Ageing and internet adoption

Since the early 2000s, scholars have argued that ‘there is a strong association between age and the so-called digital divide’ (Selwyn, Reference Selwyn2004; Neves and Amaro, Reference Neves and Amaro2012) and have coined the term ‘grey divide’ (Millward, Reference Millward2003) to describe the lack of ‘access, skills and/or knowledge that can result in older citizens being “information poor”’ (Kania-Lundholm and Torres, Reference Kania-Lundholm and Torres2015). However, research in this area has also received criticism for treating older adults as a homogenous group and neglecting the diversity in older adults’ lives and how they use digital technologies (e.g. Rasi and Kilpeläinen, Reference Rasi and Kilpeläinen2016).

Instead, there is evidence of factors associated with age – beyond income, education and gender (e.g. mistrust, low media literacy, physical and mental limitations) – that continue to hamper their advantageous use of new media (Nimrod, Reference Nimrod2017; Quan-Haase et al., Reference Quan-Haase, Williams, Kicevski, Elueze and Wellman2018). For example, Wagner et al. (Reference Wagner, Hassanein and Head2010), in their review of the literature, found that computer use among older citizens varied significantly. Czaja et al. (Reference Czaja, Charness, Fisk, Hertzog, Nair, Rogers and Sharit2006) found that the relationship between age and adoption of technology was mediated by cognitive abilities, computer self-efficacy and computer anxiety. However, the study by Gell et al. (Reference Gell, Rosenberg, Demiris, LaCroix and Patel2013), using a representative sample of adults over 64 years of age in the United States of America (USA), showed that the prevalence of internet use differs according to socio-demographic status (younger age, male sex, white race, higher education level and being married) and type of disability and impairments that limit activity. Other studies have shown evidence of the influence of social support as a central mechanism that may facilitate the use of technology among older adults, especially among those with functional limitations (e.g. Kamin et al., Reference Kamin, Beyer and Lang2019). In general, these results suggest a tendency to focus on the obstacles and facilities that can make it possible for older people to regulate their own social context with the help of technologies (Kamin et al., Reference Kamin, Lang, Kamber and Kwon2016).

Digital inequality in older people

As the pervasiveness of the internet in society increases, the access gap between rich and poor, urban and rural dwellers, old and young, or educated and uneducated people disappears. However, the dimensions related to the quality of use become important bases by which the benefits of technology are stratified (van Dijk, Reference van Dijk2005, Reference van Dijk2006). DiMaggio and Hargittai (Reference DiMaggio and Hargittai2001) identified five dimensions on which digital inequality can be articulated among older people.

The first refers to variation in the technical means (hardware and connections) by which people access the internet (van Deursen and van Dijk, Reference van Deursen and van Dijk2018). The hypothesis is that an inferior technical device reduces the benefits that users can obtain from the internet, either because they cannot access certain sites or because their experience on the Web is less rewarding and they abandon. In a recent study by van Deursen and van Dijk (Reference van Deursen and van Dijk2018), within a representative sample in the Netherlands, they observed an inverse correlation with age in terms of the diversity of devices used. In any case, there is evidence to show that having diverse material access and frequent physical access does not guarantee beneficial results (van Deursen and Helsper, Reference van Deursen and Helsper2015).

The second is the variation in the extent to which people have autonomy when using the internet, e.g. if they access it from work or home, if their use is supervised or not, or if they must compete with other users for online time (Newhagen and Bucy, Reference Newhagen, Bucy, Bucy and Newhagen2004; van Dijk, Reference van Dijk2006). On the autonomy of internet access, there is evidence of the importance of quality access points for a beneficial use of the internet, as well as an association between the socio-educational status and the quality of internet access (Hassani, Reference Hassani2006).

