Next Article in Journal
Cybersecurity Regulations and Software Resilience: Strengthening Awareness and Societal Stability
Previous Article in Journal
A Crisis, or Not a Crisis? A Case Study About the Misperception of a Local Social Media Crisis
Previous Article in Special Issue
Semiotic Fracturing of Rural Cultural Symbols in Short Video Ecosystems: A Critical Discourse Analysis of “Tǔ Wèi” Labeling and Cultural Subjectivity Construction
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Social Sciences
Volume 14
Issue 10
10.3390/socsci14100576
socsci-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

From Screens to Schooling: Associations Between Adolescent Technology Use and Gendered College Enrollment

by
MacKenzie A. Christensen
Department of Sociology, University of Oregon, Eugene, OR 97403, USA
Soc. Sci. 2025, 14(10), 576; https://doi.org/10.3390/socsci14100576
Submission received: 7 August 2025 / Revised: 18 September 2025 / Accepted: 18 September 2025 / Published: 26 September 2025
(This article belongs to the Special Issue Digitally Connected: Youth, Digital Media and Social Inclusion)
Browse Figure
Versions Notes

Abstract

Young women now surpass young men in college enrollment. While research suggests that this gender gap emerges during adolescence, less is known about the role of adolescent technology use in shaping which youth go on to pursue higher education. This study draws on two youth-focused supplements to the Panel Study of Income Dynamics to follow a cohort of adolescents into young adulthood, examining whether patterns of technology use in 2007 predict college enrollment in 2017. Using latent class analysis, I identify three distinct profiles of adolescent technology use: Connected Communicators, Web Browsers, and Digitally Disconnected. Results from multivariate analyses reveal that the relationship between adolescent technology use and college enrollment varies significantly by gender. Among boys, Connected Communicators were more likely to enroll in college than their peers in other classes. Among girls, however, Connected Communicators were less likely to enroll in college than other girls and boys in the same class. These findings suggest that the educational benefits of digital engagement may be more pronounced for boys than for girls. In the context of ongoing concerns about gender disparities in higher education, this study offers new insight into how adolescent technology use may contribute to gendered pathways to college.

1. Introduction

The gender composition of higher education has shifted dramatically in recent decades, with young women now surpassing young men in college enrollment (Buchmann et al. 2025; Reber and Smith 2023; Riegle-Crumb 2010). Research attributes this gender gap, in part, to unequal experiences during adolescence, including differences in academic achievement, peer networks, and parental expectations and involvement (Fitton et al. 2013; Lippman et al. 2008; Reber and Smith 2023). Yet less is known about whether adolescents’ digital technology use may also shape educational trajectories. This is an important oversight given that adolescence is a critical developmental stage in which young people begin to engage independently with information and communication technologies (ICTs), such as computers, social media, and video games (Fitton et al. 2013).
Today’s digital youth have grown up alongside a rise in household digital technologies that has fundamentally altered the landscape of adolescence. As a result, students arrive in classrooms with diverse digital skills, experiences, and interests that shape their educational engagement in complex ways (Livingstone et al. 2017; Rafalow 2018). Some forms of technology use, such as using ICTs for schoolwork or research, have been linked to improved academic outcomes, including higher test scores and GPA (Robinson et al. 2018; Skryabin et al. 2015; Zhang and Liu 2016). Other forms—particularly recreational activities such as gaming or social media—have been thought to distract young learners and hinder educational performance (Desimoni et al. 2024; Junco and Cotten 2012; Kirschner and De Bruyckere 2017).
Although this work offers valuable insight into the educational implications of adolescent technology use, several limitations remain. Existing research often focuses on isolated behaviors—such as school digital device use or social media use—without capturing how youth engage with multiple technologies in patterned and overlapping ways (e.g., Wang et al. 2024; Touloupis and Campbell 2024). As a result, we know less about the profiles of adolescents’ digital lives. Further, while prior studies have identified social differences in technology access and use, fewer have examined how these differences map onto holistic patterns of engagement, or how these patterns reflect broader social and cultural contexts. Finally, the long-term implications of adolescent technology use remain unclear. Most research has focused on short-term academic outcomes, leaving open the question of whether early digital experiences shape later trajectories—such as college enrollment—and whether those pathways are gendered. As youth come of age in a context of increasing digital connectivity, understanding how their digital engagement is structured, stratified, and linked to long-term outcomes is essential for understanding who goes on to pursue higher education.
In this paper, I address these gaps by drawing on two youth-focused supplements to the Panel Study of Income Dynamics (PSID) to follow a cohort of digital youth from adolescence into young adulthood. I examine whether patterns of adolescent technology use in 2007 predict college enrollment a decade later and assess whether these associations differ by gender. This study makes three key contributions to research on adolescent technology use and educational attainment. First, by moving beyond variable-centered approaches, latent class analysis (LCA) identifies distinct profiles of adolescent technology use, offering a multidimensional view of youths’ digital lives. Second, by analyzing the characteristics of latent class membership, the results reveal how patterns of digital engagement reflect broader social, cultural, and structural forces. Third, by linking adolescent technology use to college enrollment in young adulthood, the results provide critical insight into how early digital experiences shape future educational outcomes. Thus, by leveraging longitudinal data that follows youth over ten years, this study offers new insight into how adolescents’ digital engagement patterns may shape gendered pathways to higher education.

2. Background

2.1. Linking Adolescent Technology Use to Educational Trajectories

There are theoretical reasons to expect that adolescent technology use may shape future educational outcomes, including college enrollment. Specifically, the literature on digital inequalities links three interrelated levels of digital divides to unequal social outcomes: disparities in access to technology (first-level divides), differences in technology use and skills (second-level divides), and inequalities in the outcomes generated through technology use (third-level divides) (Helsper 2021).
First-level digital divides—or differences in access to digital devices and the internet—have long been recognized as key drivers of educational inequality (Van Dijk 2020). In 2007, when the adolescents in this study were coming of age, access to home computers and reliable internet varied considerably across households. Unequal access to high-quality ICTs has been shown to exacerbate academic disparities by limiting opportunities for digital learning, homework completion, and the development of digital skills (Becker 2023; Helsper 2021). As ICT access has become increasingly widespread in the U.S., research suggests that the educational benefits of access alone have diminished over time (Bulman and Fairlie 2016). Though access inequalities remain, in 2007 (the time of this study), first-level divides may have played a more central role in shaping youths’ technology use and education. Therefore, technology access may remain a necessary—if insufficient—factor shaping college trajectories among today’s youth.
Second-level digital divides, or differences in how youth use technologies, are increasingly seen as an important precursor to academic achievement (Becker 2023; DiMaggio et al. 2004; Helsper 2021; van Deursen and van Dijk 2014). A growing body of research suggests that using digital technologies for educational purposes—such as conducting online research, completing homework, or watching instructional videos—is positively associated with academic achievement and test performance (Lei and Zhao 2007; Petko et al. 2017; Robinson et al. 2018; Skryabin et al. 2015; Wang et al. 2024; Zhang and Liu 2016). For example, Robinson et al. (2018) found that American high school students who frequently engaged in educational ICT use reported higher GPAs, even after adjusting for sociodemographic characteristics and course placement.
In contrast, entertainment and communication technologies (e.g., gaming, social media, texting) have been linked to poorer academic outcomes. According to distraction theories, these technologies may undermine learning by encouraging multitasking, diverting attention, and reducing students’ motivation (Kirschner and De Bruyckere 2017; May and Elder 2018; Rosén and Gustafsson 2016; Xiao and Sun 2022). Indeed, Desimoni et al. (2024) found that increased ICT use for communication and leisure activities was negatively associated with math achievement among eighth-grade students in Italy. Given the well-documented link between academic achievement and college enrollment (Allensworth and Clark 2020), such second-level digital divides are likely to contribute to longer-term educational inequalities.
However, it is unlikely that any single form of technology use—such as educational versus recreational—fully explains adolescents’ academic trajectories. Rather, it is likely the broader patterns of digital engagement that youth develop during adolescence that may be most consequential. Young people’s technology use is multifaceted, embedded in daily routines, and shaped by their social environments (Ito 2013). Adolescents rarely engage with technology in isolated ways. Instead, they often use multiple devices, often simultaneously, for a vast array of purposes (Faverio and Sidoti 2024). For instance, some may rely on ICTs for schoolwork and information seeking, while others may focus more on entertainment-oriented practices such as gaming and social media. Still others may blend both educational and recreational uses or exhibit generally low levels of digital engagement altogether. Importantly, however, patterns of technology use reflect not only personal preferences but also the broader social, cultural, and structural influences that shape adolescents’ digital lives.

