First-level fundamentals: computer ownership is more important for internet benefits than in-home internet service

Abstract

Although computers, smartphones, and internet service are essential 21st century tools, digital equity stakeholders often focus on in-home internet service in their advocacy. Yet few scholars have compared the associated benefits of different fundamental digital resources. In response, a cross-sectional analysis of two waves of U.S. census data (2020, N = 84,206; 2023, N = 82,941) revealed that computer use (laptop/desktop) was consistently a stronger predictor of beneficial internet use (e.g., job searches, government resources, and eHealth) than smartphone access or in-home internet service, and device quality explained additional model variance. Moreover, computer use without in-home internet was more beneficial than in-home internet without computer use, and in-home internet was especially important in homes with smartphones. These findings add nuance to resources and appropriation theory and the technology maintenance framework, and underscore that internet and devices are both important, requiring policies that facilitate access to both.

There is no question that the internet has radically transformed human communication over the past 30 years. This truism is apparent in the questions that communication scholars ask, the theories they develop, and the methods that they use. This fact has extended far beyond the academic sphere, altering the way people work, attend school, and access healthcare. Yet focus on the internet as the driving factor in a communication revolution may obscure other essential components of being online. Computing devices (e.g., laptops, smartphones, tablets, etc.), social and technical infrastructure, knowledge resources, and ancillary tools (e.g., printers, routers) are all critical to benefiting from what the internet has to offer. Insufficient access to a full range of these resources impedes access to nearly all forms of private, public, and civil society today (Helsper & van Deursen, 2015; van Deursen & Helsper, 2018; Van Dijk, 2020).

A refrain amongst those dedicated to reducing the digital divide, or the “division between people who have access and use of digital media and those who do not” (Van Dijk, 2020, p. 1), is that to achieve digital equity, resources must extend beyond the provision of basic internet service. Digital equity refers to the “condition in which all individuals and communities have the information technology capacity needed for full participation in our society” (NDIA, 2024). Mnemonic labels (e.g., Assistance, Broadband Connectivity, and Devices as the ABCDs of digital equity; Katz, 2020) and bulleted lists (e.g., Siefer & Tesfaye, 2022) underscore a holistic, long-term approach to mitigating digital inequalities (Katz & Gonzalez, 2016; Reisdorf et al., 2022; Robinson et al., 2015, 2020; Van Dijk, 2020). In practice, however, it is much easier for government funding to target a single aspect of digital equity on a limited basis. In that regard, a review of federal digital equity legislation over the last thirty years finds that the majority of bills have focused on broadband related initiatives (King & Gonzales, 2023), suggesting that internet use has long been a defining criterion for closing the digital divide.

Although in-home broadband access is certainly essential for digital equity, we argue that other digital tools, such as the devices themselves, are equally essential. To better understand the relative benefits of digital resources we conducted a straightforward, albeit somewhat unconventional analysis of the two most recent waves of the Current Population Survey (CPS) and Computer and Internet Use Supplement, incorporating a first-time measure of device quality. Our findings revealed that household access to in-home internet service, a computer (i.e., desktop or laptop), and a smartphone were all positively associated with a variety of beneficial internet activities (e.g., job- and health-related uses), but the relationship between computer access and outcomes was typically stronger than that of in-home internet access or smartphone access; and the quality of those devices explained additional variance in most models. Probing further we found that computers were generally associated with beneficial internet activities even when people lacked internet at home, whereas the opposite was generally true for smartphones. We discuss possible explanations for these findings and provide recommendations for subsequent research in this area. In the short term these findings refine the resources and appropriation theory of the digital divide (Van Dijk, 2005, 2020) and also lend support for a technology maintenance approach to digital equity (Gonzales, 2014; Gonzales et al., 2016), which argues that access to high quality large-screen computing devices is needed in conjunction with in-home internet to optimize user benefits from the internet.

