# Chapter 2. Evidence on opportunities and risks for well-being in the digital age

This chapter provides a comprehensive review of the literature on the well-being impact of the digital transformation using the lens of the How’s Life? well-being framework. For each dimension of well-being and the additional dimension of ICT access and use, the key impacts of the digital transformation are discussed and illustrated with the help of available indicators, distinguishing between opportunities and risks triggered by digital technologies. For almost all dimensions of well-being, the chapter identifies both positive and negative effects, suggesting that the digital transformation often has an ambiguous influence on people’s well-being.

The statistical data for Israel are supplied by and under the responsibility of the relevant Israeli authorities. The use of such data by the OECD is without prejudice to the status of the Golan Heights, East Jerusalem and Israeli settlements in the West Bank under the terms of international law.

## Introduction

The digital transformation is often described as the third defining moment in humankind’s history, after the Neolithic Revolution and the Industrial Revolution (e.g. Harari, 2018). In a relatively small number of years, it has changed the way people work, consume, communicate and learn about the world. People now have a digital life and a digital identity. This chapter provides a comprehensive account of the way digital technologies have transformed people’s life, in both good and bad directions, in most of its dimensions.

The chapter builds on an extensive literature review presented through the lens of the How’s Life? well-being framework. Information is also provided on ICT access and usage, which acts as a channel through which all dimensions of people’s well-being are affected by digital technologies. In addition to ICT access and usage, the eleven dimensions of well-being considered in this chapter include: education and skills, income and wealth, jobs and earnings, work-life balance, health, social connections, governance and civic engagement, personal security, environmental quality, housing and subjective well-being. Each section describes the most important opportunities and risks of the digital transformation for each dimension, illustrated by indicators when available.

This review also serves a practical purpose. By identifying the aspects of the digital transformation that are the most important for people’s well-being, it provides a list of issues that should be measured. However, important data gaps still prevent us from capturing the full range of impacts of the digital transformation on people’s life. For example, while the list of important impacts in Table 1.1 contains 39 items, only 33 indicators are currently available (Table 1.2). This implies that several important impacts of the digital transformation on people’s life cannot currently be measured with comparable data.

Finally, several limitations of the analysis presented in this section should be kept in mind. First, the classification of digital impacts in terms of either risks or opportunities is not always clear-cut. For instance, having digital resources at school can constitute an opportunity to build digital education, up to a point where having too many digital resources can distract pupils from acquiring more traditional skills. In this regard, many important nuances are not reflected in the indicators used in this section. Second, the digital transformation is taking place at a very quick pace, while this review is based on evidence and indicators that are often lagging.

## ICT access and usage

For people living in OECD countries, access to the Internet and a mobile device is a pre-requisite to participating in an increasingly digitalised society and economy. Personal digital devices are necessary to benefit from the opportunities offered by digitalisation in each of the dimensions important to well-being. For example, digital skills require familiarity with ICT equipment. The ability to interact with employers, medical services, family and friends is contingent on being connected to the Internet. In short, Internet access is often a key channel through which the digital transformation is impacting upon each of the dimensions of people’s well-being in the digital age.

### Access to digital infrastructures is a prerequisite to reaping the benefits of digital technologies

A lot of progress has been made in the dissemination of Internet access in OECD countries in the last decade. In several countries, Internet access rates at home are now close to 100% (Figure 2.1). In addition, over the last ten years, cross-country differences in Internet access have narrowed markedly. Lithuania, Mexico, Turkey, and a number of other countries have experienced an increase in the share of households with Internet access of 40 percentage points or more between 2005 and 2016, partially as a result of policies in favour of rural areas and disadvantaged population groups (for example by improving telecommunication infrastructures or by providing financial incentives to support usage by disadvantaged groups, OECD, 2017a). Still, on average, more than 20% of individuals living in the OECD do not have Internet access at home. This share is particularly high in Mexico, where over half of people lack Internet access at home, and also exceeds 25% in Japan, Greece, Lithuania, the United States (Box 2.1) and Israel.

Box 2.1. Internet access in the United States

Geographic factors are an important reason for the lack of home broadband access in some countries. The rural-urban divide in Internet access is particularly pronounced in the United States, where the gap in home broadband rates is about 12-13% (Whitacre, Strover and Gallardo, 2015). Wealthy and less densely populated countries such as Norway, Finland and Sweden have achieved much higher broadband access rates. One explanation for this difference is on the supply-side, as rural areas may not be sufficiently serviced by telecommunications networks. Whitacre, Strover and Gallardo (2015) find that in 2011, 13% of the American population living in rural areas had no broadband, compared to only 2% of urban Americans. Demand-side factors also play a role: African Americans and Hispanics in the United States are less aware of the availability of broadband and of the ways to gain access, even when controlling for other demographic factors such as education, income and age (Prieger and Hu, 2007). This suggests that social capital may be an important factor in explaining the digital divide and the inequalities that result from it.

### The diversity of Internet uses brings greater benefits to individuals

Internet access at home is not always a prerequisite to use the Internet. Mobile Internet, which becomes increasingly widespread, allows people to use the Internet without a connection in their home. Indeed, the share of people using the Internet is higher than the share of households who have access to the Internet. 84% of people in OECD countries have used the Internet in the last 12 months (Figure 2.2), and 72% do so every day or almost every day.1

Substantial variation exists in the use of the Internet among different groups, with Internet use heavily shaped by socio-economic factors (Wunnava and Leiter, 2008; Kiiski and Pohjola, 2002), particularly in OECD countries with lower rates of use. Across the OECD, Internet use rates for people in the highest income quartile are 22 percentage points higher than for people in the bottom quartile. Differences in Internet may also exist within a household: a home may have Internet access, without all members of the household using it. This is particularly relevant for differences in Internet use between men and women, which are pronounced in some OECD countries. The gender gap in Internet use is 18 percentage points in Turkey, 10 in Chile and 8 in Italy.2

Age is another important determinant of internet usage. While more than 95% of young people (16-24 year-olds) in OECD countries had access to Internet in 2016, among 55-74 year-olds this share is only 60%. This is a missed opportunity, since there is evidence that the Internet can play a role in achieving positive outcomes for the elderly in dimensions such as health status and social connections (Cotten et al., 2014). In Denmark, Iceland, Luxembourg, Norway and Sweden, Internet usage rates exceed 90% even among the older cohort, while in Turkey this share is only 20%.

However, data on the share of individuals having used the Internet does not capture the sophistication with which people navigate the Internet. Figure 2.3 considers the variety of different online activities that are used by at least 50% of people in each country, giving an indication of the depth of Internet use in different countries. Not surprisingly, people in countries with high Internet penetration rates use it for a larger range of functions. In four countries (Luxembourg and three Nordic countries), nine out of ten online activities are used by a majority of the population. The variety of uses shows that the share of the population using the Internet does not fully reflect the extent to which people use the Internet for important daily tasks. For example, in Chile and Italy, while over 70% of people have used the Internet in the last 12 months, the majority of people only use the Internet for one single activity, suggesting that the sophistication of Internet use remains limited.

### There are inequalities of Internet usage even when there is equality in access

While the majority of people in OECD countries now have access to Internet, the second digital divide remains persistent (Attewell, 2001; Goldfarb and Prince, 2008). This divide in digital skills limits people’s job prospects, but its implications extend to all areas of well-being. In people’s daily lives, this inequality in digital skills manifests itself in the form of different abilities to use the Internet in a variety of ways. All time-saving opportunities, new ways to access information and social networking depend on people’s ability to take advantage of the various possibilities provided by Internet. As Internet access rates are very high in all OECD countries, differences in ability to use the Internet are a key factor of inequalities (OECD, 2010).

Figure 2.4 describes this second digital divide in the form of vertical inequalities in usage of the Internet within countries. In countries with lower Internet access and usage rates, the variety gap is lower because many of these activities have not yet been introduced on a large scale. In other countries, such as Slovenia, Italy and Portugal, there are larger gaps in the intensity at which people make use of the Internet. While some people use the Internet for a very wide range of activities, others barely use it at all. This has implications for the extent to which digital technologies can impact well-being in its different dimensions and for the inequality-enhancing effects that may emerge as a result.

Education is an important factor in determining differences in usage within countries, with tertiary-educated users performing on average 7.3 different activities online as opposed to 4.6 activities on average for people with a lower secondary education (OECD, 2016a). However, conditional on adoption, people from lower socio-economic backgrounds tend to use the internet more intensely than people in high-educated, high-income groups (Goldfarb and Prince, 2008). But while the former use the internet primarily for recreational and entertainment purposes, younger and more educated people use the Internet for more productive activities such as finding jobs, getting access to health care services or engaging in political and social activities (van Deursen and van Dijk, 2014; Putnam, 2015).

## Education and skills

Digital skills are essential for people to reap the benefits of digitalisation and are necessary to participate in a society that relies increasingly on digital platforms to interact with other people and with institutions. Many social and economic transactions now include some kind of digital component. Improved access to health care and government services and the ability to manage digital security and privacy risks all depend on mastering some level of digital skills (Box 2.2). In addition, the digital economy increasingly demands workers who are able to solve problems in technology-rich environments, but who also have the creative and interpersonal skills that foster success in this digital environment. Digital technologies are also transforming the learning experience itself, both in schools as well as in adult education, where opportunities to follow online courses allow people to engage in lifelong learning. However, digital skills are only an opportunity for those who have them, and so while they present an opportunity for people’s well-being, the digital divide in skills also presents a risk at a societal level as the digital skills gap can perpetuate existing inequalities.

Box 2.2. What types of skills are necessary in a digital society?

Three types of skills needs have emerged in the context of the digital transformation (OECD, 2017a). First, everyone in society needs to be equipped with ICT problem-solving skills as well as solid literacy, numeracy and problem-solving skills in order to be able to benefit from using digital technologies in their daily life and in the workplace. This implies making investments to reduce the skills gap between students with and without digital skills (OECD, 2016b). Second, specialised skills are needed to ensure the realisation of the societal benefits of ICTs. The production of ICT products and services and new advances in cloud computing, big data analysis, blockchain and AI are reliant on highly specialised skills. Finally, digital technologies have sparked the demand for additional skills that are complementary to digital technologies, such as creative, social and emotional skills (OECD, 2015a). These skills allow people to use digital social networks without emotional or social harm and to be aware of the risks of extreme Internet use. In the workplace, interpersonal skills and leadership skills, as well as the ability to navigate and leverage the digital economy, are also becoming more important (Deming, 2017). In the context of growing automation, these human-specific skills are growing in demand.

While all three of these skills needs are important, the focus here is on the first, i.e. ICT problem-solving skills, as these skills are the most specific to the digital transformation. ICT problem-solving skills refer to the ability to navigate technology-rich environments and use the Internet in a variety of ways. Specialised skills are important for society, but an individual does not necessarily require them in his everyday life, hence they are more important for the economy at large than for individuals’ well-being. Finally, while complementary skills bring advantages on the job market, as well as in navigating risks related to social connections and mental health, there are currently no adequate measures for them.

### Students and adults need digital skills to participate in the digital economy and society

Today, digital skills are a prerequisite to fully participate in the labour market. For instance, 95% of workers of large businesses in the OECD, and 65% of those in small businesses, already use the Internet as part of their jobs (OECD, 2016b). According to Berger and Frey (2016), ICT skills are necessary in all but two occupations in the United States. At the same time, 40% of people who use simple office software at work indicate they do not have the digital skills necessary for effectively using such tools (OECD, 2017a).

In order to measure the digital skills of adults, the Programme for the International Assessment of Adult Competencies (PIAAC) includes a task-based measure of adults’ abilities to solve the types of problems commonly faced in using ICTs in modern societies. This problem-solving task requires adults to use a variety of computer applications, such as e-mail, spreadsheets, word processors and websites that adults may encounter in their daily life. Scores are classified in different proficiency levels, where Level 1 denotes the capacity to only perform very basic tasks, whereas Level 2 and Level 3 denote medium and high skills. Adults scoring at Level 2 or 3 can solve problems that require the co-ordinated use of multiple applications, can evaluate the results of web searches and can manage unexpected outcomes.3

Figure 2.5 shows the percentage of people with a medium or high score in problem-solving skills in technology-rich environments (Level 2 or Level 3) across OECD countries. Sweden and New Zealand have the highest share of adults with medium or high digital skills, while in Turkey, Chile and Greece, among others, less than a quarter of people have this skill level. In all countries a significant age gap exists when it comes to digital skills. While younger generations (“digital natives”) are increasingly fluent in the use of digital technologies, older people are often left behind. This has severe consequences in other dimensions of well-being, since digital skills are necessary to benefit from many of the opportunities of the digital transformation. In particular, people in the elderly population are at risk of being excluded from key services in the areas of health-care and e-government, which are increasingly reliant on digitalised systems.

### The emergence of a digital skills gap

The increasing importance of digital skills means that inequalities in skills have the potential to perpetuate or even worsen existing well-being inequalities. Figure 2.6 shows vertical (i.e. total) inequalities in digital skills, as measured by the coefficient of variation of the PIAAC problem-solving test score. Countries with high coefficients of variation have lower mean scores and a wider distribution, i.e. a larger gap between those with high and low scores. Turkey, Chile, the Slovak Republic and Korea record high levels of inequality in digital skills, while New-Zealand, the Netherlands and the Nordic countries display greater homogeneity in digital problem-solving scores among adults. The digital skills gap highlights the divisive potential of the digital transformation, and the extent to which the digital transformation currently manifests itself in the form of a skills gap.

### Digital resources in classrooms can help prepare students for a digital society and economy

Besides offering new pathways for learning, schools play an important role in bridging the digital divide and ensure that all children reap the benefits of technological advances. There is evidence that children from different socioeconomic backgrounds use digital technologies differently (Hargittai and Hsieh, 2013). Digital resources in the classroom can serve as an equalising force between students who do and do not have access to digital technologies at home, allowing the latter to catch up with the digital mastery of the former. In addition, Banerjee et al. (2007) suggest that computer-assisted learning especially benefits schools where the quality of teaching is lower, meaning that differences in teacher quality across schools can partially be mitigated by the introduction of digital resources in the classroom.

As shown in Figure 2.7, there are important gaps between countries with widespread digital resources at school such as the Netherlands, New Zealand or Australia, and countries such as Mexico, Latvia or Japan where only half of the students benefit from internet-connected computers in schools (OECD, 2015b). These differences may reflect differences in financial resources across countries, but may also result from conscious choices as a result of an awareness of the risks associated with the use of technology in the classroom.

About one third of lower-secondary students in OECD countries do not use school computers connected to the Internet. In countries with a large digital divide between students with different socio-economic backgrounds differences in the penetration of digital resources at school may exacerbate these inequalities. In most OECD countries, however, parental education is not a strong determinant of students’ access to digital resources at school. This suggests that access to these resources at school is not necessarily dependent on socio-economic background. However, in some countries, e.g. Mexico, students with highly educated parents are more likely to have digital resources at school than those with low-educated parents.

### The adverse effects of digital resources in the classroom may reduce learning outcomes

Digital resources at school may also present risks for learning outcomes. The results of digital learning experiences in schools are somewhat mixed, and many studies report limited or no benefits of digital education (Bulman and Fairlie, 2016; Escueta et al., 2017). According to evidence from the Programme for International Student Assessment (PISA), while using digital resources in the classroom is beneficial for learning outcomes up to a point, too much use of digital technologies in the classroom can have negative impacts on learning outcomes (OECD, 2015b). This negative effect may be the result of greater distractions in the classroom, when students use Internet connection for chatting or playing rather than learning (McCoy, 2013). Unfortunately, no data is available on the distractive potential of digital technologies within schools.

Another way digital resources may not necessarily be conducive to improved learning outcomes relates to the lack of digital skills of teachers, which poses a constraint to computer-assisted learning. When teachers are not familiar with digital technologies, digital resources can form a distraction for both teacher and students (OECD, 2016c). On average, 20% of lower secondary education teachers in OECD countries report that their ICT skills are insufficient (Figure 2.8). In Italy, 36% of teachers report to have a high need to develop their ICT skills, as compared to 8% in the United Kingdom and 10% in Portugal and Canada.

### Online education and digital learning tools can support lifelong learning and new learning models

In addition to changing the educational experience at school, digital technologies have opened up new educational opportunities for all groups of the population through various e-learning platforms offering a wide range of lifelong learning opportunities. According to Kearns (2010), the Internet has engendered a “fourth generation” of distance learning, which allows for large-scale participation and higher- quality online learning. Examples of such online learning tools are Online Educational Resources (OERs), Massive Online Open Courses (MOOCs), digital learning materials, open data, etc. (OECD, 2015a). Such learning opportunities are particularly useful for workers who want to improve their skills in their current job or find a new job, and may thus improve job-to-job mobility (OECD, 2016b). Figure 2.9 shows that the percentage of people having undertaken an online course ranges between 20% in Canada and less than 3% in Turkey.

While online learning was widely hailed as one of the democratising forces of the Internet, these expectations have not quite been met. Some observers have suggested that most successful MOOCs students are from higher socio-economic strata of the population, implying that MOOCs may increase existing educational inequalities (Escueta et al., 2017). Indeed, the share of individuals that have taken an online course in OECD countries is almost twice as high among people in the highest household income quartile than those in the lowest. The only exception to this pattern is Finland, where more low-income people use online courses.