The third is inequality in the skill that people bring to their use of the internet. As the access to resources among citizens is generally overcome, a third variation is defined in the skills required for a productive use of ICTs (e.g. Zillien and Hargittai, Reference Zillien and Hargittai2009; Witte and Mannon, Reference Witte and Mannon2010). Recent studies in advanced countries such as Denmark (van Boekel et al., Reference van Boekel, Peek and Luijkx2017) and Germany (Schehl et al., Reference Schehl, Leukel and Sugumaran2019) have shown that older adults are a heterogeneous group, with various categories of internet users that are identified according to frequency, type and extent of use. However, other qualitative studies carried out in the USA (e.g. Heart and Kalderon, Reference Heart and Kalderon2013) and Canada (Schreuers et al., Reference Schreuers, Quan-Haase and Martin2017) showed that the range of activities conducted by older adults through technology is limited; so that their skills will be limited in the context of these activities, reducing the range of (digital) activities that could benefit their lives. This is what Schreuers et al. (Reference Schreuers, Quan-Haase and Martin2017) named the digital skills bubble, where some activities and skills are very familiar, while others are not. Thus, and according to these authors, in order for these limitations to be overcome, the bubble will have to expand, so that older adults improve their skills. But this can only occur if they have more opportunities and social support to try to use ICTs (Francis et al., Reference Francis, Kadylak, Makki, Rikard and Cotten2018).

The fourth is inequality in the support available to internet users. There are numerous studies that have attempted to understand the training and support needed to encourage digital literacy for older adults (e.g. Sayago et al., Reference Sayago, Sloan and Blat2011; Hardill and Olphert, Reference Hardill and Olphert2012; Barnard et al., Reference Barnard, Bradley, Hodgson and Lloyd2013; Friemel, Reference Friemel2016; Rasi and Kilpeläinen, Reference Rasi and Kilpeläinen2016; Quan-Haase et al., Reference Quan-Haase, Mo and Wellman2017; Hunsaker et al., Reference Hunsaker, Nguyen, Fuchs, Djukaric, Hugentobler and Hargittai2019). In general, digital literacy is acquired through experience, which goes beyond basic exposure to technology (Murray and Pérez, Reference Murray and Pérez2014). Among older people, social support – family and friends – constitute some of the main stimuli and sources of literacy for internet use (Friemel, Reference Friemel2016). However, the support given can lack immediacy, leaving older adults dependent on others’ availability to provide it (Hunsaker et al., Reference Hunsaker, Nguyen, Fuchs, Djukaric, Hugentobler and Hargittai2019). In this sense, qualitative studies have found that older people may prefer to opt for formal sources of support (Seo et al., Reference Seo, Erba, Altschwager and Geana2019), due to the burden it may entail (Peek et al., Reference Peek, Wouters, Luijkx and Vrijhoef2016) or the feeling of dependency on one's own family (Schreurs et al., Reference Schreuers, Quan-Haase and Martin2017). Hunsaker et al. (Reference Hunsaker, Nguyen, Fuchs, Djukaric, Hugentobler and Hargittai2019), based on in-depth qualitative interviews with a multinational sample of older adults, explored the support needs of older adults in the use of digital media and concluded that educational approaches that emphasise individualised support and wide availability could allow an enhanced experience tailored to a diverse range of interests.