The role of first-level factors

A guiding framework for understanding digital equity is van Dijk’s resources and appropriations theory (R&AT; Van Dijk, 2005, 2020). In this model, van Dijk describes how various demographic factors, both personal (e.g., age, health, and sex) and positional (e.g., education, employment, and geographic location), predict resources that ultimately shape an individual’s experience of digital access and use. In the model, the concept of access is divided into “first-level divide” issues, typically measured in terms of physical access, and “second-level divide” issues, typically measured in terms of skill. In this article, we will focus on physical access to technology resources, which is associated with a range of economic, political, and social outcomes (Helsper, 2021; van Deursen & Helsper, 2018; Van Dijk, 2020), and in many ways is the heart of popular understanding of digital equity. This has become particularly salient following the coronavirus disease 2019 (COVID-19) pandemic, which resulted in the federal government passing a $3.2 billion dollar Emergency Broadband Benefit (Universal Service Administrative Company, 2022), and later, an unprecedented budget of $65 billion to improve internet service nationwide (Internet for All, n.d.).

The recent influx of funding from the U.S. federal government signals a heightened awareness of the need for greater digital equity among policymakers, but scholars have long understood the value of basic, physical information and communication technology (ICT) resources. Digital divide research has demonstrated a range of benefits associated with ICT access and use, encompassing economic, political, and personal domains, such as health and education (Helsper & van Deursen, 2015; van Deursen & Helsper, 2018; Van Dijk, 2020). As one key example, studies have looked at the contributions of computer use and internet access to educational success, spanning from elementary school to college (Hampton et al., 2021; Katz et al., 2021; Schmitt & Wadsworth, 2006). ICTs are also beneficial for job seekers: Access to subsidized internet service is associated with reduced unemployment within a region (Zuo & Kolliner, 2021), and internet use is associated with increased earnings (DiMaggio & Bonikowski, 2008). Finally, ICTs have also become a critical part of contemporary healthcare. One of the most common uses of the internet is for seeking health information (Wang et al., 2021), and ICTs are commonly used to communicate directly with doctors and providers (Mold et al., 2019), especially in the form of telehealthcare (Snoswell et al., 2021). Electronic medical records are another vital digital health tool, though uptake is sometimes varied (Abd-alrazaq et al., 2019). In short, an enormous body of digital divide scholarship illustrates that across many sectors, reliable physical and material ICT access has become an indispensable means of acquiring services and staying socially connected.

We pose the first hypothesis as further validation of claims by the resources and appropriation model that physical access to technology is associated with quality-of-life outcomes of various forms. Census data assess internet use related to employment, government services, healthcare, entertainment, and personal finance. Although these are all important for quality-of-life, to necessarily narrow the scope of the article, we examined the following categories as outcomes particularly consequential for socio-economically disadvantaged Americans:

H1: In-home internet access, computer access, and smartphone access will be positively associated with beneficial internet uses, including (a) online job searches, (b) use of the internet for government resources, (c) health information seeking, (d) patient-provider communication, and (e) accessing electronic medical records.

The relative benefits of ICT

As noted, advocates of digital equity emphasize that one must have stable access to a range of digital resources to function in today’s ICT-dependent society (Katz, 2020; Robinson et al., 2020; Siefer & Tesfaye, 2022; van Deursen & van Dijk, 2019; Van Dijk, 2020). For this reason, perhaps, researchers rarely compare the benefits associated with essential digital resources. However, in a world of limited time and money, it is worth considering digital resources in isolation and in combination.

To our knowledge, no research directly compares the benefits of having home internet service with having a personal computer, presumably because these resources are ideally used in tandem. Yet given the expense and fragility associated with these tools, individuals must sometimes choose between them or are forced to live for periods with one and not the other. There is, in contrast, quite a bit of research on standalone smartphone access, or what some have called mobile-only or smartphone-only users—those who do not have in-home internet service (see “Mobile Fact Sheet”; Pew Research Center, 2021). Given the substantial expense of the combined costs of monthly home internet service, a computer, and a smartphone, many choose to forgo the first two to consolidate costs. While this approach affords flexibility in when and where people get online, mobile dependence can limit the breadth of use and skills learned (Correa et al., 2018, 2022; Napoli & Obar, 2014; Pearce & Rice, 2013). This body of work may thus offer insights into research questions on the relative importance of different forms of physical ICT access.

