Technology Access and Use, and Their Associations With Social Engagement Among Older Adults: Do Women and Men Differ?

Abstract

Objectives

To examine how information and communication technology (ICT) access and use are conceptually incorporated in the Successful Aging 2.0 framework.

Method

Using data from the 2011 National Health and Aging Trends Study (N = 6,476), we examined how ICT access and use for different purposes are associated with social engagement (i.e., informal and formal social participation) by gender. Weighted logistic regression analyses were performed.

Results

Findings revealed that men were more likely to access and use ICT than women. ICT access was positively associated with all types of women’s social engagement, but only with men’s informal social participation. Information technology (IT) use for health matters was positively associated with formal social participation for women and with informal social participation for men. IT use for personal tasks was negatively associated with formal social participation for older adults. Communication technology use was positively associated with formal and informal social participation for women and men.

Discussion

This study supports the expansion of the successful aging model by incorporating ICT access and use. Further, it assists in the identification of specific technologies that promote active engagement in later life for women and men.

Information and communication technology (ICT) is defined as using e-mail, the Internet, social networking sites, and voice/video technology on devices such as smartphones, computers, and tablets (Blaschke, Freddolino, & Mullen, 2009). A Pew Research Center study from 2013 found that 77% of U.S. older adults owned a cell phone, 59% used the Internet, and 47% had a high-speed broadband connection at home (Smith, 2014). Older adults use the Internet less than all other age groups, but their rate of use has been increasing rapidly in recent years (Perrin & Duggan, 2015). For example, about 14% of older adults used the Internet in 2000, as compared to about 58% of older adults in 2015. Research has found that older adults who are less advanced in age (Choi & Dinitto, 2013; Cody, Dunn, Hoppin, & Wendt, 1999; Morris, Goodman, & Brading, 2007; Smith, 2014), identify as White (Choi & Dinitto, 2013), have more education (Choi & Dinitto, 2013; Vroman, Arthanat, & Lysack, 2015), are not depressed (Cotten, Ford, Ford, & Hale, 2014), have higher incomes (Choi & Dinitto, 2013; Smith, 2014), and who maintain social connections (Choi & Dinitto, 2013; Cody et al., 1999; Hogeboom, McDermott, Perrin, Osman, & Bell-Ellison, 2010) are more likely to use the Internet.

Gender is also another significant factor that has affected older adults’ accessing and using ICT. In a recent survey, older women still lagged behind older men in their adoption of digital devices and Internet usage (Smith, 2014), whereas this gender gap has disappeared in the general population (Perrin & Duggan, 2015). This gap in older adulthood can be partially explained using a life-course perspective. Men and women have experienced different expectations and opportunities over the years, especially in regard to educational attainment, paid employment, and caregiving responsibilities (Settersten & Lovegreen, 1998), thus influencing exposure to ICT. Research has found that older women reported feeling tension between managing household chores and having time to use the computer (Barnett, Buys, & Adkins, 2000; Buse, 2009).

ICT can transform how older people engage in social activities. Research has pointed to the benefits and usefulness of ICT use in informal social engagement among older adults. For example, using e-mail helps older adults establish positive and time efficient interactions with family and friends (White & Weatherall, 2000). Further, ICT use helps older adults supplement and enhance their in-person connections with family or friends (Cornejo, Tentori, & Favela, 2013; Russell, Campbell, & Hughes, 2008). Older adults’ ICT use is also positively associated with formal social participation, such as attending meetings and involvement with community organizations (Cody et al., 1999; Hogeboom et al., 2010), religious participation, and volunteer work (Choi & Dinitto, 2013). Recently, Ihm and Hsieh (2015) reported that older adults’ offline social engagement (informal and formal) was positively associated with instrumental ICT use (e.g., obtaining information, services, and other resources), but not with social ICT use (e.g., using social media, posting comments on blogs).

Research has reported gender differences in social engagement, for both informal and formal participation. Seeman, Lusignolo, Albert, and Berkman (2001) found that older men reported more close social ties with family, friends, or relatives, but women reported more involvement in social group activities. Studies also found that older women were more likely than men to be involved in religious participation (Zhang, 2010) and do volunteer work (Moen & Flood, 2013). However, more research is needed to understand how older adults’ accessing and using ICT is differentially associated with their formal and informal social participation and whether or not gender inequality exists in this relationship.