2.2. Determinants of Adolescent Technology Use

While adolescent technology use may influence educational outcomes, their digital engagement styles are not uniform. Rather, patterns of technology use are shaped by a multitude of factors, including sociodemographic characteristics, family environments, peer groups, and educational contexts (e.g., Desimoni et al. 2024; Hargittai and Shafer 2006; Jackson et al. 2008; Rafalow 2018). Therefore, any link between technology use and later college enrollment may partly reflect preexisting differences in adolescents’ social backgrounds and experiences.
Demographic characteristics—including gender, age, race/ethnicity, and socioeconomic background—reflect structural inequalities that may underlie both digital engagement and educational outcomes. For instance, studies not only document clear gender differences in college enrollment (Buchmann et al. 2025; Reber and Smith 2023; Riegle-Crumb 2010) but also in technology use (Helsper 2010; Mollborn et al. 2021). Research finds that girls are more likely to use ICTs for educational and communicative purposes, whereas boys report higher rates of recreational use, especially gaming (Fraillon et al. 2014; Jackson et al. 2008; Tsai and Tsai 2010; Wasserman and Richmond-Abbott 2005; Xiao and Sun 2022). Socioeconomic status (SES) also strongly predicts both access to digital technologies and college attendance (Becker 2023; DiMaggio et al. 2004). For instance, students from higher SES backgrounds are more likely to engage in “capital-enhancing” ICT activities—such as using technology for schoolwork, information seeking, or skill development—that support academic success (Becker 2023; Notten and Becker 2017).
Educational factors further shape these patterns of digital use. College expectations, high school GPA, and academic self-efficacy are consistent predictors of future college enrollment (Allensworth and Clark 2020; Domina et al. 2011). They also likely inform how students engage with technologies. Students with greater academic confidence, who anticipate going to college, may be more likely to use ICTs for learning purposes, such as researching assignments or planning for college. Conversely, those who do not intend to pursue higher education, or those with less academic confidence, may be more inclined to use technologies for entertainment or leisure purposes. As such, educational experiences and expectations not only influence college trajectories but may also shape how adolescents interact with technologies.
Social environments also play an important role in shaping educational pathways and technology use. Peer influences—including whether their friends plan to attend college—can shape students’ academic experiences and their technology use, as adolescents often mirror the behaviors and interests of their friends (Morgan 2005; Ryan 2001). For instance, youth embedded in academically focused peer networks may be more likely to utilize digital tools for schoolwork or research, and less likely to engage with entertainment or leisure technologies. Similarly, parental expectations and communication about college not only increase the chances of college enrollment (Lippman et al. 2008; Morgan 2005; Stage and Hossler 1989) but may also influence adolescents’ relationships with ICTs. Parents who encourage their child’s educational growth may promote academic uses of ICTs while limiting recreational screen time, thus shaping their child’s digital engagement in ways that align with their academic goals. Ultimately, adolescents’ demographic characteristics, academic background, and social environments likely shape both educational and digital practices. Therefore, in examining whether patterns of technology use influence college enrollment, I account for the demographic characteristics and social contexts that may confound this relationship.

2.3. Differential Returns to Technology Use for Boys and Girls

It is also possible that girls and boys benefit differently from similar types of digital engagement—reflecting third-level digital divides (Helsper 2021). These divides refer to the social and structural conditions that influence whether individuals can translate their digital engagement into meaningful outcomes, such as college enrollment, while avoiding potential adverse effects of technology use. In other words, gendered perceptions of technology use may affect not just access or frequency of use, but the value assigned to those practices and the support they receive.
Digital activities—whether educational or recreational—are not perceived or rewarded equally across genders. Boys’ technology use, such as coding or searching the internet, is often legitimized within school and peer contexts and viewed as reflective of technological competence or future-oriented skills (Ito 2013; Sims 2014; Rafalow 2018). Even recreational activities, such as gaming or social media, are shown to generate social capital for boys, enhancing their peer status and signaling valued digital literacies that translate into academic support (Sims 2014; Rafalow 2018). For example, in an ethnographic study of middle school students, Sims (2014) found that even when boys and girls reported similar levels of technology use, teachers interpreted their digital engagement in gendered ways. Boys’ technology use—particularly around gaming and computing—was legitimized and encouraged by teachers. Whereas girls’ technology use—especially their use of social and communication technologies—was dismissed as frivolous, distracting, or unrelated to their academic success.
These gender differences reflect broader cultural beliefs that position technology as a masculine domain, where boys are presumed to have greater technological competence than girls (Correll 2004; Hargittai and Shafer 2006; Ridgeway 2011). These beliefs shape not only how educators respond to students’ digital behaviors, but also how students view their own abilities. Indeed, studies show that, regardless of actual ability, boys tend to report greater technological confidence than girls (Christensen 2023). Research also indicates that self-assessed competence is a key predictor of educational outcomes (Correll 2004; Morgan et al. 2013), suggesting that gendered beliefs may further reinforce unequal returns from youths’ digital practices.
As a result, boys may benefit more from their technology use because their engagement is more likely to be culturally legitimized, socially rewarded, and institutionally supported. In contrast, girls may find it more difficult to leverage their technology use for educational advancement, especially if their preferred forms of digital engagement are seen as less serious or academically relevant. Ultimately, digital inequalities are not just about what technologies adolescents use, but how those practices are interpreted, valued, and supported within educational and social contexts. Therefore, this study examines whether the relationship between adolescent technology use and college enrollment varies by gender.