The benefits of smartphones

There is mixed evidence benefits of smartphones only internet service. Smartphones are an invaluable tool for staying connected to social support networks, accessing healthcare service, and acquiring various forms of economic, social, and cultural resources, especially for vulnerable populations (e.g., Calvo et al., 2019; Ling, 2022; Read et al., 2021). Studies have also found that smartphones play a critical role in fostering civic participation (Mossberger et al., 2016), especially for youth, who may be drawn to the civic laboratories of mobile-based social media (Lane et al., 2018). In addition to being a relatively cost-effective way to get online, the convenience of smartphones makes them a popular computing choice for younger generations across all income groups (Pew Research Center, 2021).

Yet research also indicates that smartphones are an insufficient replacement for computers. Although some mobile internet may not change the frequency of internet use (Wang & Liu, 2018), reliance on smartphones for computing may limit the breadth of internet use and, because of this fact, constrain the development of important digital literacy and skills (Correa et al., 2018, 2022; Napoli & Obar, 2014; Pearce & Rice, 2013). Completing lengthy forms on a smartphone, such as job applications or trying to complete homework assignments, can be difficult and impede use (Gershon & Gonzales, 2021; Katz, 2017; Katz et al., 2021). Indeed, mobile-only households tend to be disproportionately under-resourced in a variety of ways (Pew Research Center, 2021), underscoring the fact that being mobile-only is often a function of inequality and may further perpetuate those inequalities.

Ideally people would possess in-home internet, a device, and a smartphone, but that is not always an option for many. Although a holistic repertoire of digital resources best positions users to optimize internet benefits (e.g., Correa et al., 2022; Katz & Gonzalez, 2016; Pearce & Rice, 2013; Reisdorf et al., 2022; Robinson et al., 2015; van Deursen & van Dijk, 2019), it is yet unclear how the benefits of these core ICT resources compare to one another. Disentangling relative benefits of different fundamental “first-level” digital equity resources (Van Dijk, 2005, 2020) contributes to the refinement of this theoretical construct, and may also inform the focus of advocacy or policymaking when stakeholders are deciding what services and subsidies to prioritize. We thus pose the following research question followed by a moderating hypothesis:

RQ1: Which of the three traditional measures of physical ICT access (in-home internet, computer, and smartphone access) has the greatest positive association with beneficial internet uses, including (a) online job searches, (b) use of the internet for government resources, (c) health information seeking, (d) patient-provider communication, and (e) accessing electronic medical records?

H2: Regardless of findings in RQ1, the association of (a) computer access and (b) smartphone access with beneficial internet uses will be stronger when individuals have in-home internet access compared to when they do not.

The Importance of Device Quality

In addition to examining the relationship between traditional measures of physical ICT access and beneficial internet uses associated with quality of life, we further extend the scope of our contribution by exploring the implications of device satisfaction. The nuances of digital disconnection, or “under connectedness” (Katz, 2017), manifest in various ways, including but not limited to: slow data speeds, data caps, device sharing, spotty internet connections, or devices that are old and unequipped to handle the data-rich tasks of a post-COVID society. This understanding of digital equity as a holistic, ongoing effort is described by the technology maintenance framework (Gonzales, 2014; Gonzales et al., 2016), which states that “as the poor increasingly achieve in-home and public access to digital technology, they will struggle to maintain that access” (Gonzales et al., 2016, p. 1423). According to technology maintenance, because low-income households do not have enough savings to repair or replace ICTs when they inevitably malfunction these households cycle through periods of intermittent disruption, or dependable instability. Scholarship in this arena underscores the need for holistic approaches to measuring digital disparities to capture these periodic disruptions (Katz, 2017; Reisdorf & Groselj, 2017; van Deursen & van Dijk, 2019). And though there have been important exceptions in which researchers delve into the nuances of intermittent computing disruptions (e.g., Goedhart et al., 2019; Katz, 2017; Matthews & Ali, 2022; van Deursen & van Dijk, 2019), quantitative surveys often focus on dichotomous measures of ICT access, which tend to imply a more permanent, fixed state of access. In a turn from this, the most recent wave of data from the 2021 CPS and Computer and Internet Use module includes an item measuring device quality by asking how well devices “work overall” on a scale of 1 (not at all) to 5 (perfectly). Including such a measure alongside more traditional, dichotomous measures of ICT access may help further predict beneficial uses of the internet. We thus pose a final hypothesis:

H3: Quality of device access will explain additional variance in beneficial internet uses, including (a) online job searches, (b) use of the internet for government resources, (c) health information seeking, (d) patient-provider communication, and (e) accessing electronic medical records, after controlling for traditional measures of physical ICT access.