Conceptual Framework

Successful aging (SA) is a concept that describes the aging process positively. The original SA model by Rowe and Kahn (1997) defined three core components: “low probability of disease and disease-related disability, high cognitive and physical functional capacity, and active engagement with life” (p. 433). In their model, successful agers embodied all three components. However, the narrowly defined criteria of the SA model have been criticized because they capture a small number of older adults in advancing age (Cho, Martin, & Poon, 2012), devalue a role of genetics (Masoro, 2001), and reflect a predominantly medical perspective (Katz & Calasanti, 2015).

Recently, the SA model has been re-examined in terms of conceptual definitions, practical measures, and factors that have not been fully incorporated (Pruncho, 2015). Rowe and Kahn (2015) suggested conceptually expanding their original model by incorporating the impacts of macrosocial changes and unexplored social factors that can influence individuals’ aging successfully. Technological innovations may be one of the most powerful and important macrosocial changes that have influenced older adults’ social engagement and lifestyles. Older adults, however, can experience an individual, structural, or cultural “lag” (Cutler, 2006) in response to technological changes. For example, an individual lag (Lawton, 1998) can occur when technological environments change more rapidly than older individuals’ capabilities to adopt and use technologies. Also, older adults, specifically women, could experience a structural lag (Riley & Riley, 1994) if they lack educational and employment opportunities to learn and use digital technologies within their life course.

In the literature, social engagement and social participation have been used interchangeably (Bennett, 2002; Morrow-Howell & Gehlert, 2012). Social engagement refers to maintaining social connections and participating in social activities (Bassuk, Glass, & Berkman, 1999). In this study, social engagement is measured by two components: informal and formal social participation. Informal social participation refers to social interactions with friends, neighbors, or relatives through in-person visiting; formal social participation refers to religious participation, meetings, and involvement with community organizations, and volunteering (Utz, Carr, Nesse, & Worthman, 2002). As noted in Ihm and Hsieh (2015), instrumental ICT use was positively associated with older adults’ offline social engagement. These findings provide strong support for examining the role of ICT use for different purposes in the context of SA.

The current study focuses on older adults’ accessing and using ICT, and its link to social engagement by gender, which expands the original SA model. Social engagement incorporates social relations and productive activity participation (Rowe & Kahn, 1997) and has been studied as a modifiable factor that has influenced health outcomes. For example, social engagement enhanced subjective physical health (Bennett, 2005), was negatively associated with functional disability (Mendes de Leon, Glass, & Berkman, 2003), and was associated with better cognitive functioning (Krueger et al., 2009). Instead of narrowly focusing on older adults who have successfully met all three criteria of the SA model, this study controls for health factors (e.g., chronic illnesses, limitations in activities or mobility, cognitive function) in explaining social engagement. By incorporating various social factors, this study mainly focuses on how digital inequality in accessing and utilizing ICT (DiMaggio & Hargittai, 2001) is connected with social engagement and how gender is differentially expressed in this relationship.

Current Study

The present study aims to investigate (a) ICT access and use among older adults by gender and (b) the relationship between ICT access and use, and informal and formal social participation by gender.

Method

Data and Study Sample

Data were drawn from the first round of the National Health and Aging Trends Study (NHATS). NHATS is a nationally representative panel study that examines health, disability, cognitive capacity, and other aspects of lifestyles of adults ages 65 and older residing in the contiguous United States. These data were sampled from the Medicare enrollment database as of September 30, 2010, using a stratified three-stage sample design and oversampling persons aged 90 or older and non-Hispanic African American individuals. With a 71% response rate, the first round of NHATS in 2011 was composed of 8,245 Medicare beneficiaries who are reinterviewed annually (Montaquila, Freedman, Edwards, & Kasper, 2012). This study used a community-based sample (N = 7,197), which excluded study participants living in nursing homes (N = 468) or other residential care facilities (N = 412), and questionnaires that were completed by staff members in residential care settings (N = 168). Additionally, the analysis excluded cases with missing data (N = 204) and cases completed by a proxy respondent (N = 517) because measures were defined as self-reported. Therefore, the current study sample was composed of 6,476 older adults residing in the community.