3. Materials and Methods

3.1. Data

To explore the association between adolescent technology use and college enrollment, I use secondary data from the 2007 Child Development Supplement (CDS-07) and the 2017 Transition to Adulthood Supplement (TAS-17) to the Panel Study of Income Dynamics (PSID), a longitudinal study of individuals living in U.S. households (McGonagle et al. 2012). Beginning in 1968, the original PSID surveyed over 18,000 individuals from 5000 U.S. families and has since recruited up to seven generations of family members. As generations developed, the PSID added two youth-focused supplements on early childhood, adolescence, and young adulthood.
The CDS began in 1997 as a longitudinal study of PSID children aged 0–12 years. By 1997, the CDS I-III began collecting longitudinal data on up to two randomly selected 0–12-year-old children (n = 3653) and their caregivers from the main PSID families, with follow-up interviews conducted in 2002/03 and 2007/08. Each wave of the CDS includes data on children’s well-being, time use, and development across social contexts. In particular, the CDS collects detailed information on schooling and educational contexts from K-12th grade, and captures future family, academic, and occupational plans and expectations.
Recognizing the developmental importance of young adulthood, the PSID initiated a new longitudinal study in 2005, designed to follow children who aged out of the original CDS cohort during their transition into early adulthood, known as the Transition to Adulthood Supplement (TAS). Following these CDS alumni, the original TAS was collected biennially from 2005 to 2015. It was intended to fill a gap between information collected on children (CDS) and adults who have reached economic independence (the main PSID). The TAS includes information on young adults’ (ages 18 to 29 years) technology use, attitudes, expectations, and transitions in schooling, work, and family. The TAS was relaunched in 2017 and collected again in 2019, including remaining eligible participants from the original TAS sample but also including other PSID sample children, regardless of prior CDS or main PSID participation. Together, the CDS and TAS track youth from adolescence into young adulthood, enabling a longitudinal analysis of how adolescent technology use shapes later college enrollment.
For this study, I use data from the CDS-07, which was collected when the youth were aged 10–18, and from the TAS-17, which was collected when the same cohort was aged 18–28. I restrict my analyses to include only those surveyed in both the CDS-07 and the TAS-17 (n = 1168). I imputed missing data by generating 20 datasets and averaging the results across imputations (Allison 2012). After imputation, I dropped cases with missing data on the dependent variable (n = 4). Therefore, following a cohort of “digital youths” from adolescence through young adulthood, my final analytic sample consists of 1164 young adults aged 18–29 in 2017.

3.2. Measures

3.2.1. Dependent Variable

College enrollment was measured using data from the TAS-17, which asked respondents whether they had ever attended a college or university. College enrollment is then coded dichotomously (1 = ever enrolled), capturing both current and past enrollment.

3.2.2. Independent Variables

The key independent variables are latent classes of adolescent technology use, which are constructed from eight technology use items measured in the CDS-07. Adolescents reported the average number of days per month they engaged in eight digital technology activities: (1) texting, (2) using the computer for schoolwork, (3) instant messaging or chat rooms, (4) using email, (5) posting personal information on websites, (6) playing computer games, (7) browsing websites, and (8) other computer-based activities. Response options were categorized as: near-daily (1 = almost every day to every day), weekly (2 = once a week to two or three times a week), and almost never (3 = once or twice a month to never). These indicators were used in a latent class analysis (LCA) to identify distinct patterns of adolescent technology use.

3.2.3. Control Variables

The multivariate models adjust for a range of variables known to be associated with both adolescent technology use and college enrollment (e.g., Becker 2023; Monaghan 2021; Musick et al. 2012; Notten and Becker 2017). Adolescent variables were measured in the CDS-07 unless otherwise noted, and young adult variables were drawn from the TAS-17.
Sociodemographic controls include age (in years), race/ethnicity (non-Hispanic white, Hispanic/Latinx, non-Hispanic Black, and other/multiracial), parental education (in years), and logged family income from 2006. Educational background measures include high school GPA (TAS-17); adolescents’ math and reading self-efficacy (1 = low to 9 = high); hours spent on homework; plans to attend a four-year college (1 = at least “pretty likely”); friends’ plans to attend college (1 = most friends plan to go); parent’s expectation that the child will earn a four-year degree (1 = yes); and how frequently adolescents discussed college with a parent (1 = not in the past month to 6 = every day)
Technology background measures include the number of household electronic devices (sum of computers, game consoles, TVs, and cell phones); parental technology limits (average frequency of setting limits on games, internet, and email use, from 0 = never to 3 = often); and parental technology encouragement (average frequency of encouraging use of email, games, and internet, from 0 = never to 4 = often). Finally, models account for several young adult characteristics measured in TAS-17, including whether respondents were married or cohabiting, had at least one child, were employed full-time or part-time, and were living with their parents (all coded as 1 = yes).

3.3. Analytic Plan

The analysis proceeded in three stages. First, latent class analysis (LCA) was used to identify unobserved subgroups of adolescents based on their patterns of digital technology use. LCA offers several advantages over other methods. Unlike variable-centered approaches that focus on associations between individual variables, LCA is person-centered—it allows distinct patterns to emerge from the data rather than imposing them a priori. Additionally, compared to factor or cluster analysis, which assumes indicators load in the same direction, LCA captures multidimensional patterns of use, meaning individuals can score high on some digital activities but low on others.
To determine the optimal number of latent classes, I relied on a combination of fit statistics (Bayesian Information Criterion [BIC] and Akaike Information Criterion [AIC]), theoretical grounding, and interpretability of the class structure. Once the best-fitting model was identified, adolescents were assigned to classes based on their highest posterior probability of membership. Each class was characterized by conditional probabilities for each technology item, reflecting the likelihood of a given response conditional on class membership. Second, after constructing technology use classes, I examined descriptive characteristics of each latent class, examining statistically significant differences across groups using t-tests.
Third, I estimated the relationship between adolescent technology use classes and college enrollment using logistic regression models. Model 1 intersects latent classes of technology use with adolescents’ gender to check whether the association varied between adolescent girls and boys. Model 2 adjusts for adolescent-level covariates to account for selection into technology classes, and Model 3 includes additional young adult covariates that may influence college enrollment (and therefore helps isolate the relationship between adolescent technology use and college enrollment). To attend to potential confounding and selection bias, models are weighted using inverse probability weights (IPWs). IPWs have been shown to effectively reduce confounding bias when comparing latent classes and when class membership is defined using modal class assignment (Schuler et al. 2014). This was done by estimating adolescents’ probabilities of being assigned to each latent class using a multinomial logistic regression model that included all adolescent covariates described above. These probabilities were used to calculate inverse probability weights, with one class designated as the reference group. I then incorporated these IPWs, along with the original survey weights, into logistic regression models predicting college enrollment in 2017. This approach helps approximate the conditions of a pseudo-randomized design by reweighting the sample to balance observed adolescent characteristics across latent classes, allowing for a more robust estimate of the association between adolescent technology use in 2007 and college enrollment in 2017 (Schuler et al. 2014).

4. Results

4.1. Latent Classes of Adolescent Technology Use

I estimated solutions from one to four latent classes (Table 1). Three latent classes had the lowest BIC and AIC scores (AIC = 15,236.94; BIC = 15,486.91), indicating that a three latent class model generated the best model fit, and the likelihood-ratio test (G2) indicated that this number of classes fit as well as the saturated model (X2 = 2004.88, p = 1.00). Based on class interpretation, I assigned a label to each latent class summarizing the participants’ conditional response probabilities1. Therefore, three latent classes are used to represent adolescent technology use patterns in 2007.
Table 2 presents item-response probabilities for eight digital activities, showing how adolescents in each latent class engaged with technology2. The largest group, about 40% of youth, were the Web Browsers. These adolescents exhibited moderate and selective engagement with digital technologies. They were especially likely to report weekly use of the computer to browse websites (62%) and email (29%), but less likely to use these tools daily. For instance, just 11% used email near-daily, and 31% browsed websites daily. This group was characterized by relatively low engagement in social or expressive digital activities—only 20% texted near-daily and 0% posted personal information online near-daily. Instead, their technology use appeared to focus more on functional or academic purposes, with 41% using the computer weekly for schoolwork, and 56% engaged in other computer activities every week.
In contrast, the Connected Communicators (34% of adolescents) demonstrated high and diverse levels of digital activity. A majority reported using websites (99%), email (53%), texting (49%), and “other” computer activities (96%) on a near-daily basis. This group also showed elevated engagement in social platforms, with 41% participating in chat rooms near-daily and 41% posting personal information online near-daily—substantially higher than the other classes. Only 19% of Connected Communicators reported “almost never” using email, compared to 61% of Web Browsers and 97% of Digitally Disconnected adolescents. This class reflects broad, multidimensional digital engagement spanning social, expressive, and educational domains.
The third class, the Digitally Disconnected (25% of adolescents), reported the lowest levels of engagement across all domains. Nearly all reported “almost never” using email (97%), chat rooms (99%), or school-related digital tools (86%). Only 3% reported even weekly use of email, and just 4% engaged in “other” computer activities near-daily. This group also had the highest probability of “almost never” texting (84%), posting personal information (99%), and browsing websites (83%). Overall, their digital engagement was relatively minimal, suggesting a noteworthy disconnection from digital technologies during adolescence. These patterns underscore the heterogeneity in adolescent digital experiences. Adolescents appear to cluster into distinct usage profiles, ranging from low-engagement to broadly connected, which may shape their academic, social, and developmental trajectories differently.