Methods

Data source

The data analysis was obtained from the CPS and Computer and Internet Use Supplement. Sponsored by the U.S. Census Bureau and the U.S. Bureau of Labor Statistics, the CPS is an open source of official government statistics on the American labor force. The CPS is conducted monthly with probability sampling of approximately 60,000 eligible households, and the Computer and Internet Use Supplement was first administered in 2000, with November 2021 and 2023 being the most recent waves of data collection. The data consist of two parts: the CPS labor force data and supplement data. The 2021 CPS dataset is a unique, and arguably conservative, test of these hypotheses as they occurred during COVID-19 when in-home internet use was at its peak. For this reason, we have also included the more recent 2023 CPS dataset. We focused our analysis exclusively on adults, as the supplement survey was administered only to individuals aged 15 and older. See Supplementary Appendix A for descriptive data on age, ethnicity, disability, citizenship, race, education, income, and labor status for both waves of data. All of these sociodemographic variables were controlled for in the subsequent analysis.

Measures

In-home internet access

The respondents were asked the following: “I am going to read a list of ways that people access the Internet from their homes, other than a mobile data plan.¹ At home, (do you/does anyone in this household) access the Internet using: (a) High-speed Internet service installed at home, such as cable, DSL, or fiber optic service, (b) Satellite Internet service, (c) Dial-up service (used in 2021 only), (d) Some other service? (note: Dial-Up service was dropped from the 2023 survey)” where each response was on a binary scale (1 = Yes; 2 = No), which was reversed (0 = No; 1 = Yes) to reflect the presence of household home internet access. To calculate in-home internet access, we totaled all positive responses to #1–3. In total, 65,723 (78.1%) and 66,529 (80.4%) respondents reported internet service at home in the 2021 and 2023 data, respectively.

Computer access

Computer access was a binary variable (0 = No; 1 = Yes) computed to indicate whether individuals had household access to personal computers based on their answers to “(Do you/Does anyone in this household, including you) use a desktop computer?” (1 = Yes; 2 = No) and “What about a laptop or notebook? (Do you/Does anyone in this household) use a laptop or notebook computer?” (1 = Yes; 2 = No). For computer access, 63,620 (75.6%) and 62,816 (75.9%) respondents reported household access to desktop computers or laptops in the 2021 and 2023 data, respectively.

Smartphone access

The respondents were interviewed with a binary question, “What about a smartphone, or a cell phone that connects to the Internet? (Do you/Does anyone in this household) use a smartphone?” (1 = Yes; 2 = No), which was also reversed (0 = No; 1 = Yes) to reflect the presence of household smartphone access. In total 70,452 (83.7%) and 71,666 (86.6%) respondents reported household access to smartphones in the 2021 and 2023 data, respectively.

Device quality

Device quality was measured with “How well do the computers and other Internet-connected devices [you use/used by this household] work overall?” (1 = They do not work at all; 5 = They work perfectly), and it was treated as a continuous measure in the analyses (M = 4.07, SD = 0.75 for 2021, and M = 4.09, SD = 0.72 for 2023). Among 74,352 valid observations of access quality, 9,969 (13.4%) respondents reported low or moderate levels (i.e., 1–3) of quality on this 5-level scale in the 2021 data, and out of 74,678 valid observations, 8,588 (11.5%) respondents reported low or moderate levels in the 2023 data.

Online job searches

The respondents were asked “In the past six months, (have you/has NAME) used the Internet to search or apply for a job? (Do you/Does NAME) use the Internet to search or apply for a job?” (1 = Yes; 2 = No), which was recorded (0 = No; 1 = Yes) to indicate whether household members used the internet for job searches. Out of 36,559 respondents, 6,484 (17.7%) engaged in this online activity in 2021; out of 36,648 respondents, 6,167 (16.8%) engaged in this online activity in 2023.