## Income and wealth

While many studies have found a positive relationship between investments in digital technologies and productivity growth (Brynjolfsson, 1993; Brynjolfsson and Hitt, 1995, 1996, 1997; Jorgenson and Stiroh, 2000), the size of this relationship varies across countries, with the United States, Korea and Japan recording higher returns of digital technologies than countries in continental Europe. GDP gains from digital technologies are essential for increasing country-wide living standards. However, what also matters for people’s well-being is whether the income generated is fairly distributed. This section focuses primarily on wage gaps between workers engaged in ICT tasks and those who are not, and on the increase in the “consumer surplus” resulting from the wider consumption choices triggered by digital technologies.

### Digital skills confer a wage premium

Workers with digital skills typically earn a higher wage as a result of these competencies (Lane and Conlon, 2016; Falck, Heimisch and Wiederhold, 2016). PIAAC data show that workers who have no experience in using ICT earn 18% less per hour than those who score at Level 1 in the digital problem-solving test when controlling for individual characteristics, such as education, age and levels of numeracy and literacy skills (OECD, 2017a). Workers with higher digital skill levels (Level 2 or 3 in PIAAC) earn 26% more than those with basic skills, although part of this effect is explained by other skills and higher education: simply having digital skills is not enough to receive a wage premium, as these skills need to be put to use in order to reap the rewards (OECD, 2017a). Finally, Grundke et al. (2018) highlight the need to combine cognitive and non-cognitive skills in order to reap the monetary benefits from digitalisation.

Figure 2.10 shows the labour market returns to ICT tasks, which are particularly large in the United States, Korea, Japan and Ireland, amounting to around 35% when ICT task intensity increases by 100%. Compared with one additional year of education (which yields a wage-return of about 8%), a doubling of ICT task intensity is equivalent to around five additional years of education. However, much lower wage-returns are recorded in Israel, Turkey and Denmark.

Wage benefits to individuals with digital skills imply a growing cleavage between high-skilled workers and those who are struggling to keep up. Technological changes have been cited as one of the main causes behind growing income inequality and the falling labour share. According to the IMF, half of the decline in the share of national income received by workers is due to technological progress (IMF, 2017). Evidence confirms that most job losses have been among middle-skill workers in sectors such as manufacturing in favour of jobs requiring more advanced skills, with higher productivity levels and associated wages (Michaels, Natraj and van Reenen, 2014; OECD, 2017b).

### Online consumption and the sharing economy may increase consumer surplus

Digital technologies have greatly improved the ease of purchasing online products, especially entertainment products such as music, e-books, movies, TV-series, often purchased at significantly lower price than in traditional forms. Across OECD countries, 45% of individuals played or downloaded music or games online in 2014. The consumer surplus4 gains realised by the consumption of online products like digital music, e-books and search engines have been estimated to be as high as USD 500 per person per year (World Bank, 2016). In addition, e-commerce allows consumers to save time and access a wider range of products online. Digital markets benefit consumers through a variety of channels, from lower prices to higher quality of the goods or services consumed (OECD, 2016c). In some countries, online consumption is becoming part of normal daily life, with over 80% of people in the United Kingdom, Sweden and Denmark having purchased goods and services on the Internet in the past year (Figure 2.11). But the use of e-commerce services is not as widespread in other countries, and just about half of individuals across the OECD have purchased online over the year.

Differences in use of online websites to buy goods and services between population groups show how digital skill inequalities and socio-economic differences exclude people from reaping the benefits of the digital economy. While the United Kingdom leads the ranking in online consumption, less than half of low educated people in the country have used e-commerce services. This gap is also large in Korea, the United States and Ireland. While both digital skills and income inequalities may account for these gaps, it is clear that low educated people do find their way to online marketplaces in some countries but not in others.

The sharing economy not only allows people to increase their consumer surplus but also enables them to sell goods and services themselves. The emergence of peer-to-peer platforms such as AirBnB, Blablacar and Craigslist have transformed entire markets (Ahmad and Shreyer, 2016). Studies have shown that peer-to-peer markets can contribute positively to consumer welfare, and that these benefits are larger for consumers below the median income (Fraiberger and Sundararajan, 2017). People in the Netherlands, Iceland and Norway lead in the OECD when it comes to selling goods and services online (Figure 2.12). Among Dutch people, 37% have engaged in a sale online in the last 3 months, compared with only 2% of Greeks. The opportunity to sell goods and products online appears to be used more by highly educated people, with an average gap of 14% between low and high educated people in OECD countries.

## Jobs and earnings

Beyond income security, employment fulfils a number of important roles for human well-being, such as time structure, social contact, a sense of purpose, a valued social position as well as an opportunity for skill use (Jahoda, 1981; Warr, 2007). For this reason, the labour market effects of the digital transformation are among the most significant for people’s well-being. The digital transformation has the potential to generate substantial changes in the composition of the labour market as jobs that require certain skillsets are replaced by a combination of technology and higher skilled labour, or even completely automated. At the same time, digitalisation yields opportunities by creating employment in new and existing industries, with greater job-to-job mobility facilitated by online job search tools. The digital economy also fundamentally changes the nature of work for many people, with fewer jobs exerting physical demands on workers but more jobs placing an emotional strain on desk-workers.

### New jobs in ICT and in other sectors become available

Despite fears for the automation, there is little evidence so far that technological change has led to a net loss of jobs. The emergence of digital technologies has gone in parallel with steadily rising employment rates in most OECD countries (OECD, 2017b). There are theoretical reasons for which technological progress may contribute to job creation (Autor and Salomons, 2018). Efficiency gains and cost-savings may induce job creation within industries by expanding the market and therefore increasing demand. Increased productivity in one sector can also have positive spillovers in other sectors, if this translates into lower prices and higher demand across the economy. While these processes may imply short-term unemployment among displaced workers, they have the potential to generate economy-wide employment gains.

Estimating the economy-wide impact of digital technologies on employment is challenging, however, because the job creating effects of technological change are often indirect. Employment in information industries is not a proxy for the wider employment gains of the digital transformation but it gives some insight into the contribution of information industries to employment (Figure 2.13). Israel and Estonia have the largest share of workers in the ICT sector, representing 6.1% and 5.5% of the labour force, respectively. Computer, electronic and optical products industries account for substantial shares of employment in Korea, Mexico and Switzerland. Employment in information industries relative to other industries has grown strongly in Estonia, Latvia and Lithuania, where ICT-related jobs have contributed to significant job creation.

### Digital technologies may destroy jobs at risk of automation

While the previous section has pointed to the lack of evidence of the negative effects of technological change on total employment so far, a number of authors have argued that ICT-based technological change will be more profound than previous instances of great technological change. This argument is mainly supported by the observation that the labour-saving potential of digital technologies is far greater than in the case of previous technological changes (Brynjolfsson and McAfee, 2011). As a result, automation may, in the future, have much more impactful consequences on the need for human labour than it has so far. For the moment, while a shift away from manufacturing jobs has been observed, this has not translated to overall losses in employment, as middle-skill jobs have been replaced by new high-skill and low-skill jobs (OECD, 2017b).

Concerns of the automation of jobs are warranted, however, at least in order to make the case for the need to invest in the most appropriate skills for the future digital economy. Thus far, estimates of the impact of automation mainly rely on expert’s predictions of the types of tasks that are likely to be replaced by machines. Previous estimates by Autor, Levy and Murnane (2003) quickly proved to be too cautious: tasks that Autor et al. considered to be out of reach for machines, such as truck driving, are already being threatened by rapid advances in machine learning and AI. More recent estimates of the potential job-displacement effects of automation have looked at job tasks rather than entire job categories (Frey and Osborne, 2013). Food preparation assistants, cleaners and helpers, labourers in mining, construction, manufacturing and transport, and assemblers are the most likely to see their job tasks automated, while teaching professionals, health professionals and personal care workers are among the least likely to lose their job to a machine. Similarly, Schwab (2016) and Susskind and Susskind (2015) consider that the work of lawyers, financial analysts, journalists, doctors or librarians could be partially or totally automated. Schwab (2016) emphasises that algorithms made available by AI are able to successfully replace human actions, even creative ones. The author presents the example of automated narrative generation, in which algorithms can conceive written texts for particular types of audience.

While Frey and Osborne (2013) predicted that almost half of all jobs in the United States may be automated in the next 10 or 20 years, their estimates may be overblown as they did not consider the variety of tasks that workers may perform, as well as differences between specific jobs based on their various tasks. Looking at specific jobs and the tasks performed, more conservative estimates suggest that about 14% of jobs across the OECD are likely to be automated (Arntz, Gregory, and Zierahn, 2016; Nedelkoska and Quintini, 2018). Even though these estimates depend on a number of assumptions about the automation of various job tasks, they highlight which countries’ labour markets are vulnerable to the risks of automation.

According to these estimates, the risk of job automation is relatively low in Norway, New Zealand, Finland and the United States and is highest in Slovakia, Lithuania and Turkey (Figure 2.14). The variation among countries is even wider when considering only jobs that are classified as “at high risk of automation” (those with a probability of being replaced by machines of 70% or more). While in Norway, Finland and Sweden, only 5-10% of jobs are at high risk of automation, this value is around 34% in the Slovak Republic and between 20 and 30% in Greece, Lithuania, Slovenia and Spain.

### The digital transformation may lead to job polarisation

In addition to the long-run risks of digital transformation on employment in the form of automation, job polarisation poses a more immediate risk (Mazzolari and Ragusa, 2013; OECD, 2017b). Job polarisation is the outcome of higher demand for high- and low- skilled jobs associated with a decline in the demand for middle-skill jobs. Between 1980 and 2014, industries where the demand for high-educated workers grew the fastest also recorded the highest decline in the demand for middle-skill workers (Michaels, Natraj and van Reenen, 2014). On one hand, middle-skilled workers are outcompeted by high-skilled workers who are needed to operate automated production systems. On the other hand, there is an increase in the supply of jobs in low-skilled service sectors, such as food service workers, security guards, janitors and cleaners, home health aides, child care workers (Autor and Dorn, 2013). The skills needed in these jobs are more difficult to automate, as they are reliant on social interactions or the type of dexterity that is not yet available in machines. Autor and Dorn (2013) found that the number of hours worked in such service occupations among low-educated workers in the United States rose by more than 50 % from 1980 to 2005. The decline in medium-skill jobs may have a range of well-being implications, such as increasing wage inequality, short-term unemployment and lower job satisfaction for workers who have no other alternative than moving to low-skilled jobs.

There is substantial evidence that this polarisation is taking place also outside the United States. OECD studies that use a variety of national and European labour force surveys show that middle-skill jobs are disappearing not just as a result of the shrinking manufacturing sector, but also within almost every industry (OECD, 2017b). Estimates of the extent of polarisation between 1997 and 2007 vary between 9 percentage points in Austria to 2 percentage points in Canada (OECD, 2017b). While both technological change and globalisation contribute to job polarisation, the respective roles of each of them is hard to disentangle. Recent OECD studies suggest that polarisation in the labour market is most strongly associated with the penetration of ICT within sectors, more so than factors associated with globalisation (OECD, 2017b; Breemersch, Damijan and Konings, 2017).

### Online job search helps job seekers find employment opportunities

The Internet has significantly improved the matching process in the labour market through new platforms for job search and recruitment (Faberman and Kudlyak, 2016). While initially there was scepticism about the effects of online job search in reducing unemployment duration (Kuhn and Skuterud, 2004), many recent studies have shown a positive impact of Internet in reducing the job search process. Kuhn and Mansour (2014) found that unemployed persons who look for jobs online found work 25% faster than comparable workers who did not use the Internet. Contacting friends and relatives, sending out resumes, filling out applications and looking for advertisements were all found to be effective channels for job search through the Internet.

Online job search has grown rapidly in countries where Internet penetration is high and access costs low. In the United States, the share of young people who looked for jobs online tripled from 24% to 74% between 1998-2000 and 2008-09 (Kuhn and Mansour, 2014). Figure 2.15 shows the percentage of individuals reporting to have used the Internet to look for a job or send a job application in the last three months. Some countries where general Internet penetration is high have very low rates of online job search. For example, Japan records the third lowest share of online job searchers. One possible explanation is that lifetime employment is still very common in Japan (Sousa-Poza and Henneberger, 2004), whereas a large share of online job seekers in the United States consists of workers with an existing job who wish to change employer (Kurt and Mansour, 2014).

The relationship between online job search and education level varies across OECD countries. In most countries, online job search is more common among high-educated workers, with the OECD average share of individuals looking for jobs online twice as high among educated workers than among low-educated workers. This gap is most pronounced in Greece, Mexico and, notably, Chile, where the difference is 26%. However, in a few countries, notably Estonia, Iceland, Luxembourg and Norway, online job search is more common among low educated workers. This means that online job search may act as an equalising force in some countries more than others.

### Workers with computer-based jobs are less subject to job strain

The digital economy has fundamentally changed the nature of work and people’s work experience. More jobs today involve computer-based tasks, and new modes of work go hand in hand with changing social expectations around the organisation of work. Between 1995 and 2015, the proportion of workers using computers at their job increased from 40% to over 60% (Salvatori, Menon and Zwysen, 2018). These changes may have both negative and positive implications for job quality. For example, computer-based jobs may allow workers to organise their work with more flexibility (Salvatori, Menon and Zwysen, 2018), and present less physical risk factors to workers. Negative associations exist particularly in the emergence of higher emotional demands associated with an increased pace of work.

Data for European countries show that the frequent use of computers, laptops and smartphones at work is significantly and positively associated with task discretion, i.e. the extent to which employees feel that they can organise their work time and methods, flexibility of working hours and lower physical demands.5 On the other hand, intense use of computers is associated with the degree to which jobs involve responding to tight deadlines. On balance, people who frequently use computers at work tend to experience higher quality of the working environment than those who do not.6 This relationship also holds across countries, with countries with more computer-based jobs experiencing lower job strain (Figure 2.16).

On average, across OECD European countries, computer-based jobs are associated with a 4% lower share of workers experiencing extended job strain (Figure 2.17). Workers in Slovakia, Ireland and France benefit particularly from reduced job strain. It should be noted that in countries where extended job strain is low, there is less scope for improvement than in countries where many workers are experiencing extended job strain.

### Jobs in the digital economy may be associated with higher stress in the workplace

The introduction of the Internet and other digital tools in the workplace has dramatically increased the flow of information that workers have to manage. Research has documented new forms of information flows in a large range of work settings, such as investment analysis, managerial decisions, price setting, physicians’ decision-making, aviation, library management and many others, and through a range of digital media, such as e-mail, intranets and push systems (Eppler and Mengis, 2004). The resulting information overload is associated with technostress: “a form of stress associated with individuals’ attempts to deal with constantly evolving ICTs and the changing physical, social, and cognitive responses demanded by their use” (Ragu-Nathan et al, 2008; see also Brod, 1984; Arnetz and Wiholm, 1997). Information overload in the work place lowers job satisfaction and self-reported health status (Ragu-Nathan et al., 2008; Misra and Stokols, 2012). A recent study also linked perceived e-mail overload to burnout and decreased work engagement (Reinke and Chamorro-Premuzic, 2014).

An analysis of the relationship between computer use at work and self-reported job stress suggests that workers with digitalised jobs do experience more stress, even when controlling for earnings, skill level, and sector of employment (Figure 2.18). Because the analysis is pooled across countries, individual country-effects are computed based on the share of workers that have computer-based jobs.7 Countries with more ‘digital jobs’, therefore, by construction have a higher share of workers experiencing job stress associated with such jobs. In these countries (Norway, Luxembourg and Denmark), up to 3% more workers may experience stress at work associated with computer-based jobs. In Turkey and the Czech Republic, where few workers use computers and digital devices at work, the impact on job stress is more limited.

## Work-life balance

Life in the digital age is faster than before the arrival of the smartphone and other tools that provide constant connectivity. People who are constantly connected complain that there are not enough hours in the day. This is somewhat paradoxical, because many applications of digital technologies aim at saving time (Wajcman, 2014): mobile technologies can aid in navigation and shorten travel times, instant messaging services allow for faster communication, and peer-to-peer services improve access to services, for example by improving their geographic reach (World Bank, 2016; OECD, 2016d). But the Internet and its applications have also increased the volume of the activities people engage in. As a result, the changing speed of life engendered by the digital transformation may have effects on people’s experience of their work-life balance and indirectly on their mental health and experienced well-being.

The Internet and mobile devices have blurred previously rigid lines between work and time spent outside the workplace. Thanks to home broadband connectivity, many people are now able to work from home. Teleworking possibilities reduce commuting times and allows workers to combine work and family life more easily, especially in multi-earner households (Eurofound and ILO, 2017; Dettling, 2016). At the same time, receiving e-mails on a computer at home, or on a mobile device anywhere, allows work to protrude into the private sphere like never before. In some cases, workers are expected to be available at any time (Mazmanian and Erickson, 2014). The ability to connect from anywhere has changed the way people experience time in general, the nature of the relationship between work and home life, and people’s family relations.