4.2. Descriptive Characteristics of Adolescent Technology Use

Table 3 presents weighted descriptive statistics and significance tests comparing adolescents across the three latent classes of technology use: Web Browsers, Connected Communicators, and the Digitally Disconnected. College enrollment rates in 2017 varied meaningfully by class. Connected Communicators had the highest enrollment rate (0.77), which was significantly higher than that of Digitally Disconnected youth (0.64, p < 0.05), but not significantly different from Web Browsers (0.74).
The classes also differed on several demographic characteristics. Connected Communicators were significantly more likely to be girls (44%) and older on average (14.17 years) than both Web Browsers (13.03 years) and Digitally Disconnected youth (12.99 years, p < 0.05 for both comparisons). They also reported higher levels of parental education (14.45 years) and family income (logged mean = 11.19) compared to Digitally Disconnected adolescents (13.49 years and 10.89, respectively; p < 0.05).
Educational indicators followed similar patterns. Connected Communicators had the highest high school GPAs (3.12), followed by Web Browsers (3.03) and Digitally Disconnected youth (2.68), with all pairwise differences significant at p < 0.05. Web Browsers reported the highest levels of math self-efficacy (4.91), significantly greater than that of Connected Communicators (4.71, p < 0.05). Time spent on homework was also highest among Web Browsers (2.11 hours) and lowest among Connected Communicators (1.58 hours, p < 0.05). In terms of college-related expectations and discussions, Digitally Disconnected adolescents consistently reported less support. They were significantly less likely than Connected Communicators to say they planned to attend college (63% vs. 79%), had friends planning to attend college (67% vs. 83%), had parents that expected them to earn a four-year college degree (58% vs. 75%), and reported less frequent discussions about college with parents (mean = 2.51 vs. 3.06; all p < 0.05).
Technology background also varied significantly. Connected Communicators lived in households with the most electronic devices (mean = 10.11), followed by Web Browsers (9.74) and Digitally Disconnected youth (8.72; p < 0.05 for all pairwise differences). Interestingly, Web Browsers reported the highest levels of parental tech limits (mean = 1.80), significantly greater than Connected Communicators (1.60, p < 0.05).
Finally, differences emerged in young adult contexts. Digitally Disconnected youth were significantly less likely to be married or cohabiting in 2017 (29%) compared to Connected Communicators (39%, p < 0.05), and more likely to live with parents (42%) compared to Connected Communicators (28%, p < 0.05). They were also less likely to be employed (65%) relative to both Web Browsers (76%) and Connected Communicators (75%; p < 0.05). Together, these findings suggest that patterns of adolescent technology use are closely intertwined with key sociodemographic, educational, and social factors that shape young people’s college-going trajectories.

4.3. Predicting College Enrollment by Adolescent Technology Use

Table 4 presents results from regression models estimating the association between adolescent technology use and college enrollment in 2017. I first estimated a model with technology use classes as the key independent variable interacted with gender (Model 1). I then added controls for adolescent characteristics (Model 2) and young adult contexts (Model 3).
Without controls (Table 4, Model 1), boys who were Web Browsers in adolescence had significantly lower odds of later college enrollment compared to boys who were Connected Communicators (b = −0.99, p < 0.05). Whereas there is no baseline difference between Digitally Disconnected boys and Connected Communicators (b = −0.55, p > 0.05). After adjusting for adolescent background characteristics—including sociodemographic characteristics, education, and technology background—the pattern of gender difference persists. Boys in the Web Browser group continued to show significantly lower odds of college enrollment relative to Connected Communicators (b = −1.31, p < 0.05). Whereas boys in the Digitally Disconnected group now showed significantly lower odds of enrollment compared to Connected Communicators (b = −1.50, p < 0.05). In the fully adjusted model, which includes adolescent and young adult factors (e.g., marriage, parenthood, parental coresidence, employment), results remained similar. Boys who were Connected Communicators in adolescence were significantly more likely to enroll in college in young adulthood than boys in the other classes (p < 0.05).
To further illustrate these gender differences, Figure 1 displays predicted probabilities of college enrollment for girls and boys within each technology use class (based on Table 4, Model 3). Among Connected Communicators, boys had significantly higher predicted probabilities of college enrollment than girls (86% vs. 72%, p < 0.05)—a 14-percentage-point gap favoring boys—suggesting that the benefits of higher digital connectivity may be especially pronounced for boys. In contrast, boys in less connected groups (Web Browsers or Digitally Disconnected) were less likely to enroll than girls in the same groups (though gender differences were not statistically significant, p > 0.05). Among girls, predicted probabilities of college enrollment also varied by technology use class. In adolescence, girls who belonged to the Connected Communicators class had lower predicted enrollment (72%) compared to their peers in the Web Browsers and Digitally Disconnected groups (both 84%, p < 0.05). This pattern suggests that high levels of digital connectivity may not provide the same educational benefits for girls as they do for boys and may even be associated with slightly lower college-going rates among girls.