Online government resources

The respondents were asked “What about accessing government services, such as registering to vote or renewing your driver’s license?” (1 = Yes; 2 = No), which was reversed (0 = No; 1 = Yes) to indicate whether household members accessed government resources online.

Out of 36,559 respondents, 13,868 (37.9%) engaged in this online activity in 2021; out of 36,648 respondents, 14,436 (39.4%) engaged in this online activity in 2023.

Online health activities

Online health activities included three questions which were each used as separate dependent variables: “Do you/Does anyone in this household) research health information online, such as with WebMD or similar services?”, “(Do/Does) (you/anyone in this household, including you) communicate with a doctor or other health professional using the Internet?”, and “What about accessing health records or health insurance records online. (Do you/Does anyone in this household) access health records or health insurance records online?” Their responses were again reversed (0 = No; 1 = Yes). In the 2021 data, 35,897 (49.2%) reported internet uses for health information seeking, 39,651 (54.3%) reported internet uses for patient-provider communication, and 39,737 (54.5%) reported internet uses accessing electronic medical records out of 72,974 respondents. In the 2023 data, 31,776 (43.2%) reported internet uses for health information seeking, 38,285 (52.1%) reported internet uses for patient-provider communication, and 44,641 (60.7%) reported internet uses accessing electronic medical records out of 73,484 respondents.

Analysis

Statistical analysis was conducted using IBM SPSS version 28.0. Given the binary nature of the five outcome variables, hierarchical logistic regression was employed. Sociodemographic variables were controlled in Step 1, and variables of interest were entered in Step 2 and/or Step 3. H1 and RQ1 were explored with 10 models, simultaneously examining all three access variables in Step 2. H2 was tested with a total of 20 logistic regression models (Step 2 = Main Effects and Step 3 = Interaction Effects), followed by split sample analysis. Finally, H3 was examined with another series of 10 logistic regression models (Step 2 = Access Variables and Step 3 = Device Quality).

Results

Sociodemographic predictors of access

Before testing the hypotheses and exploring the research questions, we first examined how sociodemographic variables influenced the three access variables in the CPS 2021 and 2023 data. The details are described in the Supplementary materials (Supplementary Appendix B) and are generally consistent with previous research on demographic predictors of digital access, particularly the strong positive relationship between income and education with all three access variables. In most cases, racial and ethnic minorities and those with a disability were also less likely to have in-home internet access or use computers, though these differences were largely absent for smartphones. Finally, age was consistently negatively associated with access, though these effects were notably small.

Resources and internet benefits

The first hypothesis (H1) predicts the positive associations between in-home internet access, computer access, and smartphone access and each of the five outcome variables. In a validation and elaboration of resources and appropriation theory (Van Dijk, 2005, 2020), our research question (RQ1) explored the relative effects of each of the traditional access variables to determine which was most closely associated with (a) online job searches, (b) use of the internet for government resources, (c) online health information seeking, (d) patient-provider communication, and (e) use of medical records. We tested ten hierarchical logistic regression models across two waves of CPS data by controlling for the effects of sociodemographic variables in the first step and comparing the effects of all traditional access variables in the second step as they jointly predict the likelihood of engaging in each of the five outcome behaviors (Table 1).

H1a was partially supported with computer access being the only form of access associated with of online job searches. H1b, H1c, H1d, and H1e were fully supported, with all three types of access being significant predictors of use of online government resources, health information seeking, patient-provider communication, and medical records.

In response to RQ1, we found that in both CPS 2021 and 2023 data, household access to large-screen computers was the strongest predictor of beneficial internet use based on the size of model coefficients, except in the case of 2021 data on patient-provider communication when smartphone ownership yielded the strongest relationship, followed by in-home internet, then computer use (see Table 1).