### Teleworking allows people to save time and combine their work and personal lives

Teleworking, on the other hand, may present an opportunity for work-life balance as it improves time management and may reduce time spent commuting. A variety of studies have found that employees who engage in telework have higher job satisfaction (Kelliher and Anderson, 2009; Brenke, 2014). Among positive effects, teleworkers report reduced commuting times, more flexibility in organising their working time, and better overall work-life balance (Eurofound and ILO, 2017). Billari, Giuntella and Stella (2017) also found that German women who have high-speed Internet access at home are better able to attain their desired number of children, by reducing the time constraints associated with combining work and parenthood. Evidence from the American Time Use Survey shows that reductions in the time spent commuting and in home production due to Internet increase labour force participation, in particular among married women (Dettling, 2016).

Teleworking requires both technological and cultural transformations in organisations, and the scope for both of these transformations varies across countries. According to Eurofound and ILO (2017), employer attitudes are an important determinant of the penetration of teleworking. The share of workers having teleworked at least once is highest in Denmark and the United States, while more than 90% of people never teleworked in Italy and Turkey (Figure 2.19). Among European countries, employers are particularly open in Nordic countries (where more than one third of workers have teleworked).

Like other opportunities of the digital transformation, the possibility to work remotely also gives rise to new inequalities. Due to the different nature of job tasks, teleworking is almost exclusively available to high-skill knowledge workers (Billari, Giuntella and Stella, 2017). Germany ranks below OECD average in terms of teleworking penetration, which is perhaps explained by cultural factors: Brenke (2014) estimates that teleworking would be theoretically possible for about 40% of German jobs, but hypothesises that teleworking is less accepted by companies than in other countries. In some other countries, particularly Luxembourg, Austria, Switzerland and Norway, teleworking is also significantly more common among male workers than among female workers.

### Constant connection to work may increase worries about work when not working

Being constantly connected to work increases risks of stress for workers. Even if the time spent at work does not change, workers may be occupied with job-related tasks even after returning home. Some studies have shown that people who check their e-mail more often experience more day-to-day stress and lower levels of positive affect (Kushlev and Dunn, 2015). A study of working adults in the United States showed that both the time spent on e-mails and the organisational expectations put on staff to monitor their e-mails after working hours lower people’s satisfaction with their work-life balance (Belkin, Becker and Conroy, 2016).

Using data from the European Working Conditions Survey it is possible to get an estimate of the relationship between the frequent use of computers at work and the share of European workers who experience work-related worry at home.8 In countries where more workers have computer-based jobs, more workers experience worries about work outside work time (Figure 2.20). Those countries with more digital jobs, such as Denmark, Luxembourg and Norway are more exposed to the potential increase in worries about work after work time.

## Health

Digitalisation can affect people’s health status through the emergence of new physical and mental health risks and through its impact on the health-care delivery system. Health risks associated with the digital transformations include mental health problems associated with the extreme use of digital technologies, especially among children and teenagers and the crowding out of other activities such as physical exercise. Health-care delivery is also affected by new digital technologies, such as electronic records, new treatment options, tele-care and teleconsultation. An important aspect of digitalisation concerns the production and use of medical data to improve the effectiveness and efficiency of health systems. As a caveat, the exchange and use of medical and health data must meet high data protection and data security standards, considering its sensitivity. How and where care is delivered is also affected by digital innovations, which challenges the traditional role of care providers, with implications for interactions among care providers and between providers and patients. The effects of these changes in health-care delivery of health inequalities are potentially large, but also less well documented.

### Extreme use of digital technologies may have negative mental health effects

The effects of mobile phones, video games, and the pervasiveness of ubiquitous screens on the mental health of children and teenagers have drawn significant attention in the public debate because they may present risks of addiction (James et al., 2017). Extreme Internet use, defined as children who spend more than 6 hours on the Internet outside of school, is becoming more common among children and teenagers, with time spent online by 15-year-olds increasing by about 40 minutes between 2012 and 2015 on average in the OECD (Hooft Graafland, 2018). Rosen, Carrier and Cheever (2013) found that the iGeneration members (the generation grown up in an environment where technology is ubiquitous) check their social media accounts on average every 15 minutes. While video games used to be the primary source of extreme use of digital technologies, the smartphone has extended this risk to a wider range of applications. A recent study found that 39% of 18- to 29-year-olds in the United States are online “almost constantly” (Pew Research Center, 2018a).

Research suggests that the Internet triggers neurological processes similar to other addictive substances and activities, i.e. experiences of short-term pleasure in the brain’s “reward center” (Cash et al., 2012). This area releases a combination of dopamine, opiates and other neurochemicals when activated, a mechanism that can be compromised over time due to the deterioration of associated receptors, requiring even more stimulation to get a similar response. Children and teenagers, for biological reasons, are more susceptible to addiction because their brain is still in development. For example, a study among 14-year-olds in Belgium found that frequent gamers had brain abnormalities compared to other teens, potentially resulting from dopamine releases associated with video games (Kühn et al., 2011). However, other researchers warn that it is premature to approach Internet disorders from an addiction perspective as it is not clear that the behaviours of Internet users share the pathological characteristics of an addiction disorder (Kardefeldt-Winther, 2017).

There is evidence of a direct link between extreme Internet use and depression and anxiety (Kotikalapudi et al., 2012), but the nature of this relationship is disputed and is likely to be bi-directional, as people with anxiety, depression and other mental health problems are also potentially more likely to spend time online. A longitudinal study run on 3 000 children in Singapore found that extreme video game use and problems such as social phobia, attention deficit disorder, anxiety and depression often occur together and are likely to be mutually reinforcing (Gentile et al., 2011). Results from the PISA study show that extreme use of the Internet among children is associated with lower life satisfaction and school results, even when controlling for socio-economic backgrounds (OECD, 2017c). Overall, the consensus is emerging that digital technologies can provide benefits to children and teenagers up to a certain point, but that extreme use can have harmful effects (Przybylski and Weinstein, 2017).

Extreme Internet use among young people is common in OECD countries. On average, 24% of 15-year-olds spend more than 6 hours a day on the Internet on weekend days, and a figure that is as high as 43% in Chile and 37% in the United Kingdom (Figure 2.21). Culture may play a role in the extent to which children spend long periods of time online, with the lowest share of extreme users among children in Japan and Korea. The level of educational achievement of parents also seem to be associated with extreme Internet use, with children of highly educated parents less likely to be extreme Internet users in most countries. There are a few exceptions to this – notably Chile, Latvia, Mexico and Lithuania – where extreme Internet use is more common among children with high educated parents, possibly reflecting an income effect.

The nature of social interactions online may also be conducive to mental health problems such as anxiety, depression, bipolar-mania and narcissism. Such problems are thought to be triggered by social comparisons made online (Sabatini and Sarracino, 2017). Social comparisons online may induce jealousy and lower self-esteem (Muise, Christofides and Desmarais, 2009; Krasnova et al., 2013; Blachnio, Przepiorka and Benvenuti, 2016). Online social media are especially conducive to sharing life experiences, which leads to a phenomenon knowns as the “fear of missing out”, i.e. the “pervasive apprehension that others might be having rewarding experiences from which one is absent” (Przybylski et al., 2013). This feeling has been found to be more common among frequent users of social media (Alt, 2015). Twenge et al. (2018) report a strong increase in suicides among teenage girls in recent years and link suicide risk to frequency of social media use. However, these are only correlations and contradictory evidence exists, e.g. a decline in teenage suicide rates between 1990 and 2015 in most OECD countries (OECD, 2018a).

There is also some evidence that the use of digital technologies can have direct or indirect negative effects on various aspects of people’s physical health. Use of digital technologies may crowd out activities that are important for people’s health, such as physical exercise or sleep. Using historical variation in telecommunication infrastructure that affected broadband penetration, Billari, Giuntella and Stella (2018) find that broadband Internet access has a direct effect on the duration and quality of people’s sleep, as people use digital devices for entertainment while facing time constraints due to family and work commitments. Moreover, it has been found that blue light, emitted by the screens on many digital devices, may disturb people’s circadian rhythms and reduce sleep quality, particularly when using digital technologies in the evening (Hatori et al., 2017).

### More effective health-care delivery due to improved communication with service providers

Digital innovation may positively impact health outcomes in two ways. First, the delivery of health care is improved by more systematic development and use of electronic records and online accessibility of health care providers. Second, health care treatments are changing rapidly with the introduction of new technologies such as remote sensors, robotics, genomics and artificial intelligence (OECD, 2017d). These new technologies, as well as greater research capacity in biology and drug development are increasing the potential for improved health outcomes. However a pre-condition for these new technologies is a trustworthy and secure infrastructure.

Use of digital technologies in the health care sector has the potential to improve the delivery of care, improving patients’ experience and achieving cost efficiencies. These “enabling technologies”, are facilitated by process innovations, new eHealth technologies (Box 2.3), and the use of Big Data in decisions about treatments (OECD, 2017e). However, evidence of the effects of eHealth on health outcomes and the experience of health care delivery is still inconclusive (Black et al., 2011; Slev et al., 2016). While eHealth has been found to have positive effects on some aspects of the patient’s experience, little or inconclusive evidence exists on the impact on treatment outcomes, although some evidence of positive effects exists with respect to self-reported health and depression among cancer patients who used e-health technology (Slev et al., 2016; Johansen et al., 2012).

Box 2.3. eHealth: Technological advances in health care

Use of digital technologies in health-care (eHealth) encompasses “the application of information and communications technologies across the whole range of functions that affect the health sector” (European Commission, 2012). This broad definition covers a number of digital applications:

Electronic health records (EHR) allow for seamless data exchange between patients, health care providers and pharmacies in order to improve operational efficiency and provide more personalized care. EHRs support both individual patients and doctors by accessing care more smoothly, but can also advance research through the use of large amounts of (anonymised) data on the effectiveness of treatments. Another aspect of EHRs is ePrescribing, which aims to exchange accurate, understandable and error-free prescription information between doctors, patients and pharmacists (Cooke et al., 2010).

Telehealth and mobile health solutions (mHealth) have a wide range of functions, from making online appointments, to accessing health information and communicating with health professionals online, to connecting patients in peer-support groups. Telehealth can also be used to monitor chronic conditions remotely, personalize treatment, and make treatment adjustments without the need to go to the hospital (McKinsey, 2014).

Wearable devices and sensors are increasingly using digital communication tools to send and receive data in order to provide targeted care. These innovations include both the incorporation of already existing technologies, such as pacemakers, into the Internet of Things, as well as new health monitoring devices such as wearable watches that record continuous information on patients’ vital health signs.

Together, eHealth technologies serve the multiple goals of improving the patient’s experience, increasing the efficiency of the health care system, freeing up time for doctors and care providers to improve patients’ health outcomes, and helping physicians make better clinical decisions using new diagnosis and treatment tools supported by large amounts of new data and intelligent systems that can perform more sophisticated analyses and overcome biases.

Unfortunately, very few indicators are currently available to measure the benefits people get from digital-related process innovations in health care. An imperfect proxy indicator is the share of people who make medical appointments online. Web-based medical appointment systems have a number of positive impacts, including improving patient satisfaction, reducing no-shows and wait times (Zhao et al., 2017). Data on the share of individuals who make medical appointments online is available for some European countries (Figure 2.22). While in Austria and Iceland very few people make medical appointments online relative to the extent of Internet access in the population, in Spain, Finland and Denmark especially this figure is much higher. Still, in all countries for which data is available, only a minority of people makes medical appointments online.

Besides improving the patient experience and generating potential for more sophisticated analyses of health data for the benefit of improved treatments, eHealth can reduce health care costs by achieving administrative efficiencies and improve outcomes by reducing wasteful spending due to over-diagnosis and unnecessary treatments (OECD, 2017f). However, implementation of eHealth systems comes at a cost, and the transition from analogue to electronic health records may be burdensome to the health sector as innovations encounter financial, legal, social and ethical barriers to implementation (Ross et al., 2016).

### Digitalisation of health technologies can contribute to better health outcomes

Digital technologies can also help improve health outcomes through breakthroughs in research and treatment options facilitated by new monitoring systems and by the use of Big Data and AI (Box 2.4). For example, the deployment of advanced genome sequencing techniques using embedded data-mining algorithms reduced the costs of generating human genome sequences from USD 1 million to USD 1 000 in five years (OECD, 2016a). These discoveries may lead to new and improved treatment options in the future. Analytics also can help drug and device developers identify how patients respond to treatments in more sophisticated ways (McKinsey, 2014). Such advances are facilitated by enormous improvements in computing capacity in combination of new analytical tools that the digital transformation enables.

The direct effects of such digital applications on health outcomes are difficult to estimate. Some studies have evaluated how innovation in medicine in general (beyond just digital innovation) has impacted objective health indicators, such as life expectancy. Cutler and McClellan (2001) found positive effects in the treatment of heart attacks, low birth weight infants, cataracts and breast cancer. According to a study from the US President’s Council of Advisors on Science and Technology (2012), medical innovations have added 5 years of life expectancy between 1980 and 2010. Others have argued that more expenditure on medical innovations does not necessarily lead to improved outcomes in many areas, and point to the high opportunity cost of health care spending that could be used in other areas important for well-being (Berndt, Fisher and Rajendrababu, 2003; Lichtenberg, 2004). Because of the complexity in estimating the direct effect of digital technologies on health outcomes, no explicit measure is presented here.

Box 2.4. How Big Data are transforming healthcare: From eHealth to iHealth

In health-care, several types of data sets are available and can be linked to each other: hospital inpatient data, information from cancer and mortality registers, prescription data, information on primary care and long-term care provision, patient outcomes and diabetes data files. Another example concerns the data generated by patients through mobile applications, which can improve knowledge about patient health status, disease progression and level of function. Combining different data sets can improve acute care analytics and predict the risk of re-admission at hospital or forthcoming complications. Moreover, better and linked data can also help planning infrastructure and workforce needs, predict demand fluctuations and help assessing the efficacy of expensive technologies.

Together, the increased availability of data allows for a progression from eHealth to intelligent health, or iHealth (Berrouiget et al., 2018). iHealth employs Big Data and intelligent systems to make smarter and more efficient decisions at the individual or population level. In the United States, Kaiser Permanente used 15 years of maternal and neonatal data to develop a risk-stratification tool that detects sepsis in neonates, and leads to a reduction in antibiotic administration within 24 hours following birth. Combating the spread of antimicrobial resistance through spatial detection is facilitated by big data analysis (Vong et al., 2017). As an example, the US FDA assesses medical technology risks using a very large database (about 178 million people), while hundreds of clinical registries on diseases and interventions have been linked in Denmark (Schmidt et al., 2015).

### Health information online can improve patient experiences

Increased availability of health information online is one of the most direct ways in which the digital transformation impacts people’s health experience. Many different platforms provide individuals with access to information about diseases, educate patients with treatment options and aid in decision-making, provide support for physical and emotional problems and allow for peer support (Slev et al., 2016).

There is mixed evidence of the health and well-being impacts of these new sources of health information. Bessiere et al. (2010) found that looking for health information online is associated with depression, potentially due to poorly conducted self-diagnoses. Another hypothesis is that (unmoderated) peer platforms may increase mental health problems as they can spread wrong information. Other studies have found evidence of positive benefits from internet support groups on depression symptoms in cancer patients and survivors (Gysels and Higginson, 2007), people with AIDS (Mo and Coulson, 2010), Parkinson’s disease (Attard and Coulson, 2012), and diabetes (van Dam et al., 2005). Hong, Pena-Purcell and Ory (2012) found no effects of online cancer support on self-reported health, but positive effects on self-reported quality of life.

There are large differences between countries in the share of people who use the Internet to seek health information (Figure 2.23). Across the OECD, 45% of Internet users look for health information online. In a number of countries, led by the Netherlands, Luxembourg and the Nordic countries, more than half of users employ the Internet as a source of health information. Use of online health information is, in many OECD countries, strongly affected by the education level of individuals. On average, the share of people with high education accessing health information online is more than double the share of those with little or no education. As a result, potential health benefits remain concentrated in this segment of the population.

## Social connections

The Internet and the smartphone have fundamentally changed the way people interact with each other. As with the arrival of previous technologies such as the television or the telephone, the effect of digital technologies on social connections has been the subject of significant debate. Putnam (2000) attributed part of the decline of social capital in the United States to an increase in time spent watching television. Several early studies noted that Internet use was associated with higher self-reported loneliness (Kraut et al., 1998; Hamburger and Ben-Artzi, 2000; Morahan-Martin and Schumacher, 2003). Since then, this causal relationship has been questioned, and new evidence has emerged that the Internet role in changing social connections may be more positive.

### Increased online interactions among friends and social networks

Two competing hypotheses exist to describe the effect of the Internet on human interactions. On the one hand, some researchers have argued that the Internet displaces social interactions from the real to the virtual world (Kraut et al., 1998; Nie and Erbring, 2002; Nie and Hillygus, 2002). An early study in the United States used a longitudinal sample of first-time computer users to show that the use of Internet crowded out family time and offline social interactions (Kraut et al., 1998).9 The two causal pathways identified by the study were, first, a displacement of social activity, with less frequent offline contact and, second, a displacement of strong ties, with strong relationships offline replaced with more superficial ones online. Dienlin, Masur and Trepte (2017) also show that mobile devices have removed pretexts for offline encounters: where people used to meet in person for sharing photos, planning events or gossiping, such functions are now moved to the virtual world.