Measures

Dependent variables

Social engagement

Informal social participation includes (a) visiting in-person with friends or family not living with the sample person either at the sample person’s home or theirs and (b) going out for enjoyment (e.g., for dinner, a movie, to hear music). Formal social participation refers to (a) attending classes, clubs, or organized activities; (b) volunteering; and (c) religious participation (1 = yes for participation in each).

Independent variables

ICT access and use

This study defined communication technology (CT) use as sending messages by e-mailing or texting and information technology (IT) use as using the Internet for any other reasons, including (a) completing three personal tasks as routine activities (i.e., grocery shopping, banking, or ordering prescriptions) and (b) dealing with three health matters (i.e., contacting medical providers, handling health insurance issues, or obtaining information about health conditions). IT use measures aggregated the number of each usage (range = 0–3). CT use was assessed by the frequency of e-mailing or texting, ranging from none (0) to most days (3). ICT access measured whether the respondent was able to access ICT at any place and knew how to use computer (1 = yes).

Control variables

Physical health conditions

Chronic health condition was measured by summing nine possible conditions diagnosed by a doctor (range = 0–5, where 5 represents five to nine conditions). Activities of daily living (ADLs) include eating, bathing, toileting, dressing, and getting out of bed. Mobility refers to getting around inside the home and going outside. Difficulties in ADLs were categorized as none, 1 or 2 ADLs, and 3–5 ADLs. Difficulty in mobility assessed whether the respondent had difficulty in any of two mobility items (1 = yes).

Mental health conditions

Cognitive function was a subjective measure of one’s memory status at the time of the interview, ranging from poor (1) to excellent (5). Depressive symptoms were assessed by two questions from the Patient Health Questionnaire 2 (PHQ-2): frequency of depressed mood and anhedonia, ranging from not at all (0) to nearly every day (3) (Kroenke, Spitzer, & Williams, 2003). The PHQ-2 has been validated as a screening tool for detecting major depression in older adults (Li, Friedman, Conwell, & Fiscella, 2007). Unlike the original questionnaire, the reference time for each question was extended from 2 weeks to 1 month in the NHATS, thus possibly contributing to the lower reliability coefficient (α = .55) in the current study. Level of depressive symptoms was categorized as none (0), mild depression (1 or 2), and major depression (3–6) from the total PHQ-2 combined score, by adopting the optimal cutoff points (Kroenke et al., 2003).

Size of social network

Size of social network was measured by counting number of persons who the respondents talked with most often about important things in the past year (range = 0–5).

Sociodemographic variables

This study included age, gender, marital status, being employed for payment, race and ethnicity, level of education, and annual income. In the NHATS surveys, a total annual income was reported as either a collected amount of possible income sources and assets or one of five bracketed ranges. For missing income cases (31%) and bracketed responses (13%), five imputed values for total annual income variable were provided (see Montaquila, Freedman, & Kasper, 2012 for details). The total annual income ranged from $0 to $5,700,000, with median income of $28,977 in the study sample ($36,000 for men and $22,000 for women). Due to high skewness, this study performed log transformations of the five imputed income variables and used them in the multivariate analyses using multiple imputation estimate methods. The wording of survey questions and construction for the variables are presented in Supplementary Table A.

Analytical Strategy

For all data analyses, the current study used weighted statistics with analytic sample weights, Taylor series linearization method, and subpopulation analyses to adjust for standard errors with complex survey design, using Stata 13.0 (StataCorp, 2013). The t statistic or the adjusted Wald F statistic estimated gender differences in sample characteristics (Table 1) and ICT access and use (Table 2). Weighted logistic regression analyses were used for estimating men’s and women’s informal social participation (Table 3), as well as their formal social participation (Table 4). Using mi estimate command, this study estimated model F tests (equal fractions of missing information) from five data sets (i.e., each data set includes a different imputed income variable, but the remaining variables are same) and adjusted coefficients and standard errors for the variability between imputations in all logistic regression analyses. The variance inflation factor tests did not indicate the presence of multicollinearity in multivariate models according to the recommended cutoff point of 2.5 (Allison, 1999).