Overall, these data were somewhat unexpected, standing in contrast to an emphasis on internet service as the focus of many digital equity policies (King & Gonzales, 2023). Thus, we further probed these effects by directly comparing cases in which, regardless of smartphone access, respondents had home internet but did not use computers (as the reference group) and those who used computers but had no home internet, using ten additional hierarchical logistic regression models (see Tables 2 and 3 for sample sizes and descriptives of these subsamples). In this way, we can directly compare the benefits of in-home internet and computer access.

Findings revealed that households that used computers but no home internet were more likely to use the internet in beneficial ways than those with home internet that did not use computers (see Table 4). The only exception was a marginally significant association with patient-provider communication in 2021 in the same direction. These findings indicate that, while both resources are very important, computer use may be of even greater importance for facilitating beneficial uses of the internet than having in-home internet access.

Interaction between in-home internet and device access

Whereas our first set of analyses explored relative differences in the benefits of internet resources, our second hypothesis (H2) explores the multiplicative effects of having a range of first level-digital resources, predicting that the benefits of both household computer access and smartphone access would each be stronger with in-home internet access. To test this hypothesis, we conducted a series of hierarchical logistic regression in which we included an interaction term for in-home internet access and (H2a) computer access for each outcome, followed by models with an interaction term for household home internet access and (H2b) smartphone access for each outcome (Table 5).

Tests of H2a revealed that having household home internet access moderated the relationship between computer access and two health-related outcomes in the 2021 data in the opposite direction and moderated the effect of computer access on medical records use in the 2023 data in the predicted direction. Split-sample analyses for both waves of data (Figure 1) reveal that for all analyses, computer access was positively associated with beneficial internet activities regardless of whether households had in-home internet. In sum, H2a was not supported; this unexpected finding speaks to the importance of supplementary forms of internet service and is elaborated further below.

Once again, it was somewhat surprising to us that computers were associated with beneficial internet use for individuals not using the internet at home. We imagine that, in these cases, benefits are due in part to laptop users transporting devices to public spaces with internet service. To test this, we distinguished laptop from desktop internet use (see Supplementary Appendix C) and, indeed, found that for individuals without in-home or mobile internet access, only laptops were associated with internet benefits in the 2021 data. In the 2023 data, laptops were more consistently associated with internet benefits, though desktop use was also associated with using the internet for government services and online medical records. These findings are discussed further below.²

Testing the second half of this hypothesis (H2b, Table 5) revealed that having household home internet access moderated the relationship between smartphone access and beneficial internet use for all but three outcomes. To further understand the results of H2b, we again decomposed these effects by conducting split sample analyses. As can be seen from Figure 2, in both waves of data smartphone access was positively associated with beneficial use when in-home internet access was available, except in the case of online job searches in 2023. However, in most cases there was either no effect or a negative effect of smartphone access on beneficial use when in-home internet access was absent; two exceptions to this were health information seeking and online medical record use in 2023. Overall, H2b was generally supported.

Device quality also matters

Finally, the last hypothesis (H3a–e) predicted that quality with one’s devices would predict additional variance in the likelihood that respondents engaged in beneficial online activities (see Supplementary Appendix D). For both waves of data, the analyses similarly produced mixed support for the hypothesis. Although device quality was not associated with the likelihood of searching for work online, it was positively associated with the likelihood of engaging in online government resources in 2023 and with health outcomes in both waves. Implications for the technology maintenance construct are discussed below.

Discussion

The resources and appropriation theory of the digital divide argues that variation in digital access shapes one’s social and economic status, which in turn shapes access to digital resources (Van Dijk, 2005, 2020). Fundamental to those digital resources is “physical access” to ICTs, often operationalized as in-home internet service, a computer, and a smartphone. Though scholars and stakeholders alike have recognized that physical access must extend far beyond those three fundamentals (Goedhart et al., 2019; Gonzales, 2014; Gonzales et al., 2016; Katz, 2017; Katz & Gonzales, 2016; Reisdorf et al., 2022; Robinson et al., 2015; van Deursen & van Dijk, 2019), they have not considered the relative differences, or interactions between internet and device access in association with beneficial internet use. We also examined device quality as an additional factor contributing to beneficial internet use and discussed those findings through the lens of the technology maintenance framework (Gonzales, 2014; Gonzales et al., 2016).