The competing hypothesis is that the Internet reinforces offline relationships and that computer-mediated communication increases offline contact and social capital (Shklovski, Kraut and Rainie (2006); Boase et al., 2006; Johnston et al., 2011; Burke, Marlow and Lento, 2010). By increasing the overall volume of communication, online communication also facilitates increases face-to-face interactions (Dienlin, Masur and Trepte, 2017). In this sense, the rise of the Internet has commonalities with the arrival of the telephone, which greatly enhanced social connections (Fischer, 1992). Various studies have supported this conclusion. A study of 1 210 Dutch adolescents found that those who spent more time using instant messengers also spend more time in face-to-face interactions (Valkenburg and Peter, 2007). A positive effect of social network use on face-to-face interactions was also found in a longitudinal study using a nationally representative sample of the German population (Dienlin, Masur and Trepte, 2017).

One way through which the Internet may enhance bridging social capital is through the formation of online communities (see, however, the discussion of disinformation and “echo chambers” further below). By connecting people with a shared interest, regardless of demographic characteristics or geographic location, the Internet allows forging of new bonds and creating new groups of association. This pattern, while destructing previously existing social networks, allows for the formation of new circles of individuals sharing various commonalities (Rainie and Wellman, 2012). For example, online weight-loss support groups allow individuals to encourage each other in achieving a shared goal (Hwang, Ottenbacher and Green, 2010). Such networks may complement real-life networks. After 9/11, Dutta-Bergman (2006) found that people who engaged in support groups online were also more involved in support communities offline.

The opportunity to create bridging social capital extends to new face-to-face encounters between individuals. The Internet emulates the “strangers on the train” phenomenon, where the transient nature of the environment allows individuals who do not know each other to feel more comfortable in engaging in conversation (Bargh and McKenna, 2004). This does not mean that these encounters remain offline. According to data from the US “How Couples Meet and Stay Together Survey”, the Internet is displacing traditional venues for meeting partners, such as the neighbourhood, the friends-circle and the workplace (Rosenfeld and Thomas, 2012). People with Internet access in the United States were found to be more likely to have a romantic partner than people without Internet access, suggesting that more people may partner thanks to new ways of finding someone online (Rosenfeld and Thomas, 2012).

The proliferation of mobile devices, coupled with Internet may reduce some of the displacement risks of the computer, as smartphone-mediated communication can take place while commuting, cooking, or being at work. Individuals’ use of social applications on mobile devices increase social capital, particularly among the younger generation (Cho, 2015). For example, in a study of Israeli students, Whatsapp was found to strengthen social capital by allowing students to keep in touch with their existing contacts (Aharony, 2015). On the other hand, a study based on the Italian Multipurpose survey showed that the smartphone can interfere with the quality of real life interactions (Rotondi, Stanca and Tomasuolo, 2017). Using natural field experiments Misra et al. (2014) showed that people rated conversations through traditional devices as significantly superior than those based on a smartphone.

Despite the mixed insights from the literature, substantial evidence supports the idea that online social contact does complement offline interactions, especially when considering the active use of social networks (Howard, Rainie and Jones, 2001; Valkenburg and Peter, 2007; Johnston et al., 2011; Aharony, 2015; Dienlin, Masur and Trepte, 2017). In addition, in European countries, data from the European Quality of Life Survey highlight a moderately strong cross-country correlation between frequent internet use and people’s satisfaction with their social life (Figure 2.24). When distinguishing between daily and weekly users, the benefits of Internet use are greater for daily users than for weekly users.

The benefits of the Internet for social connections are most likely the result of online social activity. Figure 2.25 shows the percentage of Internet users who have accessed social network sites within the last three months. Social network usage among Internet users is highest in Iceland, where almost 90% of users have accessed a social network site in the last three months, and is lowest in France, where only slightly over 40% of users did so. Age is a strong predictor of social network use. While 84% of young people (aged 16-24) in the OECD use online networking sites, the same share among 55-74 year-olds is just 31%.

### The Internet may help people overcome loneliness and social exclusion

Kraut et al. (1998) suggest that people with good pre-existing social skills will particularly benefit from online social networks, whereas those with limited social skills may only feel more excluded. By contrast, Hamburger and Ben-Artzi (2000) argue that introvert people actually benefit more from the Internet as it removes some of the barriers of traditional social interactions. Across countries, there is a strong inverse cross-country relationship between Internet use and loneliness, with people living in countries with higher levels of Internet penetration experiencing lower levels of loneliness (Figure 2.26).

One area that should be highlighted is in the potential decrease in loneliness among older adults who use digital technologies. Social isolation is a major and growing problem for the elderly, as a result of higher life expectancy in old age, lower number of offspring, and patterns of living. A growing body of evidence points to the beneficial role that the Internet and online social networks can play to overcome loneliness among the elderly. Feelings of loneliness also have detrimental effects on their health outcomes, for example in relation to dementia (Holwerda et al., 2012). The Internet can also help combating social exclusion for marginalised groups, as the anonymous nature of online interactions can help reduce the barriers to finding people with similar experiences (McKenna, Green and Gleason, 2002).10

### Discrimination against minority groups using hate speech

On the other hand, social media platforms and other online fora also provide a space for negative social interactions given the comparatively lower barrier to participation than in the case for real life interactions. Because of the Internet’s anonymous or detached nature, people may engage in negative social behaviour more easily than in real life. Online harassment, discrimination against some population groups, or even criminal offences can be facilitated by social media platforms and may be as, if not more, harmful as offline. Little data exists on the prevalence of these types of harmful online behaviours across countries (see next section on Governance and civic participation). As regards discrimination, an analysis based on machine learning of 19 million tweets in the United Kingdom identified almost 5 million cases of misogyny during the four-year period of the study (Ditch the Label, 2016).

### Cyberbullying and online harassment can negatively impact the social experiences of children

Bullying can have detrimental consequences for children’s mental health and subjective well-being and can, in extreme cases, lead to suicide (Juvonen and Graham, 2014). Cyberbullying can be more harmful than traditional forms of bullying because the reach of humiliation is expanded to a large audience online, and because words and images can remain online indefinitely (Nixon, 2014). The link between cyberbullying and mental health problems has been extensively documented (Elgar et al., 2014; Mirsky and Omar, 2015; Lindert, 2017).

Measuring the prevalence of cyberbullying is difficult. Most surveys rely on self-reported information, which face inherent problems as victims may not be willing or able to report. Figure 2.27 presents the latest data on children experiencing cyberbullying from the Health Behaviour in School-Aged Children (HBSC) survey. Although this is the best source of data on this phenomenon, these data may underestimate cyberbullying rates if children do not feel comfortable answering survey questions in the school environment.11 On average, 9% of 15-year-olds reports having experienced cyberbullying at least once in their life, with girls reporting victimisation more often than boys in all countries except in Denmark, Israel and Spain. Cyberbullying is particularly prevalent in a number of Eastern European countries and in Ireland and the United Kingdom. Conversely, children in Greece, Iceland and Germany report relatively few instances of cyberbullying.

## Governance and civic engagement

New avenues for information and communication online are changing the way that individuals and governments express themselves and communicate with each other, receive and disseminate information and interact in the provision and uptake of public services. The consequences of these changes are complex: while people’s expression of their political opinions online is a form of increased engagement in democratic and political processes, it may also increase polarisation of political views, spread misinformation and compromise people’s trust in institutions. At the same time, increased transparency of the government through open data may contribute to accountability and trust. The Internet has also opened up new ways for governments to provide services to citizens through e-government and digital government.

### Increased engagement of citizens in civic and political communities

A healthy political system requires a public sphere that allows people to express their opinions, challenge the government, and engage in policy-making processes. This public sphere can serve as a place for deliberation or exchange of ideas as well as a venue for interest groups to exert influence over the political system (Grömping, 2014). The digital transformation has extended the public sphere by allowing people to express themselves politically and engage in political communities online, both on social media as well as on dedicated platforms (e.g. forwarding e-mails, sharing opinions about politics and current events, engaging with politicians on social media pages, joining collective actions online; Di Gennaro and Dutton, 2006). The Internet has been instrumental in capturing the attention of formerly disengaged voters, as witnessed for example in both Barrack Obama and Donald Trump’s presidential campaigns as well as in political parties on all sides of the spectrum in Italy, Brazil, Israel and other countries (Heaney, Newman and Sylvester, 2011; Campante, Durante and Sobbrio, 2013).

In principle, the extension of citizens’ engagement in societal and political communities online is an opportunity of the digital transformation. Online engagement can draw more people into the political debate, as it requires fewer resources to participate and removes traditional barriers. The idea that online exposure to political debates increases political engagement finds support in studies that have found positive associations between the online exposure to political discussions and offline political participation (Gil de Zúñiga, Molyneux and Zheng, 2014). The Internet also allows people to exert pressure on political processes through online petition platforms such as Change.org12 or Avaaz13 as well as through government-backed political participation platforms such as DemocracyOS14 in Argentina (Mancini, 2015). Conversely, it has also been suggested that online political engagement may crowd out traditional forms of political participation (Christensen, 2011).

Minority groups may particularly benefit from opportunities to express their voices online. The Internet has allowed people from all walks of life not only to get news online but also to create content, motivated by a human need for self-expression (Krishnamurthy and Dou, 2008). According to traditional social psychology, people engage in protest to express grievances stemming from relative deprivation, frustration or perceived injustice (Berkowitz, 1972; Gurr, 1971; Lind and Tyler, 1988), and the Internet may serve as an exhaust to express these emotions and engender change. For example, both the #MeToo movement and the Black Lives Matter movement in the US had a strong online component. According to a large study on social media activism conducted by the Pew Research Center, over half of black social media users in the United States considered social media as personally important for expressing their political views (Pew Research Center, 2018b). In France, the Yellow Vests movement has emerged and diffused on the internet.

While there may be personal psychological benefits to political engagement online, the idea that online action has positive real world outcomes is disputed. While the Arab Spring was often cited as an example of the democratising power of the Internet (Wheeler, 2006), subsequent events in the Arab World engendered more scepticism on the role of online platforms in empowering people. While historically, social movements were characterised by offline organisation, social media make it possible for individuals to rally around a particular cause in an ad hoc, bottom-up way before making their voice heard “in real-life”. For example, the April 2018 strikes of teachers across several US states built momentum through Facebook, giving citizens agency in their relationship between themselves and with the state (Slocum, Hathaway and Bernstein, 2018).

One often-cited drawback of the Internet is that the online platforms may act as echo chambers that limit people’s exposure to alternative views. This would inhibit the Internet’s capacity as a place of deliberation in the public sphere, and lead to increased political polarisation. Such views rely on the finding that people rarely seek out information that opposes their own and that their online interactions may filter information that corresponds to their views (e.g. Grömping, 2014). But recent studies that consider multiple media outlets (as opposed to single-platform studies) have suggested that most people are not in echo chambers, and that the majority actually uses the Internet to broaden their horizon (Dubois and Blank, 2018). It is also tempting to ascribe the increasing popularity of populists to social media and new campaign strategies facilitated by the Internet. But critics point out that the causes of populism are much more complex than that, and that online campaigning strategies are used across the political spectrum, not just by populist parties (Postill, 2018). More research is actually needed to untangle the complex relationships between the use of digital technologies and the formation of political views.

The issue of political engagement online is complex and the role of digital technologies in transforming political processes cannot be easily compared across countries. It is possible, however, to consider the extent to which people engage in civic and political discussions online. Comparable data on the online expression of political opinions exists for several European countries. In all these countries, the share of individuals using the Internet for posting opinions on civic or political issues is less than a quarter of the population (Figure 2.28). Online engagement is most common in Denmark, Iceland and Luxembourg, with fewest people active in online political discussions in the Czech Republic, Austria and Slovenia. Young people are substantially more engaged in political discussions online than the older generation, suggesting that different pockets of the population may engage with political issues in different ways.

### Changes in information channels and the spread of disinformation may lower people’s trust in others and in the government

The Internet and social media have not only changed the way people express themselves, but also how they access information about societal and political issues. According to a recent survey, 38% of individuals in OECD countries used Facebook as a source of news in the previous week (Newman et al., 2018). In several countries, the digital transformation and changes in how people get information have coincided with a sharp decline in people’s trust in traditional media and the government. The strongest evidence of this decline occurred in the United States, where trust in the media fell from 72% in 1976 to around 50% at the turn of the century, and to 32% in 2016. Meanwhile, people’s trust in the federal government in the United States is at an all-time low of 18% in 2017 (Pew Research Center, 2017). This pattern does not hold, however, for all OECD countries, and the mechanisms through which new sources of information (and disinformation) impact trust in institutions are still poorly understood.

The Internet’s disruptive force in the relations between the public, the media and the government is likely to be significant. Contrary to traditional media such as newspapers and the television, the Internet allows information to be instantly updated at low cost (Best and Krueger, 2005). Lazer et al. (2018) point out that this has disrupted the role of media as providers of objective and balance information that emerged in response to the widespread use of propaganda during the First World War. While the Internet harbours an opportunity for democracy (by enabling outsiders to challenge existing political norms and give a voice to people that were previously underrepresented), it also challenges the will and ability of voters to base their political judgments on facts, as opposed to false or overly simplistic messages, which Internet tends to spread (Persily, 2017).

One widely discussed medium through which this occurs is the spread of (unintentional) misinformation and (intentional) disinformation.15 Disinformation is effective because humans are inherently ineffective at recognizing deception and show confirmation bias (Rubin et al., 2015). The combined uptake of social media and low trust in traditional media create in some countries an optimal environment for disinformation to spread (World Economic Forum, 2013). However, the impact of disinformation on democratic outcomes has not been proven. Persily (2017) points out that the observation that disinformation exists does not prove its impact, and that almost no research exists on the long run impact of disinformation. Alcott and Gentzkow (2017) calculated that for disinformation to have swayed the most recent US presidential election, a single deceptive article would need to have the same effect as 36 television ads, which is indicative of the continued dominance of the television as a source of information.

It is equally unclear whether the Internet and disinformation are (partially) responsible for declining levels for trust in institutions in some countries. One study found that the consumption of news from online sources is associated with higher trust in government, but information from social media is associated with lower trust (Ceron, 2015). In addition, while trust in government has declined over time in some countries, this is not the case for all OECD countries, and many have seen recent rebounds of trust in government recently (OECD, 2017f). There does seem to be a relationship, however, between the level of exposure to disinformation and trust in government across countries (Figure 2.29). Self-reported experiences of disinformation are higher in countries where trust in government is lower. It is unclear whether this negative relationship is the result of lower trust due to disinformation, of respondents in less trusting countries being more aware of disinformation sources, or perhaps of deeper institutional and societal factors that steer countries to different equilibria of good governance, trust and resilience against disinformation.

Variation in self-reported exposure to disinformation among OECD countries is surprisingly high, with more than a third of people indicating to have been exposed to disinformation in Greece, Mexico and Hungary, as opposed to less than 10% in Denmark, Germany and the Netherlands (Figure 2.30). This variation suggests that societal and political factors may be more conducive to the spread of disinformation. It should also be noted that the measurement of self-reported disinformation is contingent on the assumption that people’s ability to identify disinformation is equal across countries. Indeed, it is possible that people are more likely to self-report experiences of disinformation in countries with lower levels of trust in traditional media. If this is the case, self-reported exposure to disinformation may represent an environment of distrust in information sources, rather than the spread of actual misinformation per se.

### Open data allows for improved transparency and accountability of government

One way in which the digital transformation can mitigate potential declines in people’s trust in institutions is by encouraging greater openness of governments to share information that was previously hidden from the public eye. One of the most significant contributions of the digital transformation in the experience of government is through information on government processes and decisions disseminated via government websites. By opening up data on expenditures, actions and outcomes, governments can increase transparency and support greater accountability of their decisions. Beyond strengthening the interaction between government and citizens, digital technologies can also move countries towards an “open state”, where the “executive, legislature, judiciary, independent public institutions, and all levels of government” work towards an open government culture (OECD, 2017g; Box 2.5), thereby strengthening trust in public institutions and reinforcing civic engagement.

Box 2.5. The OECD Recommendation on Open Government

Digital technologies enable governments to renew their interaction with citizens to an extent and a scale impossible beforehand. The digital era may also encourage a renewal of democracy that puts direct approaches and deliberation at its heart. Online platforms allow governments to interact with citizens from all corners of the country and to publish government information, creating new possibilities for stakeholder participation and transparency. In this sense, governments are moving to what the OECD Recommendation on Open Government calls a “culture of governance that promotes the principles of transparency, integrity, accountability and stakeholder participation in support of democracy and inclusive growth” (OECD, 2017g). By adopting an open government culture, civic engagement is put at the heart of government’s interaction with their citizens, thereby enhancing their well-being. Large-scale online consultations on legislation and rule-making are now a common feature in OECD countries. Citizens may also be called upon to provide their views on investment decisions in their communities through participatory budgeting projects, empowering them to be actors in their communities’ developments.