5. Discussion

Adolescence is a particularly salient period in the life course where youth begin developing early educational and career trajectories (Fitton et al. 2013). Research examining the association between adolescent behaviors and educational attainment has yet to fully attend to the increasingly digital contexts in which youth are embedded. In this study, I utilize longitudinal data from two youth-focused supplements to the PSID, capturing a cohort of digital youth from adolescence through young adulthood, to examine the association between adolescent technology use and college enrollment during the transition to adulthood. In doing so, results extend prior research on digital and educational inequalities by considering the multidimensional nature of adolescents’ technology use and the potential implications of these patterns for youths’ future college enrollment.
Results reveal three key conclusions. First, despite assumptions that digital youth are universal adopters of ICTs (Eynon 2020), findings demonstrate that today’s young adults grew up in a period with considerable diversity in their digital experiences. Latent class analysis reveals that adolescents can be categorized into three distinct classes based on their technology use. The largest group, Web Browsers, reported moderate, functional, and selective digital engagement. Their use of digital technologies, particularly for browsing websites and completing schoolwork, was routine but not intensive, and notably lower across social and communication domains. This profile reflects a more task-oriented use of technologies, consistent with patterns of “capital-enhancing” engagement.
In contrast, Connected Communicators reported high-frequency, multidimensional digital activity that spanned academic and social domains. Their near-daily engagement with email, texting, social platforms, and school-related technologies sets them apart from recent findings suggesting a trade-off between social and educational technology use in adolescence (Kastorff et al. 2023). Rather than prioritizing one over the other, Connected Communicators engaged with educational ICTs at rates comparable to, or even exceeding, those of their less socially engaged peers. Finally, Digitally Disconnected adolescents reported minimal engagement across all indicators. Their near-absence from both academic and social digital spaces suggests a disconnect not only from specific platforms but from digital participation more broadly. Together, these profiles underscore the importance of attending to the quality and context of digital engagement in adolescence.
Second, the descriptive traits of latent classes reinforce prior research on second-level digital divides (Hargittai 2010; Helsper 2021), indicating that technology use is reflective of social and academic contexts. Connected Communicators—the most digitally engaged group—had higher GPAs, greater access to household devices, and more supportive college-going environments than their peers. These findings align with work suggesting that broader digital engagement is often concentrated among more advantaged youth with access to “capital-enhancing” forms of technologies (Becker 2023; Notten and Becker 2017). Digitally Disconnected adolescents had lower levels of academic achievement, parental education, and college aspirations, and were less embedded in peer and family networks that support college enrollment. While Web Browsers were moderately engaged, they stood out for higher math self-efficacy and stricter parental controls, suggesting a more regulated, task-oriented approach to technology.
Third, the results extend prior research on digital and educational inequalities by highlighting a critical gendered third-level digital divide—showing that girls’ and boys’ patterns of technology use are linked to unequal educational outcomes. Regression results indicate that adolescent technology use is a significant predictor of college enrollment, with its association varying by gender. Among boys, being a Connected Communicator was associated with increased odds of college enrollment, even after adjusting for adolescent characteristics and young adult contexts. In contrast, girls in the same group had lower predicted enrollment than their peers in less connected classes. These findings suggest that the relationship between adolescent technology use and college enrollment is not universal, but gendered. Specifically, boys may benefit more than girls from being highly connected during adolescence, while girls’ college enrollment appears less dependent on their patterns of technology use.
These gender differences may reflect broader orientations towards social and task-based status. That is, even when boys and girls use similar technologies, they may do so for different purposes with distinct implications. For instance, research suggests that adolescent girls use digital technologies—such as social media—to build their peer-based relational status (Nesi et al. 2018), which may not directly translate into academic achievement. In contrast, because boys tend to be more task-oriented (Ridgeway 2011), boys who engage with digital technologies may develop orientations more closely aligned with academic pursuits that may support college enrollment. At the same time, digitally disconnected boys may lack not only technological resources but also the peer networks and social capital that facilitate educational transitions.
These differences are then reinforced in daily social interactions (Ridgeway 2019). Social-psychological theories suggest that in domains perceived as socially and economically valued—such as digital technologies—boys are more likely to be viewed as competent and are consequently more rewarded for their engagement (Correll 2004; Wynn and Correll 2017). Boys who were Connected Communicators may have benefited from these assumptions, with their digital practices interpreted as signs of technical aptitude or future potential. These perceptions likely contributed to greater encouragement or institutional support, bolstering their educational trajectories. In contrast, girls may not have received the same recognition or support, even with similar patterns of use. Their use of ICTs often centers on communication and social media (Tsai and Tsai 2010), domains that are commonly overlooked in educational settings (Sims 2014). Moreover, girls are more likely to underestimate their digital abilities (Christensen 2023; Hargittai and Shafer 2006), which may undermine their academic confidence or the capital they derive from their technology use. As a result, even highly connected girls may be less able to leverage their digital engagement into college-going outcomes. They may even experience their technology use as a source of distraction, stress, or social pressure.
Together, these findings underscore how third-level digital divides—differences in the outcomes of technology use—are structured by gendered expectations, social perceptions, and access to meaningful digital opportunities. While technology has the potential to support educational advancement, it does not do so evenly. Instead, the value of adolescents’ digital practices is filtered through cultural beliefs about gender and competence, shaping how adolescents engage with technologies and who benefits most from digital connectivity during this formative life stage.

Limitations

The CDS-07 and TAS-17 are uniquely well-suited to examine the long-term implications of adolescent technology use, though several data limitations should be acknowledged. First, the observed associations between adolescent technology use and college enrollment may be partially confounded by unmeasured factors—such as personality traits or motivation—that shape both digital engagement and educational outcomes. To help account for this potential selection bias, I used inverse probability weighting (IPW) to adjust for differences in adolescents’ observed characteristics that predict both latent class membership and college enrollment. While IPWs strengthens the internal validity of the findings, unobserved confounders may still bias the results. For instance, adolescents who are less interested in technology might also be less inclined to pursue higher education, regardless of their digital access. However, if this were the primary explanation, we would expect similar patterns for girls, which was not the case.
Second, the panel structure of the CDS and TAS limits the ability to observe mediating mechanisms over time. Because adolescents age in and out of the survey waves, this study cannot directly assess the intervening processes—such as ICT skill development, self-efficacy, or shifting educational aspirations—that may explain how technology use influences later outcomes. Future research should further investigate these pathways, especially those that may differ by gender.
Third, small cell sizes across key social categories—such as gender, race/ethnicity, and socioeconomic background—limit the ability to assess intersectional patterns of inequality. Future research should draw on more diverse and expansive samples to better capture the intersection of social and digital inequalities. Further, the generalizability of these findings may be constrained by the technology measures included in the CDS-07. While this study highlights important associations between adolescent technology use and college enrollment, it does not capture the complete range of digital factors influencing educational inequalities. Other dimensions of digital engagement, such as technological skill and types of online content, likely play a role but were beyond the scope of the available data. Moreover, given the rapid pace of technological innovation, the digital environment of 2007 looks notably different than that of today’s youth, underscoring the need for more updated and comprehensive measures of digital technology use.
Ultimately, this study does not aim to establish causal effects, but rather to document the associations between adolescent technology use and later educational outcomes. Though this approach reveals important insights, it cannot capture the nuanced meanings, motivations, and contextual factors that shape these relationships. Future qualitative research should investigate these mechanisms to provide a more comprehensive understanding of how digital engagement influences young adults’ educational pathways. Nonetheless, the findings underscore the lasting implications of adolescent technology use and the need to consider not only access and engagement but also variation in the returns to digital participation.

6. Conclusions

As digital technologies become increasingly embedded in educational settings, this study sheds new light on how adolescents’ engagement with ICTs is socially patterned and consequential for long-term outcomes. By 2007, access to household digital devices was widespread, yet youth continued to use these technologies in distinct and stratified ways. Using latent class analysis, this study identified three profiles of adolescent technology use—Connected Communicators, Web Browsers, and Digitally Disconnected—and revealed that their implications for college enrollment differed notably by gender. Among boys, Connected Communicators were more likely than their peers to pursue higher education, whereas among girls, those in the same class were less likely to enroll in college. These findings suggest that the educational benefits of digital engagement are not equally shared but may be more pronounced for boys than for girls. Thus, by linking adolescent technology use to later college enrollment, this study contributes to our understanding of how digital technologies may shape educational inequalities across the life course.

Funding

This research received no external funding.

Institutional Review Board Statement

Institutional review board approval was not required as authors used secondary data.

Informed Consent Statement

Not applicable.

Data Availability Statement

The Child Development Supplement (CDS) and Transition to Adulthood Supplement (TAS) to the Panel Study of Income Dynamics (PSID) data used in this article were made available by the Institute for Social Research at the University of Michigan. The data can be downloaded from https://simba.isr.umich.edu/CDS/default.aspx, accessed on 1 November 2021.

Acknowledgments

Special thanks to Rachel Goldberg, Kristin Turney, and Judy Treas for their thoughtful feedback on earlier versions of this manuscript.

Conflicts of Interest

The author declares no conflict of interest.

Notes

1
The adopted class solution had an entropy of 0.81, indicating a high level of classification certainty. This suggests that respondents were correctly classified into their latent classes approximately 81% of the time, supporting the use of modal class assignment (Clark and Muthén 2009).
2
Because latent classes capture average response patterns rather than fully homogeneous groups, the identified categories should be interpreted as probabilistic summaries of adolescent technology use, not as uniform profiles of all individuals within each class.