The importance of computer access

Our analyses of the U.S. CPS Computer and Internet supplement from 2021 and 2023 testing H1a–e reveal that, with minor exception, individuals with in-home internet service, computers, and smartphones were more likely to engage in a variety of beneficial internet activities than those who do not have those resources, including job searches, accessing government services (e.g., driver’s license renewal, voter registration), and various forms of healthcare-related activities, including searching for health information online, communicating with providers online, and accessing online medical records. The exception to this was in 2023 when having internet or a smartphone was not associated with job searching. However, by and large these findings are aligned with decades of research on the critical role of physical and material ICT access (e.g., DiMaggio et al., 2001; Van Dijk, 2005, 2020; Warschauer, 2004) and underscore that all three of these first-level divide resources are important.

More surprisingly, our analyses of RQ1 also revealed that computer access had the strongest relationship with beneficial internet use across nearly all outcomes. To probe this further, we found that households using computers without in-home internet were more likely to be using the internet in beneficial ways than households with in-home internet and no computer use. That is, a direct comparison of computer use and in-home internet access revealed that computer use was a stronger predictor of beneficial internet use. This finding was further elaborated with the results of H2a: having a computer was associated with beneficial uses of the internet even in homes without in-home internet.³ Findings that these effects were more consistently associated with laptops than desktops add further weight to the argument that public Wi-Fi is an essential supplement for households without internet service, though we cannot rule out the possibility that these households may also be using wireless hotspots or shared Wi-Fi with neighbors to get online at home. Finally, the analysis of H3 revealed that the quality of one’s internet devices also increased the likelihood of using the internet for government resources and all three healthcare-related activities, which corroborates the argument of the technology maintenance framework. In sum, in addition to in-home internet service, these data spotlight the need for reliable access to high-quality large-screen computing devices, especially laptops. Moreover, having a device that worked well was also associated with beneficial use. In short: computers matter. That is, even though policymakers (see King & Gonzales, 2023 for discussion) and the press (e.g., Dvorak, 2020) often prioritize internet access for closing the digital divide, these findings remind us that reliable computer access is also critical.

Given these findings, more research is needed on the barriers to, and benefits of, computer use, both in-home and in public contexts. We know that device distribution has improved access to healthcare in some settings (Ferguson et al., 2024; Wray et al., 2022), and there are mixed results about the benefits of devices in educational settings (Fairlie & Robinson, 2013; Martin et al., 2024), but more work is needed to understand how low-income individuals navigate the challenges of steady computer access. Moreover, continued work is needed to understand effective interventions to give people the skills and confidence to adopt computer use. We know from previous research that public Wi-Fi and hotspots often serve as a stopgap for getting online when in-home internet is lacking and may even help to facilitate enrollment in low-cost in-home service (Horrigan et al., 2024; Rhinesmith, 2024; Strover, 2019). However, access to these spaces is sometimes limited by extra policing, hours of service, number of computers, etc. (Grimes & Porter, 2024; Strover et al., 2020; Yang et al., 2021). Given the often-overlooked complexity of computer use, additional data are needed on the relationship between an individual’s personal infrastructure of digital skills and resources, and outcomes of use.

The theoretical contribution of this work is twofold. First, scholars of digital equity have not considered the relative-benefits of internet service versus devices, instead assuming that these fundamental first-level divide resources go hand-in-hand. These data continue the tradition of previous scholarship that has argued for a nuanced conceptualization of first-level divide problems (e.g., Katz, 2017; van Deursen & van Dijk, 2019) validating resources and appropriation theory from a new perspective. Second, findings suggest that policymakers and stakeholders need to ensure that devices are not only available but also sufficiently functional. This is consistent with a technology maintenance approach to digital equity, with its emphasis on optimizing the sustainability of ICT access (Gonzales, 2014; Gonzales et al., 2016). By adding this new, holistic measure of device quality to census data, these findings further underscore the need for more nuanced conceptualizations of digital equity.