The OECD OURdata Index assesses governments’ efforts to implement open data in the three critical areas: openness, usefulness, and re-usability of government open data (Figure 2.31). The index is based on responses provided by public officials in member countries on government efforts to ensure that public sector data are available and accessible to citizens, and to spur a greater re-use of this data (Ubaldi, 2013). Among countries for which data is available, Korea, France and Japan seem to perform particularly well in providing access to open government data to citizens. The Korean government has taken a number of initiatives to improve access to open data by passing national laws on open data and organising specific events to help citizens make use of available data.

### Digital technologies can improve government service delivery

Digital government strategies provide an opportunity for governments to improve the delivery of public services and foster better ties with citizens at different stages of the policy-making cycle. This is particularly important for minority groups and people who rely heavily on government support for their livelihoods, for whom improved service delivery may require access to new information, benefits, and ways to get their voices heard. In the early stages of the Internet, e-government services focused on providing specific services to citizens, ranging from the digital collection of taxes, payments of fines and dues, applications for public benefit programmes, permits and licenses, and more (Warf, 2014). Recently, governments have been implementing digital strategies in a more integrated manner in order to encourage citizen involvement (OECD, 2017g). Digital government involves a more strategic use of digital tools, using both technological and organisational innovations in government administrations in order to improve their accountability and reliability.

Use of e-government services has become widespread in a number of OECD countries (Figure 2.32). In the Nordic countries, the Netherlands and Estonia, at least three quarters of the population reports to interact with public authorities online. In a second group of countries (e.g. Germany, Spain and the United Kingdom), e-government services exist and are used by about half of the population. Across the OECD, 46% of individuals reported to have made use of e-government services in the past year, indicating that while e-government services are more frequently used than in the past, digitalisation of government services is still to work in process.

### Exclusion from digital government services due to lack of skills

While digital government offers opportunities for improved service delivery and increased engagement of citizens in government processes, inequalities in digital skills inequalities pose a challenge for the fair distribution of these benefits, excluding from these digital services some of the groups that could benefit the most. For example, the Dutch Council of State recently issued a warning to the government that it risked excluding citizens from accessing certain public services if it completely replaced traditional service provision by digital government platforms. Spire (2018) has found similar results for France especially in rural areas. In addition, many governments face difficulties in successfully implementing digital government platforms. One World Bank study suggested that up to 87% of public ICT projects could be considered as failures or partial failures (World Bank, 2016). Not only does the delivery of poor quality services impede the well-being of citizens who cannot consume services but it also damages the social contract with the state, undermining the trust that might otherwise exist.

While reasons for not making use of e-government services include not having the need to, preferring the real life contact, not trusting the online service, lack of skills is a key barrier for individuals in accessing e-government services (Figure 2.33). In a number of European OECD countries for which this data is available, 5% of respondents did not access e-government services due to lack of skills, a share that reaches 10% in Hungary. In Estonia, a country where the government has implemented a comprehensive digital strategy, fewer than 1% of respondents indicated lacking the skills needed to use e-government services, even among people with low education. Lack of skills inhibits low educated people to access e-government services significantly more often than the more educated.

## Personal security

Personal (and public) security is one of the primary challenges posed by the digital transformation for individual well-being (Gluckman and Allen, 2018). In the absence of effective regulatory, legal and ethical frameworks, Internet users and organisations can be exposed to substantial economic, social, emotional and even physical risks. Trust in digital tools and applications are essential for reaping the well-being benefits of the digital transformation. This is particularly the case when it comes to the protection of personal data. While individuals are very concerned about their privacy online, they are not always able to protect their personal data adequately themselves. The increased pervasiveness of digitalisation in everyday life also means that digitally powered tools may interfere with people’s physical security.

### Digital security incidents may compromise people’s online safety and compromise trust

Online security risks are likely to have a more indirect impact on the well-being of individuals than offline (physical) security threats. While the economic risks of online security threats are significant, the primary well-being risk of cyber-security risks is in acting as a deterrent for people to take advantage of the benefits that the digital economy offers. If people do not feel secure online, they will be more reluctant to engage in the digital economy, inhibiting this from unlocking its full potential. In other words, diminished trust may impede the effectiveness of digital solutions. According to various surveys in North America and Europe, trust and concerns about digital security are a growing concern for individuals (OECD, 2015c).

A similar trade-off takes place at the societal level, where digital opportunities are weighed against the security risks that they present. In the health-care sector threats to digital security stand in the way of taking full advantage of the benefits of digitalisation (Büschel et al., 2014). Electronic health records and digitally enabled health devices have much potential to improve health outcomes; however, their uptake partly depends on the ability of health care providers to secure new digital medical devices. While no malicious cases have been reported yet, there are concerns over the potential of cyber-security threats to essential medical devices such as automatic insulin pumps or pacemakers (OECD, 2013; Sametinger et al., 2015). The benefits of such devices for individuals thus depend on the capacity of health care providers and governments to guarantee their security.

Governments are also increasingly conscious of the necessity of securing essential infrastructures against privacy threats. Several malicious online practices have targeted governments, businesses and individuals, motivated by profit-making, activism (“hacktivism”), political goals, espionage and sabotage (OECD, 2012). In a 2014 OECD survey, governments identified such security threats as their second highest priority area in the realm of the digital economy, out of 31 possible areas (OECD, 2015c). While such cyber-attacks primarily target large organisations, individuals also face indirect consequences, which may again compromise their trust.

The measurement of cyber-security risks is challenging as online criminal activity may go unnoticed and because there is no centralised reporting mechanism for small-scale online security incidents. To measure individual experiences of cyber-security threats, self-reports remain the most reliable technique, despite possible limitations in how respondents understand these questions. In addition, high self-reports may reflect the efforts of respondents to raise awareness on cyber-security issues, rather than high prevalence of online security threats per se. Figure 2.34 shows the share of individuals who report having experienced an online security incident in the last three months. On average, about one in five people in OECD countries reported to have experienced a cyber-security incident, with higher shares in France, Luxembourg and Hungary. People in New Zealand, the Czech Republic and the Netherlands report the least number of incidents.

### Individuals are at risk of data privacy violations in various domains

One aspect of digital security that is particularly important for individuals and that may have direct consequences for their well-being is the protection of people’s personal data. Concerns about privacy online have naturally emerged from the increasing amount and diversity of personal data generated as a result of digitalisation, and of the increasing fluidity of data across geographies, organisations and systems (Büschel, 2014). Surveys indicate that this is a main concern for individuals in most OECD countries. A large majority of Europeans, according to a Eurobarometer poll, are worried that their personal information is not kept secure by websites, and 85% of them think that the risk of becoming a victim of cybercrime is increasing (European Commission, 2016). Similarly, 91% of Americans think that consumers have lost control of their personal information and data (OECD, 2015c).

Data privacy concerns stem from various forms of abuse of personal data, including national identity data (see Box 2.6), which occur both outside and within the law. The past years have seen multiple cases where personal data of large numbers of individuals were exposed as a result of a malicious attack, poor data security, or accidental publication of user data. Such data breaches have become increasingly prevalent, with the UK government estimating that 81% of large British organisations suffered a security breach in 2014 (UK Department for Business Innovation and Skills, 2014), and the Canadian Office of the Privacy Commissioner reporting that the number of data breaches more than doubled during the 2013-14 fiscal year (OECD, 2017h).

However, these breaches of data are not the only reason why people are growing increasingly concerned about the security of their personal data. The improved capacity of companies to use big data analytics to make predictions about people’s preferences and behaviour can lead to psychological, emotional, economic or social harm to individuals (Kshetri, 2014). As an example, Facebook Likes have been shown to predict personal characteristics such as sexual orientation, ethnicity, religious and political views, personality traits, intelligence, happiness and potential addictions (Kshetri, 2014; Crawford and Schultz, 2013). Such “predictive privacy harm”, if used for the wrong reasons, can carry significant risk for people’s physical safety and mental health, contributing to people’s growing unease about the potential abuse of their personal data.

Box 2.6. Digital identity

People’s digital identity is one of the areas where the tension between government’s concerns about security and citizen’s privacy is most obvious. Governments often needs simple means of confirming the identity of a citizen, and in many countries this has led to the establishment of a central database of citizen’s information. As the digital transformation allows the delivery of increasingly sophisticated and secure services that remove the need for face-to-face interaction, these databases have become common in many countries. However, they are also an attractive target for criminals.

In India, the national ID database Aadhaar, contains biometric identity data for more than 1.1 billion citizens. Anyone in the database can use their data, or thumbprint, to access private services like bank accounts or companies like Amazon; whilst membership is optional, those who are not enrolled cannot access basic government services. Despite repeated criticism about vulnerabilities of the platform, and an inadequate approach to information security, the Indian government has so far failed to enact legislation to protect the data of its citizens (Dixon, 2017).

Within the European Union, the electronic IDentification, Authentication and trust Services (eIDAS) regulation provide standards to enable interoperability of identity. This model enables the reuse of an identity verified according to one government’s approach in accessing a service provided in another country, thereby supporting freedom of movement across the Union.

Digitalisation and inter-connection of national identity databases present a well-being trade-off for citizens who lack trust in the security of their private data. When government services become exclusively available to citizens who have a digital identity, this compromises people’s feelings of online security or exclude them from important government services.

While privacy concerns are well-founded in the light of big data breaches and potential privacy harm, the extent of people’s worries has been challenged by the observation of a paradox between the concerns that they indicate in survey questions and their online behaviour. This privacy paradox has been illustrated by empirical evidence showing that individuals are often willing to sell personal information for relatively small rewards (Kokolakis, 2017). One study has found that Internet users value items of their browsing history at about EUR 7, much less than they value offline personal information such as age and address (Carrascal et al., 2013). People generally felt positive about their personal information being used to improve services, but negative at the idea that this information might be sold by service providers.

Privacy preferences also show a large heterogeneity across countries, age groups, and over time. Cultural norms may also affect the degree to which people value privacy across countries (Ardichvili et al., 2006). For example, East Asians are more careful about disclosing sensitive personal information than Westerners (Lin et al., 2013). People in more individualistic societies (such as the United States) have also been found to have lower privacy preferences than people in more collectivistic ones (such as Germany; Bellman et al, 2004). Moreover, privacy preferences are likely to change over time, with the generation of digital natives having different preferences for sharing information than older generations (Elahi, 2009).

Data-driven innovation has also increased the risk that privacy breaches could inflict economic, psychological and social harm to individuals. At the same time, the speed of the digital transformation and the rise of Big Data (together with a lagged regulatory response) do not allow strong conclusions on the well-being impacts of privacy concerns. While some researchers have argued that privacy concerns could diminish in the future, others have made the case that survey results on privacy concerns show “an undervaluation of privacy as a social value” (Hallinan, Friedewald and McCarthy, 2012). Given the incomplete understanding of self-reports, this section focuses on cross-country measures on privacy violations experienced by individuals.

Figure 2.35 shows that personal experiences of abuse of private information, as measured by self-reported violations, is relatively rare in most OECD countries: on average, 3% of individuals, report having experienced an online privacy infringement incident, with higher shares (above 5%) in Korea and Chile. These self-reported measures inform about the prevalence of digital security incidents, but not on their severity.

### Blockchain may enhance the safety of transactions and information exchange

While the digitalisation of information exchange in areas like health, government and economic interactions can pose threats to individuals’ security and privacy, technologies themselves offer potential solutions to improve digital security. One emerging technology that could improve security gains in the future is Blockchain (i.e. distributed ledger technologies), a decentralised database of information allowing networks of actors to safely conduct transactions and exchange information between each other, without the risk of corruption of data. This is because evolutions of the data are available to all blocks of the chain, and each transaction is recorded and must be approved by every node in the network. In addition, the encryption of the data makes any infringements of personal data highly unlikely. The benefits of these technologies can have numerous different applications. For example, blockchain technologies have already been used to record ownership of land parcels, in public procurement processes and even in ensuring the integrity of election processes. Finally, blockchain has been at the core of crypto-currencies, which have proved to be useful already for cross-border payments, and may have a utility for cross-machine payments in the internet of things. On the other hand, crypto-currencies have also been used for criminal purposes such as money laundering and tax evasion.

## Environmental quality

The environmental impact of the digital transformation can take a number of forms, both positive and negative. From a measurement perspective, estimating this impact is challenging for several reasons. Establishing a link between digital technologies and environmental outcomes such as air pollution is difficult, given the many different contributing factors to such outcomes. This section therefore focuses on impacts that relate to the use of resources in human consumption and production systems that have a negative environmental impact. This assessment is informed by a large body of literature that suggests that the production and consumption of technological products and the associated resources that are required to power these processes both have a substantial ecological impact and contribute to observed changes in the climate system (Cook et al., 2016; IPCC, 2013).

### Higher energy efficiency and de-materialisation of consumer products can lower energy and resource use

The impacts of digital technologies on the environment can be classified in three ways (Berkhout and Hertin, 2004). First, direct impacts of the increased use of digital technologies refer mostly to the increased use of resources associated with the production and consumption of digital products and are therefore mostly negative. Second, indirect effects stem from the improved efficiency and de-materialisation of technological, but also on the demand effects associated with falling prices and the proliferation of ICT devices used in daily life. Finally, the digital transformation may induce structural societal and behavioural effects that result from fundamental changes in society and the economy. This section presents evidence mainly on the direct and indirect effects of the digital transformation, suggesting that the current and predicted impacts of the digital transformation are likely to weigh more heavily on the environment than its effects in relieving existing pressures.

Digitalisation of production processes and consumer goods allows for substantial efficiency gains. Modern production systems rely on a variety of digitally-enabled technologies such as electronic sensors, microprocessors, optimising algorithms that reduce resource costs between 1 and 2% per year (Berkhout and Hertin, 2004). Similarly, Computer Assisted Design has brought about large efficiency gains since the 1990s, for example by reducing the use of aluminium in drink cans by about 50% (Berkhout and Hertin, 2004). More recently, improved analytics facilitated by Big Data have allowed efficiency gains in organisational processes (Bengtsson and Agerfalk, 2011). Other emerging technologies, such as 3D printing and industrial robots, which both rely heavily on advanced intelligent systems, are projected to generate further resource efficiency gains (IEA, 2017).

Efficiency gains also take place in the consumption of goods and of energy by consumers. The heating, lighting and powering of residential and commercial buildings uses up over two-thirds of all electricity used in industrialised countries, and smarter systems powered by sensors allow for reductions in this area (Berkhout and Hertin, 2004). The digital transformation has also led to the de-materialisation of parts of the entertainment industry as consumption of music, books and films increasingly rest on virtual media. Another example of de-materialisation due to digital technology is the replacement of a large range of individual products, such as the digital camera, radio, music player, calculator, flash light and the telephone by the smartphone. Transport is another area where energy efficiency gains are expected in the future as a result of the uptake of automated, connected, electric and shared (ACES) mobility (IEA, 2017). Such technologies could reduce energy usage by improved navigation and driving efficiencies.

### Digital technologies generate rebound effects that increase energy use

However, it is unclear whether the relative efficiency gains in resource use described above more than offset the impact of absolute increases in the demand for new and existing products. Starting with the former, digital technologies simply involve the creation of a range of new producer and consumer products that require physical and energy resources. Between 2006 and 2016, the number of Americans that owned multiple ICT devices has grown from 73% to 95% (Baldé et al., 2017). ICT products consist of a large number of components, from micro-chips, semiconductors and circuit boards to liquid crystal displays and batteries. A typical personal computer may contain 1 500 to 2 000 components sourced from around the world (Berkhout and Hertin, 2004).

The expansion of the Internet of Things and of the networks and data centres that support it also lead to growing energy demands, with estimates of this increase depending on scope and underlying assumptions. Osburg and Lohrmann (2017) estimate that ICT-related power consumption will increase in Germany from 59.6 TWh in 2010 to over 90 TWh in 2020. Others have calculated that the electricity consumed by digital devices is growing more than twice as fast than global electricity demand (van Heddeghem et al., 2014) and that globally, ICT electricity consumption will rise to 21% of total consumption in 2030, a four-fold absolute rise since 2010 (Andrae and Edler, 2015). Blockchain is also increasing electricity demand; the computing power for encryption associated with the Bitcoin network (blockchain’s most well-known application) now approaches the electricity consumption of Ireland (De Vries, 2018).

In addition, energy and resource savings associated to the consumption of immaterial goods are likely to increase demand for material (technology) products. This rebound effect results from lower resource prices due to greater efficiency and new demand stimulated by the better management of time, money, labour and infrastructure (Berkhout and Hertin, 2004). For example, projections of the impacts of ACES mobility vary from a 45% decrease in road transportation energy demand to a doubling of demand, depending on the indirect demand side effects that allow private travellers to spend more time in their car as a result of increased comfort and reduced driver burden (Wadud, MacKenzie and Leiby, 2016). Another example that illustrates the complexity of understanding the environmental impacts of digital technologies are e-books, where environmental gains depend on the quantity of paper books replaced by an electronic model, the duration of use of the device, and other factors (Gensch, Prakash and Hilbert, 2017).