References

  1. Allensworth, Elaine M., and Kallie Clark. 2020. High School GPAs and ACT Scores as Predictors of College Completion: Examining Assumptions about Consistency across High Schools. Educational Researcher 49: 198–211. [Google Scholar] [CrossRef]
  2. Allison, Paul D. 2012. Handling Missing Data by Maximum Likelihood. In SAS Global Forum 2012. Paper 312. Mumbai: AS Institute Inc., pp. 1–21. [Google Scholar]
  3. Becker, Birgit. 2023. Social Disparities in Adolescents’ Educational ICT Use at Home: How Digital and Educational Inequalities Interact. In The Research Handbook on Digital Sociology. Cheltenham: Edward Elgar Publishing, pp. 293–306. [Google Scholar]
  4. Buchmann, Claudia, Rachel E. Dwyer, and Ming Yao. 2025. The deepening gender divide in credentials, 2000–2020: Continuity, change, and implications. RSF: The Russell Sage Foundation Journal of the Social Sciences 11: 154–77. [Google Scholar] [CrossRef]
  5. Bulman, George, and Robert W. Fairlie. 2016. Technology and Education: Computers, Software, and the Internet. In Handbook of the Economics of Education. Amsterdam: Elsevier, vol. 5, pp. 239–80. [Google Scholar]
  6. Christensen, MacKenzie A. 2023. Tracing the Gender Confidence Gap in Computing: A Cross-National Meta-Analysis of Gender Differences in Self-Assessed Technological Ability. Social Science Research 111: 102853. [Google Scholar] [CrossRef]
  7. Clark, Shaunna L., and Bengt Muthén. 2009. Relating Latent Class Analysis Results to Variables Not Included in the Analysis. Los Angeles: University of California. [Google Scholar]
  8. Correll, Shelley J. 2004. Constraints into Preferences: Gender, Status, and Emerging Career Aspirations. American Sociological Review 69: 93–113. [Google Scholar] [CrossRef]
  9. Desimoni, Marta, Donatella Papa, Cristina Lasorsa, Michela Milioni, and Rosalba Ceravolo. 2024. Computer User Profiles in Early Adolescence and Digitally Assessed Mathematics: A Latent Class Analysis. Computers in Human Behavior Reports 13: 100369. [Google Scholar] [CrossRef]
  10. DiMaggio, Paul, Eszter Hargittai, Coral Celeste, and Steven Shafer. 2004. Digital Inequality: From Unequal Access to Differentiated Use. In Social Inequality. New York: Russell Sage Foundation, pp. 355–400. [Google Scholar]
  11. Domina, Thurston, AnneMarie Conley, and George Farkas. 2011. The Link Between Educational Expectations and Effort in the College-for-All Era. Sociology of Education 84: 93–112. [Google Scholar] [CrossRef]
  12. Eynon, Rebecca. 2020. The Myth of the Digital Native: Why It Persists and the Harm It Inflicts. In Education in the Digital Age: Healthy and Happy Children. Edited by Tracey Burns and Francesca Gottschalk. Paris: OECD Publishing. [Google Scholar]
  13. Faverio, Michelle, and Olivia Sidoti. 2024. Teens, Social Media and Technology 2024. Pew Research Center, December 12. [Google Scholar]
  14. Fitton, Vanessa A., Brian K. Ahmedani, R. Dean Harold, and Edward D. Shifflet. 2013. The Role of Technology on Young Adolescent Development: Implications for Policy, Research, and Practice. Child and Adolescent Social Work Journal 30: 399–413. [Google Scholar] [CrossRef]
  15. Fraillon, Julian, John Ainley, Wolfram Schulz, Tim Friedman, and Eveline Gebhardt. 2014. Preparing for Life in a Digital Age: The IEA International Computer and Information Literacy Study International Report. Berlin and Heidelberg: Springer Nature. [Google Scholar]
  16. Hargittai, Eszter. 2010. Digital Na(t)ives? Variation in Internet Skills and Uses among Members of the “Net Generation”. Sociological Inquiry 80: 92–113. [Google Scholar] [CrossRef]
  17. Hargittai, Eszter, and Steven Shafer. 2006. Differences in Actual and Perceived Online Skills: The Role of Gender. Social Science Quarterly 87: 432–48. [Google Scholar] [CrossRef]
  18. Helsper, Ellen. 2010. Gendered Internet Use Across Generations and Life Stages. Communication Research 37: 352–74. [Google Scholar] [CrossRef]
  19. Helsper, Ellen. 2021. The Digital Disconnect: The Social Causes and Consequences of Digital Inequalities. London: SAGE Publications. [Google Scholar]
  20. Ito, Mizuko. 2013. Hanging Out, Messing Around, and Geeking Out: Kids Living and Learning with New Media. Cambridge: MIT Press. [Google Scholar]
  21. Jackson, Linda A., Yong Zhao, Alex Kolenic, Hiram E. Fitzgerald, R. Dean Harold, and Alois von Eye. 2008. Race, Gender, and Information Technology Use: The New Digital Divide. Cyberpsychology & Behavior 11: 437–42. [Google Scholar]
  22. Junco, Reynol, and Shelia R. Cotten. 2012. No A 4 U: The Relationship Between Multitasking and Academic Performance. Computers & Education 59: 505–14. [Google Scholar]
  23. Kastorff, Tobias, Michael Sailer, and Katja Stegmann. 2023. A Typology of Adolescents’ Technology Use Before and During the COVID-19 Pandemic: A Latent Profile Analysis. International Journal of Educational Research 117: 102136. [Google Scholar] [CrossRef]
  24. Kirschner, Paul A., and Pedro De Bruyckere. 2017. The Myths of the Digital Native and the Multi-tasker. Teaching and Teacher Education 67: 135–42. [Google Scholar] [CrossRef]
  25. Lei, Jing, and Yong Zhao. 2007. Technology Uses and Student Achievement: A Longitudinal Study. Computers & Education 49: 284–96. [Google Scholar] [CrossRef]
  26. Lippman, Laura, Lina Guzman, Julie Dombrowski Keith, Akemi Kinukawa, Rebecca Shwalb, and Peter Tice. 2008. Parent Expectations and Planning for College. Statistical Analysis Report. NCES 2008-079. Washington, DC: National Center for Education Statistics. [Google Scholar]
  27. Livingstone, Sonia, Dafna Lemish, Sun Sun Lim, Monica Bulger, Patricio Cabello, Magdalena Claro, Tania Cabello-Hutt, Joe Khalil, Kristiina Kumpulainen, Usha S. Nayar, and et al. 2017. Global Perspectives on Children’s Digital Opportunities: An Emerging Research and Policy Agenda. Pediatrics 140: S137–S141. [Google Scholar] [CrossRef] [PubMed]
  28. May, Kevin E., and Andrew D. Elder. 2018. Efficient, Helpful, or Distracting? A Literature Review of Media Multi-tasking in Relation to Academic Performance. International Journal of Educational Technology in Higher Education 15: 13. [Google Scholar] [CrossRef]
  29. McGonagle, Katherine A., Robert F. Schoeni, Narayan Sastry, and Vicki A. Freedman. 2012. The Panel Study of Income Dynamics: Overview, recent innovations, and potential for life course research. Longitudinal and Life Course Studies 3: 188. [Google Scholar] [CrossRef]
  30. Mollborn, Stefanie, Paula Fomby, Jessica A. Goode, and Anna Modile. 2021. A Life Course Framework for Understanding Digital Technology Use in the Transition to Adulthood. Advances in Life Course Research 47: 100379. [Google Scholar] [CrossRef]