The fifth is variation in the purposes for which people use technology. The regulation of social contexts entails a number of challenges for older adults. The lifespan theoretical model of selection, optimisation and compensation (Baltes and Baltes, Reference Baltes, Baltes, Baltes and Baltes1990) is useful for understanding the purposes of older people when using the internet. From the outset of selection, some argue that it is essential for good ageing to use technology proactively in order to shape social contexts in accordance with age-specific needs and goals (Rohr and Lang, Reference Rohr and Lang2009). Socio-emotional selectivity serves as an example – that is, when feeling near to the end of life, older adults proactively engage in those relationships that they perceive as meaningful in their life (Carstensen et al., Reference Carstensen, Isaacowitz and Charles1999). The qualitative study by Melenhorst et al. (Reference Melenhorst, Rogers and Bouwhuis2006) concluded that although cost can be an obstacle for older adults in selecting an online activity, the absence of benefit is more likely to be the key factor. For example, Gatto and Tak (Reference Gatto and Tak2008) reported that perceived benefits of internet use among older adults were primarily associated with opportunities for social interaction, despite the obstacles they had to face. Nimrod (Reference Nimrod2013) suggested that older adults selectively participated in interactions that were in accordance with their emotional and psychosocial needs, which may be helpful in addressing age-related challenges in the areas of health, family and retirement. The optimisation principle refers to the role of technology to improve regulation of the social contexts of older people. There is empirical evidence that ICTs in old age are not only used for the compensation of loss (e.g. isolation, infrequent contact) but also for the optimisation and improvement of contact (e.g. Hogeboom et al., Reference Hogeboom, McDermott, Perrin, Osman and Bell-Ellison2010; Smith, Reference Smith2014; Rainie and Zickuhr, Reference Rainie and Zickuhr2015). Finally, the compensation principle refers to the ways in which technology is used to cope with the experiences of loss associated with ageing. For example, in a longitudinal study with Japanese older adults, researchers found that cell phone use reduces depression in older women (Minagawa and Saito, Reference Minagawa and Saito2014).

Technology support services for seniors (TSSS)

Inequality in the opportunities to access internet-based services and to benefit from their use reflects how public and institutional policies are responding to the challenge of universal internet access (Garnham, Reference Garnham1997). In order to compensate for differences in internet access among older citizens, experts and international organisations have proposed the promotion of universal internet access for the population through TSSS (Lacruz and Galofré, Reference Lacruz, Galofré and Cabezudo Rodríguez2011; WHO, 2016). Kamin et al. (Reference Kamin, Lang, Kamber and Kwon2016) define TSSS as community-based services promoted by public institutions to facilitate material access to the internet and literacy support for older people. Likewise, these services can attend to diverse social contexts and be located in different organisational contexts. For example, in Spain, universities have done this through programmes such as the University Programs for Seniors, facilitating internet use by, and digital literacy for, older adults through courses. Likewise, within the formal education system in Spain, adult education programmes also develop media and digital literacy courses. Other public policies have installed publicly owned telecentre networks that aim to facilitate and streamline internet access through networked telecentres and libraries, from senior centres and nursing homes (Alcalá, Reference Alcalá2014).

Previous studies on the effectiveness of TSSS have shown that physical access to the internet is insufficient if it is not linked to digital training programmes (Millward, Reference Millward2003; Morris, Reference Morris2007; European Commission, 2010). Therefore, promoting access opportunities implies making the necessary resources available to older people to facilitate internet use and to enable them to obtain benefits (Sourbati, Reference Sourbati2009). For example, the training programmes from Older Adults Technology Services, a non-profit organisation based in New York City, supported the selective shaping of social contexts by optimising resources and by providing means to compensate potential risks (e.g. isolation, lack of support). The participants used more communication tools (e.g. email, texting), which allowed them to create new opportunities for enhanced contact (Kamin et al., Reference Kamin, Lang, Kamber and Kwon2016).

The results of empirical studies that have examined the adoption of new ICTs by older people have highlighted the importance of social relationships in shaping access to technologies (Warschauer, Reference Warschauer2003), and suggest that internet access should be situated locally, within particular organisational and community contexts, and adapted to the rhythm of the users (Seo et al., Reference Seo, Erba, Geana and Lumpkins2017). Community workers in public libraries, senior centres or nursing homes can facilitate or constrain internet access (Sourbati, Reference Sourbati2009).

Participants

In order to respond to the research questions, we considered four types of TSSS used by retired adults in Spain. The participants were retired adults (>54 years) who used the internet – at least once in the previous 3 months – and were regular users of TSSS. The TSSS were situated in four organisational contexts:

• • Nursing homes (NH), which are homes or establishments in which social support activities are developed for older people through collective housing; temporary or permanent use that includes food, health care, hygiene and comfort, promoting co-existence and leisure time for elderly residents. Participating NH had classrooms with internet access and held computer workshops regularly.