### Increased waste of electronic products

Electronic waste or e-waste is one measurable impact of the digital transformation in the dimension of Environmental Quality. While smartphones have now replaced a previous generation of digital cameras, calculators and other electronics, this advantage is reduced by the larger number of digital devices that are used by individuals, businesses and governments, and by the rate at which digital devices are replaced. The environmental impact of producing digital equipment is significant, and is much higher than the cost of its use – the manufacturing of a smartphone accounts for 73% of its carbon emissions (Greenpeace, 2017). Osburg and Lohrmann (2017) estimate that replacing an electric device with a device that is 10% more efficient will offset the environmental impact from the new device only after 33-89 years. However, the average replacement cycle of smartphones is estimated at about 21.6 months per device in the US and 20.4 months in a number of European countries (Kantar Worldpanel, 2016).

These consumption patterns represent a substantial environmental burden. The Global E-Waste Monitor collects data on e-waste generated per capita across countries (Figure 2.36). Globally, e-waste has increased over time in absolute numbers as well as per capita, and this trend is projected to continue. In 2016, 44.7m metric tonnes of e-waste were generated with only 20% of all e-waste being collected and recycled (Baldé et al., 2017). Available data suggests that e-waste generation is highest in the United Kingdom, Denmark and the Netherlands, where individuals produce almost 25kg of e-waste per person per year. In Turkey, Mexico and Chile, e-waste generation is below 10kg. The partnership behind the Global E-Waste Monitor also notes the poor quality of official statistics on e-waste and calls for improved international harmonisation on this front.

## Housing

Housing is important for well-being to meet basic human needs such as shelter from bad weather and to provide people with security and privacy. In addition, poor housing conditions, whether due to overcrowding or in the form of health hazards due to poor building standards, can be detrimental for people’s health and mental well-being (Balestra and Sultan, 2013). The digital transformation can do little to improve or detract from the function of housing in facilitating these needs. However, smart home appliances can increase the efficiency of people’s home management, contribute to energy savings and comfort in the home. It is arguable, however, that some of these functions are not essential for people’s well-being, and that the contribution of digital technologies to better housing conditions remains to be analysed.

### Smart Home Technologies can improve house management

The Internet of Things (IoT) technology is expected to change people’s life at home through the interconnection of familiar appliances (e.g. washing machines, television, sound systems) that are made “smarter” through the inclusion of sensors and other AI softwares. The diffusion of this technology is still in its infancy and related data is not available yet, which is why no indicators of the impacts of digitalisation on housing are included in Table 1.2.

OECD (2018b) outlines the key benefits and risks associated with IoT in the “smart home”. Benefits for users include: 1) smart residential systems are more convenient as some household tasks can be automated; 2) they improve energy efficiency as they are able to cut on unnecessary energy consumption (e.g. at night); 3) they provide enhanced home security and safety regarding physical threats; 4) and they allow for a high degree of customisation as devices respond to user preferences. On the other hand, IoT devices raise some risks for smart home residents, such as data privacy, cybersecurity threats, limited interoperability, the need for lifetime product support, complex supply chains, liability regimes, and product safety.

## Subjective well-being

Digital technologies have transformed people’s lives in every dimension, but have they all contributed to a better life in the eyes of people themselves? This section discusses the positive and negative associations between digital technologies, in particular the Internet, and the life satisfaction component of subjective well-being (Box 2.7). The empirical evidence suggests that people with access to the Internet enjoy a higher life satisfaction than people without access to the Internet, even when controlling for income and education. The results presented in this section, however, should be interpreted with caution: positive associations are based on cross-sectional data and do not imply a causal relationship between digital technology use and subjective well-being, even if the empirical framework controls for a significant number of individual characteristics such as income, education, age, gender and labour force status. In particular, these results do not claim that well-being has increased as a result of the emergence of digital technologies. Rather, they show evidence that people who are more digitally connected report higher levels of life satisfaction.

### Internet access is associated with higher life satisfaction

It is difficult to assess the long-term impact of the Internet and digital technologies on life satisfaction due to the lack of long-term panel data that includes variables on Internet access and use. However, a number of studies have attempted to estimate this relationship using cross-sectional data (Dolan, Peasgood and White, 2008; Kavetsos and Koutroumpis, 2013; Graham and Nikolova, 2012; Lohmann, 2015). These studies find a consistently positive relationship between internet use and life satisfaction at the individual level. Most of these studies look at the relationship between life satisfaction and Internet access or use variables. Penard, Poussing and Suire (2013) expanded the Luxembourg European Values Survey with more detailed questions on the frequency of Internet use, finding a generally positive relationship between Internet use and life satisfaction, and no difference in the positive effect between heavy and light users of the Internet.

Box 2.7. Subjective well-being measures and the digital transformation

To better understand the potential effects of digital technologies on subjective well-being, it is important to distinguish between three components of subjective well-being: life satisfaction, an evaluation that people make of their life as a whole, affect, a term used to describe people’s emotional states (positive and negative) at a point in time, and Eudaimonia, which refers to people’s ability to reach their potential and their assessment of the meaning and purpose of their life (OECD, 2013).

Life satisfaction

Life satisfaction is a measure of people’s satisfaction with their life as a whole. It is closely related to the economist’s concept of utility, but affected by the way people recall life experiences. It is a useful measure to compare experienced quality of life between different population groups or across countries. A body of research has shown that most differences in life satisfaction between nations are explained by differences in objective life conditions, such as health outcomes, education, personal relationships and income (Diener, Inglehart and Tay, 2013).

Affect

Affect captures the joys and sorrows of day-to-day life and is most closely related to what people may describe as happiness at a given moment in time. Positive and negative affect measure how people experience live at a given moment, rather than how they remember it. Differently from life satisfaction, affect is a multi-dimensional measure and has at least two distinct dimensions: positive and negative.

Eudaimonia

Eudaimonic well-being refers to people’s psychological flourishing and the extent to which they can attain a degree of self-actualization. This is the least studied component of subjective well-being, and few studies have explored the relationship between Internet access and eudaimonic well-being. Eudaimonic well-being is less well understood than the other components; it is not clear, for example, whether it is a uni-dimensional concept or represents a range of related concepts (OECD, 2013).

In these studies, multiple pathways are described through which Internet access may affect life satisfaction. Newly accessible goods and services providing indirect and direct benefits are a potential reason for which the Internet may increase life satisfaction (Hong, 2007; Penard, Poussing and Suire., 2013). The benefits of social networking sites on social relationships are another often mentioned potential source of increased life satisfaction (Valenzuela, Park and Kee, 2009; Pittmann and Reich, 2016; Apaolaza et al., 2013). Chan (2015) shows that voice and online communication on mobile phones also have positive associations with subjective well-being through increased bonding and bridging social capital. Finally, there may be indirect pathways through which the Internet would increase subjective well: increased flexibility of work, improved access to medical and governmental services, the ability to find romantic relationships online, or opportunities of gain new knowledge and skills through online courses.

An analysis using microdata from the 2013 EU-SILC Well-being module suggests a positive relationship between Internet use and life satisfaction in European countries, consistent with findings from previous studies. The EU-SILC Well-being module includes a self-reported question on Internet access alongside measures of subjective well-being and a large range of demographic covariates, which allow estimation of the life satisfaction gains associated with Internet access. A full explanation of the empirical strategy and results is in Annex 2.A. The analysis shows that people with Internet access report a life satisfaction 0.28 points higher (on a 0-10 scale) than those who lack access to the Internet. Figure 2.37 reports the effect on life satisfaction when the population moves from zero to the current level of Internet access. By construction, countries with the highest number of self-reported Internet connections, such as Iceland, the Netherlands or Norway, rank highest in terms of life satisfaction benefit.16, 17

## References

Aharony, N. (2015), “What’s App: A social capital perspective”, Online Information Review, Vol. 39, No. 1, pp. 26-42.

Ahmad, N. and P. Schreyer (2016), “Measuring GDP in a digitalised economy”, OECD Statistics Working Papers, No. 2016/07, OECD Publishing, Paris, https://doi.org/10.1787/5jlwqd81d09r-en.

Alcott, H. and M. Gentzkow (2017), “Social media and fake news in the 2016 election”, The Journal of Economic Perspectives, Vol. 31, No. 2, pp. 211-236.

Alt, D. (2015), “College students’ academic motivation, media engagement and fear of missing out”, Computers in Human Behavior, Vol. 49, pp. 111-119.

Andrae, A.S.G. and T. Edler (2015), “On global electricity usage of communication technology: Trends to 2030”, Challenges, Vol. 6, No. 1, pp. 117-157.

Apaolaza, V. et al. (2013), “The relationship between socializing on the Spanish online networking site Tuenti and teenagers’ subjective wellbeing: The roles of self-esteem and loneliness”, Computers in Human Behavior, Vol. 29, No. 4, pp. 1282-1289.

Ardichvili, A. et al. (2006), “Cultural influences on knowledge sharing through online communities of practice”, The Journal of Knowledge Management, Vol. 10, No. 1, pp. 95-107.

Arnetz, B.B. and C. Wiholm (1997), “Technological stress: Psychophysiological symptoms in modern offices”, The Journal of Psychosomatic Research, Vol. 43, No. 1, pp. 35-42.

Arntz, M., T. Gregory and U. Zierahn (2016), “The risk of automation for jobs in OECD countries: A comparative analysis”, OECD Social, Employment and Migration Working Papers, No. 189, OECD Publishing, Paris, https://doi.org/10.1787/5jlz9h56dvq7-en.

Attard, A. and N.S. Coulson (2012), “A thematic analysis of patient communication in Parkinson’s disease online support group discussion forums”, Computers in Human Behavior, Vol. 28, No. 2, pp. 500-506.

Attewell, P. (2001), “The first and second digital divides”, Sociology of Education, Vol. 75, No. 3, pp. 252-259.

Autor, D.H. and D. Dorn (2013), “The growth of low-skill service jobs and the polarization of the US labor market”, The American Economic Review, Vol. 103, No. 5, pp. 1553-1597.

Autor, D.H., F. Levy and R.J. Murnane (2003), “The skill content of recent technological change: An empirical exploration”, The Quarterly Journal of Economics, Vol. 118, No. 4, pp. 1279-1333.

Autor, D.H. and A. Salomons (2018), “Is automation labor-displacing? Productivity growth, employment, and the labor share”, Brookings Papers on Economic Activity Conference Draft, March 8-9, 2018.

Baldé, C. et al. (2017), “The global e-waste monitor 2017: Quantities, flows and resources, international telecommunication union”, www.itu.int/en/ITU-D/Climate-Change/Documents/GEM%202017/Global-E-waste%20Monitor%202017%20.pdf (accessed on 20 July 2018).

Balestra, C. and J. Sultan (2013), “Home sweet home: The determinants of residential satisfaction and its relation with well-being”, OECD Statistics Working Papers, No. 2013/05, OECD Publishing, Paris, https://doi.org/10.1787/5jzbcx0czc0x-en.

Banerjee, A.V. et al. (2007), “Remedying Education: Evidence from two randomized experiments in India”, The Quarterly Journal of Economics, Vol. 122, No. 3, pp. 1235-1264, https://doi.org/10.1162/qjec.122.3.1235.

Bargh, J.A. and J.Y.A. McKenna (2004), “The Internet and social life”, The Annual Review of Psychology, Vol. 55, pp. 573-590.

Belkin, L.Y., W.J. Becker and S.A. Conroy (2016), “Exhausted, but unable to disconnect: After-hours email, work-family balance and identification”, Academy of Management Proceedings, Vol. 2016, No. 1.

Bellman, S. et al. (2004), “International differences in information privacy concerns: A global survey of consumers”, Information Society, Vol. 20, No. 5, pp. 313-324.

Bengtsson, F. and P.J. Agerfalk (2011), “Information technology as a change actant in sustainability innovation: Insights from Uppsala”, The Journal of Strategic Information Systems, Vol. 20, No. 1, pp. 96-112, https://doi.org/10.1016/j.jsis.2010.09.007.

Berger, T. and C.B. Frey (2016), “Did the computer revolution shift the fortunes of U.S. cities? Technology shocks and the geography of new jobs”, Regional Science and Urban Economics, Vol. 57, Issue C, pp. 38-45.

Berkhout, F. and J. Hertin (2004), “De-materialising and re-materialising: Digital technologies and the environment”, Futures, Vol. 36, No. 8, pp. 903-920.

Berkowitz, L. (1972), “Frustrations, comparisons, and other sources of emotion arousal as contributors to social unrest”, The Journal of Social Issues, Vol. 28, No. 1, pp. 77-91.

Berndt, D.J., J.W. Fisher and R.V. Rajendrababu (2003), “Measuring healthcare inequities using the Gini Index”, Proceedings of the 36th Hawaiian International Conference on System Sciences.

Berrouiguet, S. et al. (2018), “From eHealth to iHealth: Transition to participatory and personalized medicine in mental health”, The Journal for Medical Internet Research, Vol. 20, No. 1, https://doi.org/10.2196/jmir.7412.

Bessiere, K. et al. (2010), “Effects of Internet use on health and depression: A longitudinal study”, The Journal of Medical Internet Research, Vol. 12, No. 1.

Best, S.J. and B.S. Krueger (2005), “Analyzing the representativeness of internet political participation”, Political Behavior, Vol. 27, No. 2, pp. 183-216.

Billari, F.C., O. Giuntella and L. Stella (2018), “Broadband internet, digital temptations, and sleep”, The Journal of Economic Behavior and Organization, Vol. 153, pp. 58-76.

Billari, F.C., O. Giuntella and L. Stella (2017), “Does broadband Internet affect fertility?”, IZA Discussion Paper, No. 10935.

Blachnio, A., A. Przepiorka and M. Benvenuti (2016), “Cultural and personality predictors of Facebook intrusion: A cross-cultural study”, Frontiers in Psychology, Vol. 7.

Black, A.D. et al. (2011), “The impact of eHealth on the quality and safety of health care: a systematic overview”, PLOS Medicine, Vol. 8, No. 1.

Boarini, R. et al. (2012), “What makes for a better life?: The determinants of subjective well-being in OECD countries – Evidence from the Gallup World Poll”, OECD Statistics Working Papers, No. 2012/03, OECD Publishing, Paris, https://doi.org/10.1787/5k9b9ltjm937-en.

Boase, J. et al. (2006), “The strength of Internet ties”, Pew Research Center, www.pewinternet.org/2006/01/25/the-strength-of-internet-ties/ (accessed on 17 December 2018).

Breemersch, K., J. Damijan and J. Konings (2017), “Labour market polarization in advanced countries: Impact of global value chains, technology, import competition from China and labour market institutions”, OECD Social, Employment and Migration Working Papers, No. 197, OECD Publishing, Paris, https://doi.org/10.1787/06804863-en.

Brenke, K. (2014), “Heimarbeit: Immer Weniger Menschen in Deutschland Gehen Ihrem Beruf von zu Hause aus Nach” (Working from home: Less and less people in Germany do their job from their homes), IDIW Wochenbericht, No. 8.

Brod, C. (1984), Technostress: The Human Cost of the Computer Revolution, Addison-Wesley Publishing Company, Reading, MA.

Bruni, L. and L. Stanca (2008), “Watching alone: Relational goods, television and happiness”, The Journal of Economic Behavior and Organisation, Vol. 65, No. 3-4, pp. 506-528.

Brynjolfsson, E. (1993), “The productivity paradox of information technology”, Communications of the ACM, Vol. 36, No. 12.

Brynjolfsson E. and L.M. Hitt (1997), “Information technology and organizational design: Evidence from microdata”, MIT Working Paper.

Brynjolfsson E. and L.M. Hitt (1996), “Paradox lost? Firm-level evidence on the returns to information spending”, Management Science, Vol. 42, No. 4, pp. 541-558.

Brynjolfsson, E. and L.M. Hitt (1995), “Beyond computation: Information technology, organizational transformation and business performance”, The Journal of Economic Perspectives, Vol. 14, No. 4, pp. 23-48.

Brynjolfsson, E. and A. McAfee (2011), Race Against the Machine: How the Digital Revolution is Accelerating Innovation, Driving Productivity, and Irreversibly Transforming Employment and the Economy, Digital Frontier Press, Lexington, MA.

Bulman, G. and R.W. Fairlie (2016), “Technology and education: Computers, software, and the Internet” in Hanushek, E.A., S. Machin and L. Woessmann (eds.), Handbook of the Economics of Education, Vol. 5, Elsevier, North Holland.

Burke, M., C. Marlow and T. Lento (2010), “Social network activity and social well-being”, Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, pp. 1909-12, https://doi.org/10.1145/1753326.1753613.

Büschel, I. et al. (2014), “Protecting human health and security in digital Europe: How to deal with the ‘privacy paradox’?”, Science and Engineering Ethics, Vol. 20, No. 3, pp. 639-658.

Campante, F., R. Durante and F. Sobbrio (2013), Politics 2.0: The Multifaceted Effect of Broadband Internet on Political Participation, National Bureau of Economic Research, Cambridge, MA, http://dx.doi.org/10.3386/w19029.