  31. Monaghan, David B. 2021. Predictors of College Enrollment Across the Life Course: Heterogeneity by Age and Gender. Education Sciences 11: 344. [Google Scholar] [CrossRef]
  32. Morgan, Stephen L. 2005. On the Edge of Commitment: Educational Attainment and Race in the United States. Stanford: Stanford University Press. [Google Scholar]
  33. Morgan, Stephen L., Dafna Gelbgiser, and Kim A. Weeden. 2013. Feeding the Pipeline: Gender, Occupational Plans, and College Major Selection. Social Science Research 42: 989–1005. [Google Scholar] [CrossRef] [PubMed]
  34. Musick, Kelly, Jennie E. Brand, and Dwight Davis. 2012. Variation in the Relationship Between Education and Marriage: Marriage Market Mismatch? Journal of Marriage and Family 74: 53–69. [Google Scholar] [CrossRef]
  35. Nesi, Jaclyn, Sophia Choukas-Bradley, and Mitchell J. Prinstein. 2018. Transformation of adolescent peer relations in the social media context: Part 2—Application to peer group processes and future directions for research. Clinical Child and Family Psychology Review 21: 295–319. [Google Scholar] [CrossRef] [PubMed]
  36. Notten, Natascha, and Birgit Becker. 2017. Early Home Literacy and Adolescents’ Online Reading Behavior in Comparative Perspective. International Journal of Comparative Sociology 58: 475–93. [Google Scholar] [CrossRef]
  37. Petko, Dominik, Adrian Cantieni, and Daniel Prasse. 2017. Perceived Quality of Educational Technology Matters: A Secondary Analysis of Students’ ICT Use, ICT-Related Attitudes, and PISA 2012 Test Scores. Journal of Educational Computing Research 54: 1070–91. [Google Scholar] [CrossRef]
  38. Rafalow, Matthew H. 2018. Disciplining Play: Digital Youth Culture as Capital at School. American Journal of Sociology 123: 1416–52. [Google Scholar] [CrossRef]
  39. Reber, Sarah, and Emily Smith. 2023. College Enrollment Disparities: Understanding the Role of Academic Preparation. Washington, DC: Brookings. [Google Scholar]
  40. Ridgeway, Cecilia L. 2011. Framed by Gender: How Gender Inequality Persists in the Modern World. Oxford: Oxford University Press. [Google Scholar]
  41. Ridgeway, Cecilia L. 2019. Status: Why Is It Everywhere? Why Does It Matter? New York: Russell Sage Foundation. [Google Scholar]
  42. Riegle-Crumb, Catherine. 2010. More Girls Go to College: Exploring the Social and Academic Factors Behind the Female Postsecondary Advantage Among Hispanic and White Students. Research in Higher Education 51: 573–93. [Google Scholar] [CrossRef]
  43. Robinson, Laura, Øyvind Wiborg, and Jan Schulz. 2018. Interlocking Inequalities: Digital Stratification Meets Academic Stratification. American Behavioral Scientist 62: 1251–72. [Google Scholar] [CrossRef]
  44. Rosén, Monica, and Jan-Eric Gustafsson. 2016. Is Computer Availability at Home Causally Related to Reading Achievement in Grade 4? A Longitudinal Difference in Differences Approach to IEA Data from 1991 to 2006. Large-Scale Assessments in Education 4: 5. [Google Scholar] [CrossRef]
  45. Ryan, Allison M. 2001. The Peer Group as a Context for the Development of Young Adolescent Motivation and Achievement. Child Development 72: 1135–50. [Google Scholar] [CrossRef]
  46. Schuler, Megan S., Jeannie-Marie S. Leoutsakos, and Elizabeth A. Stuart. 2014. Addressing Confounding When Estimating the Effects of Latent Classes on a Distal Outcome. Health Services and Outcomes Research Methodology 14: 232–54. [Google Scholar] [CrossRef]
  47. Sims, Christo. 2014. From Differentiated Use to Differentiating Practices: Negotiating Legitimate Participation and the Production of Privileged Identities. Information, Communication & Society 17: 670–82. [Google Scholar]
  48. Skryabin, Maxim, Jie Zhang, Lili Liu, and De Zhang. 2015. How the ICT Development Level and Usage Influence Student Achievement in Reading, Mathematics, and Science. Computers & Education 85: 49–58. [Google Scholar] [CrossRef]
  49. Stage, Frances K., and Don Hossler. 1989. Differences in Family Influences on College Attendance Plans for Male and Female Ninth Graders. Research in Higher Education 30: 301–16. [Google Scholar] [CrossRef]
  50. Touloupis, Thanos, and Michael Campbell. 2024. The role of academic context-related factors and problematic social media use in academic procrastination: A cross-sectional study of students in elementary, secondary, and tertiary education. Social Psychology of Education 27: 175–214. [Google Scholar] [CrossRef]
  51. Tsai, Ming-Je, and Chin-Chung Tsai. 2010. Junior High School Students’ Internet Usage and Self-Efficacy: A Re-Examination of the Gender Gap. Computers & Education 54: 1182–92. [Google Scholar]
  52. van Deursen, Alexander J., and Jan A. van Dijk. 2014. The Digital Divide Shifts to Differences in Usage. New Media & Society 16: 507–26. [Google Scholar]
  53. Van Dijk, Jan. 2020. The Digital Divide. Hoboken: John Wiley & Sons. [Google Scholar]
  54. Wang, Fang, Xiaoli Ni, Mengzhu Zhang, and Jingjie Zhang. 2024. Educational Digital Inequality: A Meta-Analysis of the Relationship Between Digital Device Use and Academic Performance in Adolescents. Computers & Education 213: 105003. [Google Scholar] [CrossRef]
  55. Wasserman, Ira M., and Marie Richmond-Abbott. 2005. Gender and the Internet: Causes of Variation in Access, Level, and Scope of Use. Social Science Quarterly 86: 252–70. [Google Scholar] [CrossRef]
  56. Wynn, Alison T., and Shelley J. Correll. 2017. Gendered Perceptions of Cultural and Skill Alignment in Technology Companies. Social Sciences 6: 45. [Google Scholar] [CrossRef]
  57. Xiao, Feiya, and Li Sun. 2022. Profiles of Student ICT Use and Their Relations to Background, Motivational Factors, and Academic Achievement. Journal of Research on Technology in Education 54: 456–72. [Google Scholar] [CrossRef]
  58. Zhang, Danhui, and Luman Liu. 2016. How Does ICT Use Influence Students’ Achievements in Math and Science over Time? Evidence from PISA 2000 to 2012. Eurasia Journal of Mathematics, Science and Technology Education 12: 2431–49. [Google Scholar] [CrossRef]
Socsci 14 00576 g001
Figure 1. Predicted probabilities of college enrollment by girls’ and boys’ technology use classes (n = 1164). Note: Predicted probabilities based on Model 3 of Table 4 (and averaged across 20 imputed datasets). All values were held constant at their mean. * p < 0.05 for differences between girls and boys.
Figure 1. Predicted probabilities of college enrollment by girls’ and boys’ technology use classes (n = 1164). Note: Predicted probabilities based on Model 3 of Table 4 (and averaged across 20 imputed datasets). All values were held constant at their mean. * p < 0.05 for differences between girls and boys.