Results

Sample Characteristics

Gender differences in dependent and control variables are presented in Table 1. Women comprised about 56% of the sample. The vast majority of the sample is engaged in informal social participation. Women were more likely to engage in socializing with family or friends than men (90.1% vs. 86.1%). Eighty-one percent of the sample went out for enjoyment; no gender differences were found. For formal social participation, respondents were most likely to attend religious services (57.7%), followed by attending classes, clubs or organized activities (38.4%), and participating in volunteer work (26.5%). Women were more likely to engage in all the aforementioned formal activity participation except volunteer work.

Gender differences were more evident in the following control variables. Women were more likely to report difficulties in 1 or 2 ADLs (21.3% vs. 16.9%) and mobility (23.1% vs. 15.1%) compared to men. Women were more likely to have mild depression (31.8% vs. 25.5%) and major depression (13.5% vs. 11.9%) than men. Women were more than twice likely to be unmarried or not partnered than men (53.2% vs. 23.6%). Men were more likely to work for payment than women (25.2% vs. 15.5%).

Gender Differences in Older Adults’ ICT Access and Use

Table 2 presents the details of ICT access and use by gender. Men were more likely to access ICT (89.8% vs. 85.7%, p < .001) than women. In this study, ICT access among older women and men was measured by devices ownership and computer literacy—80% owned a cell phone, 68.6% owned a computer, and 66.8% were computer literate. It is noted that about 5.8% of older adult who own a computer did not know how to use it, whereas 6.9% of older adults who did not own a computer have used one elsewhere. Data on devices ownership and computer literacy are not shown in Table 2.

About 56% of older adults had not sent messages by texting or e-mailing in the past month, whereas 24% used CT on most days and 13% on some days. Men used CT more frequently than women (M = 1.1 vs. M = 1.0, p = .012). Older adults were most likely to use the Internet for paying bills or online banking (20.8%), followed by obtaining information about health conditions (16.9%), and shopping for groceries or personal items (15.0%). Men were more likely to use the Internet for dealing with personal tasks (M = 0.5 vs. M = 0.4, p < .001) and health matters (M = 0.4 vs. M = 0.3, p < .001) compared to women.

Gender Differences in Older Adults’ Informal Social Participation

Table 3 presents odds ratios (OR) to predict older adults’ informal social participation. ICT access was associated with an increased likelihood of women’s visiting with family or friends (OR = 1.6, p = .002) and going out for enjoyment (OR = 1.3, p = .018); however, it was only associated with men’s going out for enjoyment (OR = 1.4, p = .036). CT use was positively associated with men’s and women’s going out for enjoyment (OR = 1.4, OR = 1.3, both p < .001), but only with men’s visiting with family or friends (OR = 1.2, p = .04). IT use for health matters was also positively associated with men’s visiting with family or friends (OR = 1.5, p = .008).

Gender Differences in Older Adults’ Formal Social Participation

Table 4 presents the OR to predict older adults’ formal social participation. ICT access was associated with an increased likelihood of all three types of formal social participation for only women, respectively (OR = 1.9, p < .001; OR = 1.5, p = .016; OR = 1.3, p = .041). CT use was associated with an increased likelihood of both men’s and women’s attending clubs, classes, or organized activities (OR = 1.3, OR = 1.3, both p < .001) and volunteering (OR = 1.3, OR = 1.3, both p < .001). IT use for health matters was positively associated with only women’s attending clubs, classes, or organized activities (OR = 1.2, p = .044) and volunteering (OR = 1.2, p = .048). IT use for personal tasks was associated with a decreased likelihood of men’s and women’s religious participation (OR = 0.8, p = .002; OR = 0.9, p = .038).

Discussion

This study was implemented to broaden the conceptual framework of the successful aging model 2.0 (Rowe & Kahn, 2015) by incorporating ICT access and use. Specifically, it examined gender differences in ICT access and use among older adults and in the relationship between social engagement and ICT access and use.

Findings revealed a gap between access and use of ICT among community-dwelling older adults. Whereas the majority (88%) of older adults were able to access ICT or knew how to use a computer, their rates of utilization were quite low, ranging from 44% (e-mailing/texting) to 7.6% (contacting medical providers). This discrepancy between older adults’ access to ICT and their utilization of ICT underscores the need to help older adults overcome digital inequality.

Our study found that men were more likely than women to access and use digital technology for all different purposes, including communication, completing personal tasks, and handling health matters. These findings corroborate the outcomes of other national surveys (e.g., Smith, 2014) and may point to the influence of gender-based life-course trajectories in education, employment opportunities, or caregiving responsibilities (Moen & Flood, 2013; Settersten & Lovegreen, 1998). Some studies on older technology users’ activities have revealed that women were more likely than men to use digital technology for health- and hobby-related information (e.g., Karavidas, Lim, & Katsikas, 2005). However, other studies have found no gender differences among older Internet users for e-mail communication with family or looking for health and wellness information (Vroman et al., 2015), or using the Internet for personal tasks, health-related tasks, or e-mailing/texting (Choi & Dinitto, 2013). These mixed findings in the literature imply that gender inequality in ICT use may be diminishing among older adults who were able to access and use ICT.

Gender differences were also found in the relationships between social engagement and ICT access and use. For example, ICT access was positively associated with all types of social engagement for women, but only with men’s going out for enjoyment. These findings highlight how ICT access and use benefit social engagement (Ihm & Hsieh, 2015), and expand our understanding of gender inequality. These findings may be partly due to differences in marital status or difficulties in ADLs or mobility by gender. Being married or partnered had the highest odds of men’s engaging in informal social activities. Difficulties in ADLs or mobility were negatively associated with all types of social engagement for women, but only with volunteering for men. It is possible that men could maintain social relationships with their spouse or a partner regardless of accessing to ICT. Women who were not able to access to ICT, however, might be more likely to be socially isolated due to experiencing physical challenges. Special attention is needed to develop ways to reduce digital inequality and social isolation among women.

Men were more likely to use ICT for a variety of purposes (Smith, 2014), but less likely to engage in most types of social engagement (Ihm & Hsieh 2015; Moen & Flood, 2013; Zhang, 2010), except for going out for personal enjoyment. Gender differences, however, existed in the relationships between ICT use patterns and social engagement. CT use was positively associated with older adults’ informal and formal social engagement. This finding highlights the benefits of online communication because it may have the potential to foster in-person interactions with family or friends (Russell et al., 2008; White & Weatherall, 2000), and expands offline social networks organized around leisure interests or volunteer work (Russell et al., 2008).

IT use for dealing with health matters, however, was associated with different types of social engagement by gender. Specifically, when older adults used the Internet for health matters, women were more likely to attend clubs or organized activities and do volunteer work (formal participation), whereas men were more likely to socialize with family or friends in-person (informal participation). These findings expand our understanding about gender differences in social engagement; men engaged in more close ties with family, friends, or relatives, but women were more likely to participate in social group activities (Seeman et al., 2001). Technology use for health information gathering might help women and men to maintain their preferred types of social engagement. It, however, is unclear how such gender-based differences existed in the relationships between IT use for health matters and social engagement, as previous research on health information technology use has primarily focused on health service utilization (e.g., Choi, 2011). More research is warranted to unpack the complex relationships between health information technology use, social engagement, and health outcomes for the conceptual expansion of the SA model.

Unexpectedly, IT use for personal tasks was negatively associated with men’s and women’s religious service attendance, which contradicts the finding by Ihm and Hsieh (2015) that instrumental ICT use is associated with various types of offline social engagement. Although Ihm and Hsieh’s measure of instrumental ICT is similar to the one used in this study, combining religious service attendance and other types of activities into one global measure of social engagement and differences in sample characteristics may partially explain the divergent outcomes. Our sample was predominantly non-Hispanic White (82%), whereas Ihm and Hsieh did not report the racial or ethnic composition of their sample. Therefore, it is unclear how their findings could be attributed to race or ethnicity, which are important factors when explaining digital inequalities and religious participation. More research is needed to explicate the relationships between ICT use patterns and religious participation using diverse racial groups of older adults and detailed measures of ICT activities.

The gender-based outcomes obtained by this study expand our understanding of the relationship between older adults’ ICT access and use, and informal (Cornejo et al., 2013; Ihm & Hsieh, 2015; Russell et al., 2008; White & Weatherall, 2000) and formal social participation (Choi & Dinitto, 2013; Cody et al., 1999; Hogeboom et al., 2010; Ihm & Hsieh, 2015). They also offer insights into how we can possibly promote SA through involvement in social activities. For instance, the development and implementation of age- and gender-specific interventions that reduce inequalities in technological skill and use (van Dijk, 2012) and promote comfort and confidence with ICT (Wagner, Hassanein, & Head, 2010) among older adults may enhance their social engagement.

NHATS data have several key limitations. For the measure of informal and formal social participation, we could only determine if the respondents participated in such activities, but not their level of participation (e.g., number of times per month). Also, for visiting with family or friends either at their own homes or others, there might have been different purposes for these visits (i.e., family visits for caregiving vs. visiting friends for enjoyment). Another important variable for measuring informal social participation, socializing with family or friends by phone (Utz et al., 2002), was not included in this survey.

With limited questions on ICT use in the survey, it was only possible to capture participation in each type of IT use at a given time frame, not the frequencies of each activity, as was done for CT use. It is also noted that the measure of CT may be conceptualized differently. We defined e-mailing or texting as CT use because the questions do not clarify their purposes. However, they could also be a form of informal or formal social engagement. Future research is needed to further explain how e-mailing or texting could be perceived as another form of social engagement. Additionally, tablet and social network site use were not included in the first and second round of NHATS. The current study measured ordering or refilling prescription drug as one of the personal tasks because it was considered a simple one to execute online. IT use for health matters was considered to require more complex cognitive functioning. Choi and Dinitto (2013), however, measured ordering or refilling prescriptions as one of the health-related tasks according to item characteristics. More research is warranted to examine a latent construct including measurement errors.

We have limited information on racial minority groups and immigrants who are not enrolled in Medicare, due to restrictions imposed by the Personal Responsibility and Work Opportunity Reconciliation Act (Reyes & Hardy, 2014). The study sample also did not include baby boomers who turned 65 years in 2011. Future research is warranted to study ICT use and social engagement using surveys with more racial and ethnic minorities or diverse cohorts of older adults, including baby boomers (e.g., the forthcoming 2015 NHATS data).

Finally, a causal relationship between ICT use and social engagement cannot be inferred due to the use of cross-sectional data. Because we focused on incorporating ICT access and use in the SA model, the current study examined associations between ICT access and use, and informal and formal social engagement, using the baseline 2011 data for the following 2012−2014 NHATS. The findings from our study can provide researchers with preliminary groundwork to investigate the causal and/or bidirectional relationships between ICT use and social engagement, using a longitudinal research design.

This study makes an important contribution to expanding the SA model by incorporating the role of digital technology. It also broadens our understanding of how older women’s and men’s ICT access and use are associated with their informal and formal social participation. By promoting active engagement in life through ICT use, older adults will have greater opportunities to reduce social isolation and improve their overall psychosocial well-being, although there is a possibility they could feel disempowered by encountering barriers and concerns associated with safety issues and a lack of digital skills (Hill, Betts, & Gardner, 2015). The development of tailored strategies to engage older adults in social activities using digital technology could begin by taking into account a considerable gap between those who were able to access ICT and those who actually used ICT. Researchers, e-commercials, health care institutions, and technology program developers should reconsider how they provide adequate opportunities to meet the needs of older adults, thus reducing such a gap.

More specifically, a special consideration is warranted for older women who experienced the disadvantages of technological advances to a greater degree than men. Older women who were not able to access digital devices or learn about computer skills over the years have a higher chance of being socially isolated. A community-based educational effort to promote technology access and literacy for older adults is urgently needed to bolster their informal and formal social engagement. Given that about one-third of study sample lacked computer ownership or literacy, this group also needs to be prioritized for having more computer access and educational opportunities in the community. The general recommendation for ICT use is an imperative part of older adult education.

Supplementary Material

Supplementary data is available at The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences online.

Conflict of Interest

The authors declare that they have no conflict of interest.

Acknowledgments

J. Kim and H. Y. Lee planned the study; J. Kim, H. Y. Lee, M. C. Christensen, and J. R. Merighi wrote and revised the manuscript; and J. Kim conducted the statistical analysis.

References

Author notes