Carrascal, J.P. et al. (2013), “Your browsing behavior for a Big Mac: Economics of personal information online”, Proceedings of the 22nd International Conference on World Wide Web, pp. 189-200.

Cash, H. et al. (2012), “Internet addiction: A brief summary of research and practice”, Current Psychiatry Reviews, Vol. 8, No. 4, pp. 292-298.

Ceron, A. (2015), “Internet, news, and political trust: The difference between social media and online media outlets”, The Journal of Computer-Mediated Communication, Vol. 20, No. 5, pp. 487-503.

Chan, M. (2015), “Mobile phones and the good life: Examining the relationships among mobile use, social capital and subjective well-being”, New Media & Society, Vol. 17, No. 1.

Cho, J. (2015), “Roles of smartphone app use in improving social capital and reducing social isolation”, Cyberpsychology, Behavior and Social Networking, Vol. 18, No. 6, https://doi.org/10.1089/cyber.2014.0657.

Christensen, H.S. (2011), “Political activities on the Internet: Slacktivism or political participation by other means?”, First Monday, Vol. 16, No. 2.

Cook, J. et al. (2016), “Consensus on consensus: A synthesis of consensus estimates on human-caused global warming”, Environmental Research Letters, Vol. 11, No. 4.

Cooke, C.E. et al. (2010), “Potential value of electronic prescribing in health economic and outcomes research”, Patient Related Outcome Measures, Vol. 1: pp. 163-178, https://doi.org/10.2147/PROM.S13033.

Cotten, S.R. et al. (2014), “Internet use and depression among retired older adults in the United States: A longitudinal analysis”, The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, Vol. 69, No. 5, pp. 763-771.

Crawford, K. and J. Schultz (2013), “Big data and due process: Toward a framework to redress predictive privacy harms”, Boston College Law Review, Vol. 55, No. 1.

Cutler, D. and M. McClellan (2001), “Is technological change in medicine worth it?”, Health Affairs, Vol. 20, No. 5, pp. 11-29.

Deming, D.J. (2017), “The growing importance of social skills in the labor market”, The Quarterly Journal of Economics, Vol. 132, No. 4, pp. 1593-1640.

Dettling, L.J. (2016), “Broadband in the labor market: The impact of residential high-speed Internet on married women’s labor force participation”, ILR Review, Vol. 70, No. 2.

De Vries, A. (2018), “Bitcoin’s growing energy problem”, Joule, Vol. 2, No. 5, pp. 801-805.

Di Gennaro, C. and W. Dutton (2006), “The Internet and the public: Online and offline political participation in the United Kingdom”, Parliamentary Affairs, Vol. 59, No. 2, pp. 299-313.

Diener, E., R. Inglehart and L. Tay (2013), “Theory and validity of life satisfaction scales”, Social Indicators Research, Vol. 112, No. 3.

Dienlin, T., P.K. Masur and S. Trepte (2017), “Reinforcement or displacement? The reciprocity of FtF, IM and SNS communication and their effects on loneliness and life satisfaction”, The Journal of Computer-Mediated Communication, Vol. 22, No. 2.

Ditch The Label (2016), “The annual bullying survey 2016”, www.ditchthelabel.org/wp-content/uploads/2016/04/Annual-Bullying-Survey-2016-Digital.pdf (accessed on 6 November 2018).

Dixon, P. (2017), “A failure to ‘do no harm’ – India’s Aadhaar biometric ID program and its inability to protect privacy in relation to measures in Europe and the U.S.”, Health and Technology, Vol. 7, No. 4, pp. 539-567.

Dolan, P. and R. Metcalfe (2012), “Measuring subjective well-being: Recommendations on measures for use by national governments”, The Journal of Social Policy, Vol. 41, No. 2, pp. 409-427.

Dubois, E. and G. Blank (2018), “The echo chamber is overstated: The moderating effect of political interest and diverse media”, Information, Communication & Society, Vol. 21, No. 5, pp. 729-745.

Dutta-Bergman, M.J. (2006), “Community participation and Internet use after September 11: Complementarity in channel consumption”, The Journal of Computer-Mediated Communication, Vol. 11, No. 2.

Elahi, S. (2009), “Privacy and consent in the digital era”, Information Security Technology Report, Vol. 14, No. 3, pp. 113-118.

Elgar, F.J. et al. (2014), “Cyberbullying victimization and mental health in adolescents and the moderating role of family dinners”, JAMA Pediatrics, Vol. 168, No. 11, pp. 1015-22.

Eppler, M.J. and J. Mengis (2004), “The concept of information overload: A review of literature from organization science, accounting, marketing, mis and related discipline”, The Information Society, Vol. 20, No. 5, pp. 325-344.

Escueta, M. et al. (2017), “Education technology: An evidence-based review”, NBER Working Paper Series, No. 23744.

Eurofound and ILO (2017), “Working anytime, anywhere: The effects on the world of work”, Publications Office of the European Union, Luxembourg, and the International Labour Office, Geneva.

European Commission (2018), “Action plan against disinformation”, https://ec.europa.eu/commission/sites/beta-political/files/eu-communication-disinformation-euco-05122018_en.pdf (accessed on 23 January 2019).

European Commission (2016), “E-Privacy”, Flash Eurobarometer, No. 443.

European Commission (2012), “eHealth action plan 2012-2020 – Innovative healthcare for the 21st century”, https://ec.europa.eu/health/sites/health/files/ehealth/docs/com_2012_736_en.pdf (accessed on 23 January 2019)

Faberman, R.J. and M. Kudlyak (2016), “The intensity of job search and search duration”, Federal Reserve Bank of San Francisco Working Paper Series, No. 13.

Falck, O., A. Heimisch and S. Wiederhold (2016), “Returns to ICT skills”, OECD Education Working Papers, No. 134, OECD Publishing, Paris, http://dx.doi.org/10.1787/5jlzfl2p5rzq-en.

Ferrer-i-Carbonell, A. and P. Frijters (2004), “How important is methodology for the estimates of the determinants of happiness?”, The Economic Journal, Vol. 114, No. 497, https://doi.org/10.1111/j.1468-0297.2004.00235.x.

Fischer, C. (1992), America Calling: A Social History of the Telephone to 1940, University of California Press, Berkeley, CA.

Fraiberger, S.P. and A. Sundararajan (2017), “Peer-to-peer rental markets in the sharing economy” NYU Stern School of Business Research Paper, http://dx.doi.org/10.2139/ssrn.2574337.

Frey, B.S. and A. Stutzer (2005), “Happiness research: state and prospects”, The Review of Social Economy, Vol. 63, No. 2, pp. 207-228, https://doi.org/10.1080/00346760500130366.

Frey, B.S, C. Benesch and A. Stutzer (2007), “Does watching TV make us happy?”, The Journal of Economic Psychology, Vol. 28, pp. 283-313.

Frey, C.B. and M.A. Osborne (2013), “The future of employment: How susceptible are jobs to computerisation”, The Oxford Martin Programme on Technology and Employment Working Paper.

Gensch, C.-O., S. Prakash and I. Hilbert (2017), “Is digitalisation a driver for sustainability?”, in Osburg, T. and C. Lohrmann (eds), Sustainability in a Digital World, Springer Publishing Company, Cham.

Gentile, D. et al. (2011), “Pathological video game use among youths: A two-year longitudinal study”, Pediatrics, Vol. 127, No. 2.

Gil de Zúñiga, H., L. Molyneux and P. Zheng (2014), “Social media, political expression, and political participation: Panel analysis of lagged and concurrent relationships”, The Journal of Communication, Vol. 64, No. 4, pp. 612-634, http://dx.doi.org/10.1111/jcom.12103.

Gluckman, P. and K. Allen (2018), “Understanding wellbeing in the context of rapid digital and associated transformations: Implications for research, policy and measurement”, The International Network for Government Science Advice, Auckland, www.ingsa.org/wp-content/uploads/2018/10/INGSA-Digital-Wellbeing-Sept18.pdf.

Goldfarb, A. and J. Prince (2008), “Internet adoption and usage patterns are different: Implications for the digital divide”, Information Economics and Policy, Vol. 20, No. 1, pp. 2-15.

Goolsbee, A. and J. Guryan (2006), “The impact of Internet subsidies in public schools”, The Review of Economics and Statistics, Vol. 88, No. 2, pp. 336-347.

Graham, C. and M. Nikolova (2012), “Does access to information technology make people happier?”, Global Economy and Development Working Paper, No. 53.

Greenpeace (2017), From Smart to Senseless: The Global Impact of 10 Years of Smartphone, Greenpeace, www.greenpeace.org/usa/wp-content/uploads/2017/03/FINAL-10YearsSmartphones-Report-Design-230217-Digital.pdf (accessed on 6 November 2018).

Grömping, M. (2014), “Echo chambers, partisan Facebook groups during the 2014 Thai election”, Asia Pacific Media Educator, Vol. 24, No. 1, pp. 39-59.

Grundke, R. et al. (2018), “Which skills for the digital era? Returns to skills analysis”, OECD Science, Technology and Industry Working Papers, No. 2018/09, OECD Publishing, Paris. http://dx.doi.org/10.1787/9a9479b5-en.

Gurr, T.R. (1971), Why Men Rebel, Princeton University Press, Princeton, NJ.

Gysels, M. and I.J. Higginson (2007), “Interactive technologies and videotapes for patient education in cancer care: systematic review and meta-analysis of randomised trials”, Support Care Cancer, Vol. 15, No. 1, pp. 7-20.

Hallinan, C., M. Friedewald, and P. McCarthy (2012), “Citizens’ perceptions of data protection and privacy in Europe”, Computer Law and Security Review, Vol. 28, No. 3, pp. 263-272.

Hamburger, Y.A. and E. Ben-Artzi (2000), “The relationship between extraversion and neuroticism and the different uses of the Internet”, Computers in Human Behavior, Vol. 16, No. 4, pp. 441-449, https://doi.org/10.1016/S0747-5632(00)00017-0.

Harari, Y.N. (2018), 21 Lessons for the 21st Century, Penguin Random House.

Hargittai, E. and Y.P. Hsieh (2013), “Digital inequality”, in Dutton, W.H. (ed.), The Oxford Handbook of Internet Studies, Oxford University Press, Oxford.

Hatori, M. et al. (2017), “Global rise of potential health hazards caused by blue light-induced circadian disruption in modern aging societies”, Npj Aging and Mechanisms of Disease, Vol. 3, No. 9.

Heaney, M.T., M.E. Newman and D.E. Sylvester (2011), “Campaigning in the Internet age”, in Craig, S.C. and D.R. Hill (eds.), The Electoral Challenge: Theory Meets Practice, CQ Press, Washington, D.C.

Helliwell, J.F. (2008), “Life satisfaction and quality of development”, NBER Working Papers, No. 14507.

Holwerda, T.J. et al. (2012), “Feelings of loneliness, but not social isolation, predict dementia onset: Results from the Amsterdam Study of the Elderly (AMSTEL)”, The Journal of Neurology, Neurosurgery and Psychiatry, Vol. 85, No. 2, http://dx.doi.org/10.1136/jnnp-2012-302755.

Hong, Y. (2007), “The recent growth of the Internet and changes in household-level demand for entertainment”, Information Economics and Policy, Vol. 19, No. 3-4, pp. 304-318, https://doi.org/10.1016/j.infoecopol.2007.06.004.

Hong, Y., N.C. Pena-Purcell and M.G. Ory (2012), “Outcomes of online support and resources for cancer survivors: A systematic literature review”, Patient Education and Counseling, Vol. 86, No. 3, pp. 288-296.

Hooft Graafland, J. (2018), “New technologies and 21st century children: Recent trends and outcomes”, OECD Education Working Papers, No. 179, OECD Publishing, Paris, https://doi.org/10.1787/e071a505-en.

Howard, P.E.N., L. Rainie and S. Jones (2001), “Days and nights on the Internet: The impact of a diffusing technology”, American Behavioral Scientist, Vol. 45, No. 3, pp. 383-404.

Hwang, K.O., A.J. Ottenbacher and A.P. Green (2010), “Social support in an Internet weight loss community”, The International Journal of Medical Informatics, Vol. 79, No. 1, pp. 5-13.

IEA (2017), Digitalization & Energy, IEA, Paris, https://doi.org/10.1787/9789264286276-en.

IMF (2017), “Promoting inclusive growth”, IMF Annual Report 2017, International Monetary Fund.

IPCC (2013), “Summary for policymakers”, in Stocker, T.F. et al., Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge and New York, NY.

Jahoda, M. (1981), “Work, employment, and unemployment: Values, theories, and approaches in social research”, The American Psychologist, Vol. 36, No. 2, pp. 184-191, http://dx.doi.org/10.1037/0003-066X.36.2.184.

James, C. et al. (2017), “Digital life and youth well-being, social connectedness, empathy, and narcissism”, Pediatrics, Vol. 140, No. S2.

Johansen, M.A. et al. (2012), “Electronic symptom reporting between patient and provider for improved health care service quality: A systematic review of randomized controlled trials. Part 1: State of the art”, The Journal of Medical Internet Research, Vol. 14, No. 5.

Johnston, K. et al. (2011), “Social capital: The benefit of Facebook ‘friends’”, Behaviour and Information Technology, Vol. 32, No. 1, pp. 24-36.

Jorgenson, D.W. and K.J. Stiroh (2000), “Raising the speed limit: U.S. economic growth in the information age”, Brookings Papers on Economic Activity, No. 1, pp. 125-211.

Juvonen, J. and S. Graham (2014), “Bullying in schools: The power of bullies and the plight of victims”, The Annual Review of Psychology, Vol. 65, pp. 159-185.

Kantar Worldpanel (2016), An Incredible Decade for the Smartphone: What’s Next? The Future of Mobile is in Combining Devices, Content, and Services, Kantar Worldpanel, www.kantarworldpanel.com/dwl.php?sn=news_downloads&id=1361 (accessed on 18 December 2018).

Kardefelt-Winther, D. (2017), “Conceptualizing Internet use disorders: Addiction or coping process?”, Psychiatry and Clinical Neurosciences, Vol. 71, No. 7, pp. 459-466.

Kavetsos, G. and P. Koutroumpis (2013), “Technological affluence and subjective well-being”, The Journal of Economic Psychology, Vol. 32, No. 5, pp. 742-753.

Kearns, S.K. (2010), E-learning in aviation, Routledge, London.

Kelliher, C. and D. Anderson (2009), “Doing more with less? Flexible working practices and the intensification of work”, Human Relations, Vol. 63, No. 1, pp. 83-106, https://doi.org/10.1177/0018726709349199.

Kiiski, S. and M. Pohjola (2002), “Cross-country diffusion of the Internet”, Information Economics and Policy, Vol. 14, No. 2, pp. 297-310.

Kokolakis, S. (2017), “Privacy attitudes and privacy behaviour”, Computers & Security, Vol. 64, No. C., pp. 122-134, https://doi.org/10.1016/j.cose.2015.07.002.

Kotikalapudi, R. et al. (2012), “Associating Internet usage with depressive behavior among college students”, IEEE Technology and Society Magazine, Winter 2012.

Krasnova, H. et al. (2013), “Envy on Facebook: A hidden threat to users’ life satisfaction?”, Wirtschaftsinformatik Proceedings 2013, No. 92.

Kraut, R. et al. (1998), “Internet paradox. A social technology that reduces social involvement and psychological well-being?”, American Psychology, Vol. 53, No. 9, pp. 1017-1031.

Krishnamurthy, S. and W. Dou (2008), “Note from special issue editors”, The Journal of Interactive Advertising, Vol. 8, No. 2, pp. 1-4.

Kshetri, N. (2014), “The emerging role of Big Data in key development issues: Opportunities, challenges and concerns”, Big Data & Society, Vol. 1, No. 2, https://doi.org/10.1177/2053951714564227.

Kuhn, P. and M. Skuterud (2004), “Internet job search and unemployment durations”, The American Economic Review, Vol. 94, No. 1, pp. 218-232.

Kuhn, P. and H. Mansour (2014), “Is Internet job search still ineffective?”, The Economic Journal, Vol. 124, No. 581, pp. 1213-33.

Kühn, S. et al. (2011), “The neural basis of video gaming”, Translational Psychiatry, Vol. 1, No. 11.

Kushlev, K. and E.W. Dunn (2015), “Checking email less frequently reduces stress”, Computers in Human Behavior, Vol. 43, pp. 220-228.

Lane, M. and G. Conlon (2016), “The impact of literacy, numeracy and computer skills on earning and employment outcomes”, OECD Education Working Papers, No. 129, OECD Publishing, Paris, http://dx.doi.org/10.1787/5jm2cv4t4gzs-en.

Lazer, D. et al. (2018), “The science of fake news”, Science, Vol. 359, No. 6380, pp. 1094-96.

Lichtenberg, F.R. (2004), “The expanding pharmaceutical arsenal in the war on cancer”, NBER Working Paper, No. 10328.

Lin, J. et al. (2013), “A comparative study of location-sharing privacy preferences in the United States and China”, Personal and Ubiquitous Computing, Vol. 17, No. 4, pp. 697-711.

Lind, E.A. and T.R. Tyler (1988), The Social Psychology of Procedural Justice, Plenum Press, New York, NY.