The risks of smartphone dependence

Analysis of H2 revealed that the benefits of personal computers, especially high-functioning ones, stand in contrast to the more complicated role that smartphones play in contributing to online benefits. Scholarship on mobile phones has lauded these devices as critical tools for getting telephony and information into the hands of people who have often lack easy access to such resources. Transient citizens or those living in regions without wireline access, for example, have benefited enormously from the informational and social affordances of mobile computing (Calvo et al., 2019; Pearce & Rice, 2013; Read et al., 2021). On the other hand, relying on smartphones as the primary form of internet access in the home limits the range of internet activities (Correa et al., 2018; Pearce & Rice, 2013; Reisdorf et al., 2022), which also limits one's ability to improve their digital skills (Correa et al., 2022). These data build upon the previous work, indicating that in homes with smartphones, having in-home internet access greatly amplifies the likelihood that people are using the internet in beneficial ways. One plausible explanation is that, after holding key demographic factors constant (e.g., education, age, etc.), smartphone web use is not only more cumbersome than a computer (e.g., accessing medical records, which require log-ins and navigating unfamiliar interfaces) but also more distracting by offering a ready source of other activities (e.g., texting, email, and social media; see Reisdorf et al., 2022), though more work is needed to explore this explanation effect.

Policy implications, limitations, and future directions

As digital equity activists search for new solutions to the now-obsolete Affordable Connectivity Program (Federal Communications Commission, 2023), these findings urge a clear-eyed approach to the importance of devices. For instance, the Affordable Connectivity Program from the FCC previously included a $100 subsidy for computer purchases, but, to our knowledge, little research has examined the success of that component of the program. As a source for hot-spot loans, public internet, and devices, libraries are instrumental resources to address this multitude of necessary resources (Rhinesmith, 2024; Strover, 2019; Strover et al., 2020), though funding limits often mean demand exceeds supply, which will likely worsen if recent efforts to close the Institute for Museum and Library Services, the federal agency responsible for supporting libraries, are successful. Finally, schools have also played a major role nationwide in increasing computer access for students, but these low-cost devices are often quite limited in their functionality and are only available to students. Yet this sticky problem of facilitating device access is likely one worth pursuing, according to these data.

The primary limitation of these data rests in the fact that they are cross-sectional, and thus causal relationships cannot be determined. Although the relationship between independent and dependent variables is not likely bi-directional, it is quite possible that, despite a range of sociodemographic control variables, alternative extraneous variables may partially explain these findings. As one possible indication of this, individuals that use computers with no in-home internet are younger, wealthier, and better educated than those with in-home internet that are not using a computer (see Table 3 notes). These data likely may be pointing to spurious variables that reduce the chances of using a computer and using the internet in beneficial ways that should be further investigated.

Related to this, these data are limited in the range of factors we can measure. We could not examine frequency of use, digital skills, etc., as they are not currently used in CPS data. We also acknowledge that measures of computer and smartphone access do not specify personal ownership. That is, some respondents who have answered in the affirmative likely do not own computers but rather are consistently accessing them through alternative means (e.g., libraries, friends, etc.). We suggest that CPS consider revising this question, or adding a supplementary question that would distinguish shared, community, and personal device ownership, as that would better inform policy making.

Conclusion

Despite the limitations of cross-sectional data, these findings underscore the role that computers play in ensuring successful internet access. Although this may seem obvious, U.S. policy has historically emphasized provisions of internet service (King & Gonzales, 2023), a conversation that was magnified for important reasons during COVID-19. We do not mean to diminish the importance of in-home internet service, particularly in homes with smartphone users. Internet access is critical for daily life in industrialized contexts, and the enormously expensive nature of broadband infrastructure often requires government intervention. But, as a result, policies may at times overlook the essential nature of having physical and operational access to a reliable computer. Researchers should continue to explore the barriers that people face in successfully accessing a computer. With a substantial percent of the U.S. population still without in-home internet or computer access, there is still work to be done.

Supplementary material

Supplementary material is available at Journal of Computer-Mediated Communication online.

Data availability

The data underlying this article are available at the U.S. National Telecommunications and Information Administration: https://www.ntia.gov/page/download-ntia-internet-use-survey-datasets.

Funding

This work was funded in part by an Academic Senate Grant from the University of California, Santa Barbara.

Conflicts of interest: None declared.

Notes

References