Socsci 14 00576 g001
Table 1. Latent class model fit statistics (n = 1164).
Table 1. Latent class model fit statistics (n = 1164).
Solution Log LikelihooddfG2AICBIC
Class 1−8452.7516p = 1.00016,937.5017,017.49
Class 2−7728.32133p = 1.00015,522.6415,687.62
Class 3−7568.4750p = 1.00015,236.9415,486.91
Note: Class 4 did not converge. AIC = Akaike information criterion; BIC = Bayesian information criterion.
Table 2. Item-response probabilities conditional on class membership.
Table 2. Item-response probabilities conditional on class membership.
Class 1Class 2Class 3
MeasuresMeanWeb Browsers Connected CommunicatorsDigitally Disconnected
Emailing
Near-daily0.240.110.530.00
Weekly 0.230.290.280.03
Almost never 0.520.610.190.97
Chat rooms
Near-daily0.160.030.410.00
Weekly 0.150.190.180.01
Almost never 0.690.790.410.99
Texting
Near-daily0.290.200.490.11
Weekly 0.060.050.070.05
Almost never 0.650.750.430.84
Posting personal info
Near-daily0.150.000.410.00
Weekly 0.150.160.220.01
Almost never 0.700.840.370.99
Playing games
Near-daily0.270.330.240.21
Weekly 0.260.300.230.26
Almost never 0.460.380.540.53
Browse websites
Near-daily0.510.310.990.07
Weekly 0.290.620.010.11
Almost never 0.190.070.000.83
School work
Near-daily0.240.250.340.01
Weekly 0.370.410.450.13
Almost never 0.390.340.200.86
Other activities
Near-daily0.510.350.960.04
Weekly 0.290.560.040.12
Almost never 0.200.090.010.85
N1164467411286
Source: Child Development Supplement 2007.
Table 3. Descriptive statistics by adolescent technology use classes.
Table 3. Descriptive statistics by adolescent technology use classes.
Full SampleAdolescent Technology Use Classes
MeasuresMeanSEWeb BrowsersConnected CommunicatorsDigitally Disconnected
College enrollment (TAS-17)0.73 0.740.77 c0.64 a
Adolescent characteristics (CDS-07)
Boy 0.51 0.510.44 c0.60 a
Age (in years)13.41(0.09)13.03 a14.17 bc12.99 a
Race/ethnicity
Non-Hispanic white 0.60 0.600.65 c0.54 a
Non-Hispanic Black 0.15 0.150.150.16
Hispanic 0.17 0.180.140.21
Other/multiracial0.05 0.040.050.06
Parental years of education14.03(0.23)14.0114.45 c13.49 a
Family income (logged) 11.01(0.05)10.91 a11.19 bc10.89 a
Educational background (CDS-07)
High school GPA (TAS-17)2.97(0.06)3.03 c3.12 c2.68 ab
Math self-efficacy 4.81(0.04)4.91 a4.71 b4.76
Reading self-efficacy 5.05(0.04)5.095.074.96
Hours spent on homework1.89(0.11)2.11 a1.58 b1.93
Child plans to go to college0.71 0.70 a0.79 bc0.63 a
Friends plan to go to college 0.77 0.780.83 b0.67 a
Parents expect a college degree0.69 0.70 c0.75 c0.58 ab
Discuss college with parents 2.76(0.06)2.68 a3.06 bc2.51 a
Technology background (CDS-07)
Number of household electronics 9.61(0.21)9.74 c10.11 c8.72 ab
Parental tech limits1.69(0.05)1.80 a1.60 b1.64
Parental tech encouragement 0.92(0.04)0.950.930.86
Young adult contexts (TAS-2017)
Married or cohabiting 0.33 0.29 a0.39 bc0.29 a
Has at least one child0.20 0.160.230.22
Parental coresidence 0.35 0.38 a0.28 bc0.42 a
Employed (full or part-time)0.73 0.76 c0.75 c0.65 ab
N1164 467411286
Note: Results based on survey-weighted, imputed data. Standard errors are in parentheses. Superscripts indicate significance at p < 0.05 from: a Connected Communicators, b Web Browsers, c Digitally Disconnected. Source: Panel Study of Income Dynamics, 2007 Child Development Supplement (CDS-07); 2017 Transition to Adulthood Supplement (TAS-17).
Table 4. Logistic regression models estimating the relationship between latent classes of technology use and girls’ and boys’ college enrollment (n = 1164).
Table 4. Logistic regression models estimating the relationship between latent classes of technology use and girls’ and boys’ college enrollment (n = 1164).
Model 1Model 2Model 3
+ Adolescent Characteristics+ Young Adult Contexts
Technology use classes (ref = Connected Communicators)
Web Browsers0.62 *(0.29) 0.68 +(0.38) 0.68 +(0.39)
Digitally Disconnected 0.13(0.41) 0.66 +(0.38) 0.66(0.41)
Boy (ref = girl)0.19(0.36) 0.90 *(0.36) 0.82 *(0.37)
Technology classes × Gender
Web Browsers × Boys−0.99 *(0.47) −1.31 *(0.53) −1.40 *(0.56)
Digitally Disconnected × Boys −0.55(0.59) −1.50 *(0.63) −1.50 *(0.63)
Adolescent characteristics (CDS-07)
Age (in years) −0.03(0.07) 0.03(0.07)
Race/ethnicity (ref = non-Hispanic white)
Non-Hispanic Black −0.51 +(0.30) −0.67 *(0.30)
Hispanic 0.70 +(0.38) 0.67 +(0.39)
Non-Hispanic other 0.21(0.56) 0.09(0.57)
Parental years of education 0.06(0.05) 0.05(0.04)
Family income (logged) 0.06(0.12) 0.05(0.13)
Educational background (CDS-07)
High school GPA 0.58 ***(0.13) 0.56 ***(0.14)
Math self-efficacy 0.15(0.14) 0.17(0.14)
Reading self-efficacy 0.07(0.14) 0.05(0.14)
Hours spent on homework 0.02(0.03) 0.02(0.04)
Child plans to go to college 0.69 *(0.33) 0.61 +(0.34)
Friends plan to go to college 0.39(0.25) 0.39(0.26)
Parents expect a college degree 1.26 ***(0.26) 1.24 ***(0.26)
Discuss college with parents −0.04(0.09) −0.04(0.09)
Technology background (CDS-07)
Number of electronic devices 0.02(0.04) 0.01(0.03)
Parental tech limits 0.07(0.12) 0.03(0.12)
Parental tech encouragement 0.22(0.17) 0.30 +(0.16)
Young adult contexts (TAS-2017)
Married or cohabiting −0.71 *(0.28)
Has at least one child −0.52 +(0.26)
Parental coresidence −0.49 +(0.25)
Employed (full or part-time) −0.23(0.23)
Constant 0.87 ***(0.18) −5.16 *(2.16) −4.67 *(2.21)
Notes: Results present regression coefficients with standard errors in parentheses. All models were weighted using survey and inverse probability weighting. *** p < 0.001, * p < 0.05, + p < 0.01. Source: Panel Study of Income Dynamics, 2007 Child Development Supplement (CDS-07); 2017 Transition to Adulthood Supplement (TAS-17).
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Christensen, M.A. From Screens to Schooling: Associations Between Adolescent Technology Use and Gendered College Enrollment. Soc. Sci. 2025, 14, 576. https://doi.org/10.3390/socsci14100576

AMA Style

Christensen MA. From Screens to Schooling: Associations Between Adolescent Technology Use and Gendered College Enrollment. Social Sciences. 2025; 14(10):576. https://doi.org/10.3390/socsci14100576

Chicago/Turabian Style

Christensen, MacKenzie A. 2025. "From Screens to Schooling: Associations Between Adolescent Technology Use and Gendered College Enrollment" Social Sciences 14, no. 10: 576. https://doi.org/10.3390/socsci14100576

APA Style

Christensen, M. A. (2025). From Screens to Schooling: Associations Between Adolescent Technology Use and Gendered College Enrollment. Social Sciences, 14(10), 576. https://doi.org/10.3390/socsci14100576

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop