracted, and confused: Emotions that promote creativity and learning in a 28-month-old child using an iPad. Journal of Intelligence, 10(4), 118.
18. Kilani, H., Markov, I. V., Francis, D., & Grigorenko, E. L. (2024). Screens and Preschools: The Bilingual English Language Learner Assessment as a Curriculum-Compliant Digital Application. Children, 11(8), 914.
19. Kolloff, K., Ger, E., & Roebers, C. M. (2025). The role of metacognitive experience: Can feedback training affect kindergarteners' monitoring accuracy? Metacognition and Learning, 20(1), 1-32.
20. Kopp, C. B. (1982). Antecedents of self-regulation: a developmental perspective. Developmental psychology, 18(2), 199.
21. Kubsch, M., Strauß, S., Grimm, A., Gombert, S., Drachsler, H., Neumann, K., & Rummel, N. (2025). Self-regulated Learning in the Digitally Enhanced Science Classroom: Toward an Early Warning System. Educational Psychology Review, 37(2), 1-38.
22. Madigan, S., Browne, D., Racine, N., Mori, C., & Tough, S. (2019). Association between screen time and children’s performance on a developmental screening test. JAMA pediatrics, 173(3), 244-250.
23. McManis, L. D., & Gunnewig, S. B. (2012). Finding the education in educational technology with early learners. Young children, 67(3), 14-24.
24. McPake, J., Plowman, L., & Stephen, C. (2013). Pre‐school children creating and communicating with digital technologies in the home. British Journal of Educational Technology, 44(3), 421-431.
25. Neumann, M. M. (2016). Young children's use of touch screen tablets for writing and reading at home: Relationships with emergent literacy. Computers & Education, 97, 61-68.
26. Neumann, M. M. (2018). Using tablets and apps to enhance emergent literacy skills in young children. Early Childhood Research Quarterly, 42, 239-246.
27. Pekrun, R., & Linnenbrink-Garcia, L. (2014). Introduction to emotions in education. In International handbook of emotions in education (pp. 1-10). Routledge.
28. Plass, J. L., & Kaplan, U. (2016). Emotional design in digital media for learning. In Emotions, technology, design, and learning (pp. 131-161). Elsevier.
29. Plass, J. L., Homer, B. D., & Kinzer, C. K. (2015). Foundations of game-based learning. Educational psychologist, 50(4), 258-283.
30. Rideout, V. (2021). The Common Sense Census: Media Use by Kids Age Zero to Eight in America, A Common Sense Media Research Study,[United States], 2013, 2017.
31. Shapiro, E. S., & Gebhardt, S. N. (2012, 2012/09/01). Comparing Computer-Adaptive and Curriculum-Based Measurement Methods of Assessment. School Psychology Review, 41(3), 295-305. https://doi.org/10.1080/02796015.2012.12087510
32. Strouse, G. A., & Ganea, P. A. (2017). Parent–toddler behavior and language differ when reading electronic and print picture books. Frontiers in psychology, 8, 677.
33. Taber, K. S. (2025). Mediated learning leading development—The social development theory of Lev Vygotsky. In Science education in theory and practice: An introductory guide to learning theory (pp. 275-292). Springer.
34. Tan, J., Mao, J., Jiang, Y., & Gao, M. (2021). The influence of academic emotions on learning effects: A systematic review. International journal of environmental research and public health, 18(18), 9678.
35. Tan, M., Kilani, H., Markov, I., Hein, S., & Grigorenko, E. L. (2023). Assessing Cognitive Skills in Early Childhood Education Using a Bilingual Early Language Learner Assessment Tool. Journal of Intelligence, 11(7), 143. https://www.mdpi.com/2079-3200/11/7/143
36. Teaching Strategies, L. (2022). Teaching Strategies GOLD. Bethesda: Teaching Strategies LLC.
37. Vygotsky, L. (1978). Mind in society: The development of higher psychological processes (Vol. 86). Harvard university press.
38. Vygotsky, L. (1994a). The development of academic concepts in school aged children. The Vygotsky reader, 355-370.
39. Vygotsky, L. (1994b). Tool and symbol in child development. The Vygotsky reader.
40. Wartella, E., Rideout, V., Lauricella, A. R., & Connell, S. (2013). Parenting in the age of digital technology. Report for the center on media and Human development school of communication Northwestern University, 297-309.
41. Yun, C., Melnick, H., & Wechsler, M. (2021). High-Quality Early Childhood Assessment: Learning from States' Use of Kindergarten Entry Assessments. Learning Policy Institute.
42. Zaranis, N. (2016). The use of ICT in kindergarten for teaching addition based on realistic mathematics education. Education and Information Technologies, 21, 589-606.
43. Zhukova, M. A., Kilani, H., Garcia, J. L., Tan, M., Parell, P. D., Hein, S., & Grigorenko, E. L. (2020). How Does a Toddler Experience Digital Media? A Case Study of a 28-Month Old Child Learning to Use the iPad. Grantee Submission, 10(3), 5-23.
44. Zimmerman, B. J. (2002). Becoming a self-regulated learner: An overview. Theory into practice, 41(2), 64-70.