• • Senior community centres (SCC), which are non-residential social services facilities intended to promote co-existence among older people, encouraging participation and social integration. They offer sociocultural, occupational, artistic and recreational activities. These centres house classrooms with computers and internet access, and computer workshops are held regularly.

• • University Programs for Seniors (UPS), which arose in the 1970s, were designed to improve the basic skills of older adults in order to promote their social and cultural participation and personal development (Villar et al., Reference Villar, Pinazo, Triadó, Celdrán and Solé2010). UPS offer digital literacy workshops.

• • Adult education programmes (AEP), in which formal educational plans are offered to obtain the official Compulsory Secondary Education diploma. They also offer non-formal education plans. In their educational plans, subjects in the field of communication are offered.

In the absence of reliable data concerning the population of older adults attending these TSSS, we conducted non-probability quota sampling to obtain a similar number of subjects in each TSSS. The authors established quotas of 140 adults over 54 and internet users in the four types of organisational context (N = 560). The selection of the centres and participants was random among the users of the four TSSS in three cities of autonomous communities which have a low, medium and high per capita income and ageing index (INEbase, 2019a, 2019b): Huelva, Asturias and Madrid. Although the systems are not mutually exclusive (one user can be a user of multiple supports), the participants were exclusive users of one of them.

In order to obtain an unbiased sample within the same type of TSSS, our team of researchers administered a maximum of five questionnaires with the prior authorisation of the entity's administrator. We administered the questionnaires during the third week of October 2018.

The mean age of the respondents was 67.78 years old. Broken down by type of context, the mean age of the respondents in NH was 73.28, in SCC 69.74, in UPS 63.98 and in AEP 63.91. Regarding gender, 56.7 per cent were female, 56.4 per cent being female in NH, 55.6 per cent in SCC, 64.0 per cent in UPS and 50.4 per cent in AEP.

Measurements

In response to the research questions, we controlled for variables relating to the demographic characteristics of the sample (age, gender and co-existence) and socio-economic status (educational level and monthly income). Researchers in the area of digital inequality have found that age, gender, education and income are associated with internet use (e.g. van Deursen and van Dijk, Reference van Deursen and van Dijk2018) and with capital-enhancing activities and digital skills (Hargittai and Dobransky, Reference Hargittai and Dobransky2017). Covariates included chronological age, re-codified into a categorical variable with three categories based on the retirement age (55–64, 65–74 and over 74 years of age), gender (female and male), co-existence (alone and accompanied), educational level (primary, secondary and tertiary studies) and monthly income, taking as a reference the pension bracket in Spain (less than €600, €601–1,200, €1,201–1,800 and more than €1,800).

In order to measure the literacy support received from the TSSS, the Eurostat (2019c) scale was used, with values from 1 (never) to 4 (very frequently). This scale measured the degree of support provided by the institution's educators to older people in order to learn to use the internet (α = 0.95). The variable literacy support was the mean value of the responses of the respondents to all the items on the scale.

To operationalise the dimensions of digital inequality (DiMaggio and Hargittai, Reference DiMaggio and Hargittai2001), we used the following measures:

• • Technical means refer to the availability of adequate equipment to carry out a given online activity (van Deursen and van Dijk, Reference van Deursen and van Dijk2018): we used a dichotomous scale (no = 0, yes = 1) to ask the respondents for the number of internet access devices (smartphone, tablet, laptop, desktop computer, others). The variable technical means was the sum of the responses (minimum = 1 and maximum = 5).

• • Autonomy to access the internet refers to the freedom to use technology when and where one wants without constraint from others, such as lines of library patrons or employer supervision (Hargittai and Hinnant, Reference Hargittai and Hinnant2008; Hargittai and Dobransky, Reference Hargittai and Dobransky2017). We used a dichotomous scale (no = 0, yes = 1) to assess the number of places from where the respondent accessed the internet at least once a week (home, the street, social centres, educational centres and community services). Finally, the variable autonomy of access was the sum of the places of access reported by the respondents (minimum = 1 and maximum = 5).

• • Digital skills refer to skills that enable users to derive the full benefits that access can provide (Hargittai and Hinnant, Reference Hargittai and Hinnant2008). To measure the digital skills, we used a synthesis of the scale by Goldhammer et al. (Reference Goldhammer, Naumann and Keßel2013) – skills for access, management and dissemination of information – related to basic skills, with values from 1 (not true) to 4 (completely true) (α = 0.95). The digital skills variable was the mean value of the responses to each item.

• • Purpose of social connectedness refers to the need for psychological connection with others and to create positive human interaction (Schulz et al., Reference Schulz, Wahl, Matthews, De Vito Dabbs, Beach and Czaja2015). The search for social connectedness is based on the evidence that sustains that older people use the internet to interact socially, either for reasons of emotional selectivity (e.g. Nimrod, Reference Nimrod2013), to optimise their social contacts (e.g. Hogeboom et al., Reference Hogeboom, McDermott, Perrin, Osman and Bell-Ellison2010; Smith, Reference Smith2014; Rainie and Zickuhr, Reference Rainie and Zickuhr2015) or to compensate for the losses derived from ageing (e.g. Minagawa and Saito, Reference Minagawa and Saito2014). We used a selection of items from the scales used by Papacharissi and Rubin (Reference Papacharissi and Rubin2000) – α values greater than 0.78 – and by Ku et al. (Reference Ku, Chu and Tseng2013) – α values greater than 0.85 – on the gratifications obtained through the use of the internet, with values of 1 (disagree) to 4 (totally agree). The variable purpose of social connectedness was the mean value of the responses of the respondents to all the items on the scale.

Table A1 in the Appendix includes more details about each scale's wording and descriptive results as well as the convergent validity of the variables analysed (average variance, factorial loading, composite reliability and Cronbach's alpha of each scale).

Data analysis

In order to answer the research questions, we used sequential binary logistic regression (BLR). The primary objective of this technique is to model how categorical variables influence the probability of occurrence of a dichotomous event. To avoid the problem derived from having categorical predictors with limited cases in each category, we ran descriptive statistics on each of their predictors and we contracted categories. We also made a diagnosis of multicollinearity, which verified that socio-demographic variables and literacy support were not strongly related, showing tolerance values higher than 1 and Variance Inflation Factor lower than 10. Finally, our research team verified outliers or cases that are not well explained by their model. For this, we specified the residuals and eliminated the outliers.

This study will seek to identify the variables that predict the probability of belonging or not to a singular cluster in four types of TSSS (i.e. NH, SCC, UPS and AEP). The dependent variables were technical means (0 = just one access device, 1 = more than one), autonomy of access (0 = up to two access points, 1 = more than two), digital skills (0 = up to the 75th percentile, 1 = above the 75th percentile) and social connectedness (0 = up to the 75th percentile, 1 = above the 75th percentile).

We sequentially included two blocks of independent variables in the models: (a) the socio-demographic characteristics of TSSS users and (b) the frequency of literacy support. The socio-demographic variables were age (55–64, 65–74 and over 74), gender (female and male), co-existence (alone or accompanied), educational level (primary, secondary or university) and monthly income (up to €1,200 or more than €1,200). We then categorised the level of support received for digital literacy into tertiles (Tertiles 1, 2 and 3), with Tertile 3 representing a level of support greater than two-thirds of the respondents.

Table 1 shows the distribution of the respondents among the different categories of the dependent and independent variables in the four organisational contexts. It is noteworthy that in general terms UPS users tend to be younger and of medium-high socio-economic status. NH and AEP users tend to have a medium-low socio-economic level. SCC users are usually older and of an upper-middle level. In addition, it should be noted that the older users are found in NH and SCC. Regarding the provision of literacy support, it should be noted that TSSS users in contexts with a lower socio-economic status (i.e. NH and AEP) reported that they had more support available than in other contexts (i.e. UPS and SCC).

Table 1. Descriptive analysis

Notes: NH: nursing homes. SCC: senior community centres. UPS: University Programs for Seniors. AEP: adult education programmes.