Lindert, J. (2017), “Cyber-bullying and its impact on mental health”, The European Journal of Public Health, Vol. 27, No. 3.

Lohmann, S. (2015), “Information technologies and subjective well-being: Does the Internet raise material aspirations?”, Oxford Economic Papers, Vol. 67, No. 3, pp. 740-759, https://doi.org/10.1093/oep/gpv032.

Mancini, P. (2015), “Why it is time to redesign our political system”, European View, Vol. 14, No. 1, pp. 69-75.

Mazmanian, M. and I. Erickson (2014), “The product of availability: Understanding the economic underpinnings of constant connectivity”, Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, pp. 763-772.

Mazzolari, F. and G. Ragusa (2013), “Spillovers from high-skill consumption to low-skill labor markets”, Review of Economics and Statistics, Vol. 95, No. 1, pp. 74-86.

McCoy, B. (2013), “Digital distractions in the classroom: Student classroom use of digital devices for non-class related purposes”, The Journal of Media Education, Vol. 4, No. 5.

McKenna, K.Y.A., A.S. Green and M.E.J. Gleason (2002), “Relationship formation on the Internet: What’s the big attraction?”, The Journal of Social Issues, Vol. 58, No. 1.

McKenna, K.Y.A. and J.A. Bargh (1998), “Coming out in the age of the Internet: Identity ‘demarginalization’ through virtual group participation”, The Journal of Personality and Social Psychology, Vol. 75, No. 3, pp. 681-694.

McKinsey, (2014), “Healthcare’s digital future”, McKinsey & Company, www.mckinsey.com/industries/healthcare-systems-and-services/our-insights/healthcares-digital-future (accessed on 6 November 2018).

Michaels, G., A. Natraj and J. van Reenen (2014), “Has ICT polarized skill demand? Evidence from eleven countries over 25 years”, The Review of Economics and Statistics, Vol. 96, No. 1, pp. 60-77.

Mirsky, E.L. and H.A. Omar (2015), “Cyberbullying in adolescents: The prevalence of mental disorders and suicidal behavior”, Pediatrics Faculty Publications, No. 170.

Misra, S et al. (2014), “The Iphone effect: The quality of in-person social interactions in the presence of mobile devices”, Environment and Behavior, Vol. 48, No. 2, pp. 275-298.

Misra, S. and D. Stokols (2012), “Psychological and health outcomes of perceived information overload”, Environment and Behavior, Vol. 44, No. 6.

Mo, P.K.H and N.S. Coulson (2010), “Living with HIV/AIDS and use of online support groups”, The Journal of Health Psychology, Vol. 15, No. 3, pp. 339-350.

Morahan-Martin, J. and P. Schumacher (2003), “Loneliness and social uses of the Internet”, Computers in Human Behavior, Vol. 19, No. 6, pp. 659-671.

Muise, A., E. Christofides and S. Desmarais (2009), “More information than you ever wanted: Does Facebook bring out the green-eyed monster of jealousy?”, CyberPsychology & Behavior, Vol. 12, No. 4.

Nedelkoska, L. and G. Quintini (2018), “Automation, skills use and training”, OECD Social, Economic and Migration Working Papers, No. 202, OECD Publishing, Paris, https://doi.org/10.1787/2e2f4eea-en.

Newman, N. et al. (2018), Reuters Institute Digital News Report, Reuters Institute for the Study of Journalism, http://media.digitalnewsreport.org/wp-content/uploads/2018/06/digital-news-report-2018.pdf?x89475 (accessed on 6 November 2018).

Nie, N.H. and L. Erbring (2002), “Internet and society: A preliminary report”, IT & Society, Vol. 1, No. 1, pp. 275-283.

Nie, N.H. and D.S. Hillygus (2002), “The impact of Internet use on sociability: Time-diary findings”, IT & Society, Vol. 1, No. 1, pp. 1-20.

Nixon, C.L. (2014), “Current perspectives: The impact of cyberbullying on adolescent health”, Adolescent Health, Medicine and Therapeutics, Vol. 5, pp. 143-158.

OECD (2018b), “Consumer policy and the smart home”, OECD Digital Economy Papers, No. 268, OECD Publishing, Paris, https://doi.org/10.1787/e124c34a-en.

OECD (2017a), OECD Digital Economy Outlook 2017, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264276284-en.

OECD (2017b), OECD Employment Outlook 2017, OECD Publishing, Paris, https://doi.org/10.1787/empl_outlook-2017-en.

OECD (2017c), PISA 2015 Results (Volume III): Students’ Well-Being, PISA, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264273856-en.

OECD (2017d), Tackling Wasteful Spending on Health, OECD Publishing, Paris, https://doi.org/10.1787/9789264266414-en.

OECD (2017e), Health at a Glance 2017: OECD Indicators, OECD Publishing, Paris, https://doi.org/10.1787/health_glance-2017-en.

OECD (2017f), “Trust in government”, in Government at a Glance 2017, OECD Publishing, Paris, https://doi.org/10.1787/gov_glance-2017-76-en.

OECD (2017g), “Recommendation of the Council on open government”, OECD Council Recommendation, www.oecd.org/gov/Recommendation-Open-Government-Approved-Council-141217.pdf.

OECD (2017h), “OECD workshop ‘Improving the measurement of digital security incidents and risk management’, Draft summary of main points”, OECD Working Party on Security and Privacy in the Digital Economy, www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=DSTI/CDEP/SPDE(2017)19&docLanguage=En.

OECD (2017i), OECD Guidelines on Measuring the Quality of the Working Environment, OECD Publishing, Paris, https://doi.org/10.1787/9789264278240-en.

OECD (2016a), Innovating Education and Educating for Innovation: The Power of Digital Technologies and Skills, Educational Research and Innovation, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264265097-en.

OECD (2016b), “Skills for a digital world”, OECD, Paris, www.oecd.org/els/emp/Skills-for-a-Digital-World.pdf.

OECD (2016c), “Teachers’ ICT and problem-solving skills: Competencies and needs”, Education Indicators in Focus, No. 40, OECD Publishing, Paris, https://doi.org/10.1787/5jm0q1mvzqmq-en.

OECD (2016d), “Protecting consumers in peer platform markets: Exploring the issues”, OECD Digital Economy Papers, No. 253, OECD Publishing, Paris, https://doi.org/10.1787/5jlwvz39m1zw-en.

OECD (2015a), Skills for Social Progress: The Power of Social and Emotional Skills, OECD Skills Studies, OECD Publishing, Paris, https://doi.org/10.1787/9789264226159-en.

OECD (2015b), Students, Computers and Learning: Making the Connection, PISA, OECD Publishing, Paris, https://doi.org/10.1787/9789264239555-en.

OECD (2015c), OECD Digital Economy Outlook 2015, OECD Publishing, Paris, https://doi.org/10.1787/9789264232440-en.

OECD (2013), ICTs and the Health Sector: Towards Smarter Health and Wellness Models, OECD Publishing, Paris, https://doi.org/10.1787/9789264202863-en.

OECD (2012), “Cybersecurity policy making at a turning point: Analysing a new generation of national cybersecurity strategies for the Internet economy”, OECD Digital Economy Papers, No. 211, OECD Publishing, Paris, https://doi.org/10.1787/5k8zq92vdgtl-en.

OECD (2010), Are the New Millennium Learners Making the Grade? Technology Use and Educational Performance in PISA 2006, Educational Research and Innovation, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264076044-en.

Osburg, T. and C. Lohrmann (2017), Sustainability in a Digital World: New Opportunities through New Technologies, Springer Publishing Company, Cham.

Penard, T., N. Poussing and R. Suire (2013), “Does the Internet make people happier?”, The Journal of Socio-Economics, Vol. 46, pp. 105-116.

Persily, N. (2017), “The 2016 US election: Can Democracy survive the Internet?”, The Journal of Democracy, Vol. 28, No. 2, pp. 63-76.

Pew Research Center (2018a), “About a quarter of U.S. adults say they are ‘almost constantly’ online”, Pew Research Center, www.pewresearch.org/fact-tank/2018/03/14/about-a-quarter-of-americans-report-going-online-almost-constantly/ (accessed on 6 November 2018).

Pew Research Center (2018b), “Activism in the social media age”, Pew Research Center, www.pewinternet.org/2018/07/11/activism-in-the-social-media-age/ (accessed on 6 November 2018).

Pew Research Center (2017), “Public trust in government: 1958-2017”, Pew Research Center, www.people-press.org/2017/12/14/public-trust-in-government-1958-2017/ (accessed on 6 November 2018).

Pittman, M. and B. Reich (2016), “Social media and loneliness: Why an Instagram picture may be worth more than a thousand Twitter words”, Computers in Human Behaviour, Vol. 62, pp. 155-167.

Postill, J. (2018), “Populism and social media: A global perspective”, Media, Culture & Society, Vol. 40, No. 5, pp. 754-765.

Prieger, J.E. and W. Hu (2007), “The broadband digital divide and the nexus of race, competition, and quality”, Information Economics and Policy, Vol. 20, No. 2, pp. 150-167.

Przybylski, A.K. and N. Weinstein (2017), “A large-scale test of the Goldilocks hypothesis: Quantifying the relations between digital-screen use and the mental well-being of adolescents”, Psychological Science, Vol. 28, No. 2, pp. 204-215, https://doi.org/10.1177/0956797616678438.

Przybylski, A.K. et al. (2013), “Motivational, emotional and behavioural correlates of fear of missing out”, Computers in Human Behavior, Vol. 29, No. 4, pp. 1841-1848.

Putnam, R.D. (2000), Bowling Alone: The Collapse and Revival of American Community, Simon & Schuster, New York, NY.

Ragu-Nathan, T.S. et al. (2008), “The consequences of technostress for end users in organizations: Conceptual development and empirical validation”, Information Systems Research, Vol. 19, No. 4, pp. 417-433.

Rainie, L. and B. Wellman (2012), Networked: The New Social Operating System, MIT Press, Cambridge, MA.

Reinke, K. and T. Chamorro-Premuzic (2014), “When email use gets out of control”, Computers in Human Behavior, Vol. 36, Issue C, pp. 502-509.

Rosen, L.D., L.M. Carrier and N.A. Cheever (2013), “Facebook and texting made me do it: Media-induced task-switching while studying”, Computers in Human Behavior, Vol. 29, No. 3, pp. 948-958.

Rosenfeld, M.J. and R.J. Thomas (2012), “Searching for a mate: The rise of the Internet as a social intermediary”, The American Sociological Review, Vol. 77, No. 4, pp. 523-547.

Ross, J. et al. (2016), “Factors that influence the implementation of e-health: A systematic review of systematic reviews”, Implementation Science, Vol. 11, No. 146.

Rotondi, V., L. Stanca and M. Tomasuolo (2017), “Connecting alone: Smartphone use, quality of social interactions and well-being”, The Journal of Economic Psychology, Vol. 63, pp. 17-26.

Rubin, V.L. et al. (2015), “Deception detection of news: Three types of fakes”, ASIST&T Conference, Saint Louis, MO.

Sabatini, F. and F. Sarracino (2017), “Online networks and subjective well-being”, Kyklos, Vol. 70, No. 3, pp. 456-480.

Salvatori, A., S. Menon and W. Zwysen (2018), “The effect of computer use on job quality: Evidence from Europe”, OECD Social, Employment and Migration Working Papers, No. 200, OECD Publishing, Paris, https://doi.org/10.1787/1621d67f-en.

Sametinger, J. et al. (2015), “Security challenges for medical devices”, Communications of the ACM, Vol. 58, No. 4, pp. 74-82.

Schmidt, M. et al. (2015), “The Danish national patient registry: A review of content, data quality, and research potential”, Clinical Epidemiology, Vol. 7, pp. 449-490.

Schwab, K. (2016), The Fourth Industrial Revolution. World Economic Forum, Geneva, www.weforum.org/about/the-fourth-industrial-revolution-by-klaus-schwab.

Senik, C. (2014), “The French unhappiness puzzle: The cultural dimension of happiness”, PSE Working papers, No. 2011-34.

Shklovski, I., R. Kraut and L. Rainie (2006), “The Internet and social participation: Contrasting cross-sectional and longitudinal analyses”, The Journal of Computer-Mediated Communication, Vol. 10, No. 1.

Slev, V.N. et al. (2016), “Effects of eHealth for patients and informal caregivers confronted with cancer: A meta-review”, The International Journal of Medical Informatics, Vol. 87, pp. 54-67.

Slocum, A., R. Hathaway and M. Bernstein (2018), “Striking signs: The Diverse discourse of the 2018 West Virginia teachers’ strike”, English Education, Vol. 50, No. 4, pp. 365-374.

Sousa-Poza, A. and F. Henneberger (2004), “Analyzing job mobility with job turnover intentions: An international comparative study”, The Journal of Economic Issues, Vol. 38, No. 1, pp. 113-137.

Spire, A. (2018), Résistances à l’impôt, attachement à l’État, Seuil, Paris.

Susskind, R. and D. Susskind (2015), The Future of the Professions. How Technology Will Transform the Work of Human Experts, Oxford University Press, Oxford.

Twenge, J.M. et al. (2018), “Increases in depressive symptoms, suicide-related outcomes, and suicide rates among U.S. adolescents after 2010 and links to increased new media screen time”, Clinical Psychological Science, Vol. 6, No. 1, pp. 3-17, https://doi.org/10.1177/2167702617723376.

Ubaldi, B. (2013), “Open government data: Towards empirical analysis of open government data initiatives”, OECD Working Papers on Public Governance, No. 22, OECD Publishing, Paris, https://doi.org/10.1787/5k46bj4f03s7-en.

UK Department for Business Innovation and Skills (2014), “Information and security breaches survey 2014: Technical report”, https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/307296/bis-14-767-information-security-breaches-survey-2014-technical-report-revision1.pdf (accessed on 6 November 2018).

US President’s Council of Advisors on Science and Technology (2012), “Report to President on propelling innovation in drug discovery, development, and evaluation”, www.brookings.edu/wp-content/uploads/2012/11/PCAST-Executive-Summary.pdf (accessed on 23 January 2019)

Valenzuela, S., N. Park and K. Kee (2009), “Is there social capital in a social network site? Facebook use and college students’ life satisfaction, trust, and participation”, The Journal of Computer-Mediated Communication, Vol. 14, No. 4, pp. 875-901.

Valkenburg, P.M. and J. Peter (2007), “Online communication and adolescent well-being: Testing the stimulation versus the displacement hypothesis”, The Journal of Computer-Mediated Communication, Vol. 12, No. 4.

van Dam, H.A. et al. (2005), “Social support in diabetes: A systematic review of controlled intervention studies”, Patient Education and Counseling, Vol. 59, No. 1, pp. 1-12.

van Deursen, A. and J. van Dijk (2014), “The digital divide shifts to differences in usage”, New media & Society, Vol. 16, No. 3, pp. 507-526, https://doi.org/10.1177/1461444813487959.

van Heddeghem, W. et al. (2014), “Trends in worldwide ICT electricity consumption from 2007 to 2012”, Computer Communications, Vol. 50, pp. 64-76.

Vong, S. et al. (2017), “Using information technology to improve surveillance of antimicrobial resistance in South East Asia”, The BMJ, No. 358, Suppl. 1, https://doi.org/10.1136/bmj.j3781.

Wadud, Z., D. MacKenzie and P. Leiby (2016), “Help or hindrance? The travel, energy and carbon impacts of highly automated vehicles”, Transportation Research Part A: Policy and Practice, Vol. 86, pp. 1-18, https://doi.org/10.1016/j.tra.2015.12.001.

Wajcman, J. (2014), Pressed for Time: The Acceleration of Life in Digital Capitalism, University of Chicago Press, Chicago.

Warf, B. (2014), “E-government in the OECD: A comparative geographic analysis”, in Baum, S. and A. Mahizhnan, E-Governance and Social Inclusion: Concepts and Cases, Information Science Reference, Hershey, PA.

Warr, P. (2007), Work, Happiness and Unhappiness, Lawrence Erlbaum Associates Publishers, Mahwah, NJ.

Wheeler, D. (2006), “Empowering publics: Information technology and democratization in the Arab world-lessons from Internet café’s and beyond”, Oxford Internet Institute Research Report, No. 11.

Whitacre, B., S. Strover and R. Gallardo (2015), “How much does broadband infrastructure matter? Decomposing the metro-non-metro adoption gap with the help of the National Broadband Map”, Government Information Quarterly, Vol. 32, No. 3, pp. 261-269.

World Bank (2016), “Digital dividends”, World Development Report 2016, International Bank for Reconstruction and Development/World Bank.

World Economic Forum (2013), Global Risks 2013, Eight Edition, Insights Report – An Initiative of the Risk Response Network, World Economic Forum, Cologny/Geneva, www3.weforum.org/docs/WEF_GlobalRisks_Report_2013.pdf (accessed on 6 November 2018).

Wunnava, P.V. and D. Leiter (2008), “Determinants of inter-country Internet diffusion rates”, IZA Discussion Paper, No. 3666, http://dx.doi.org/10.1111/j.0042-7092.2007.00700.x.

Zhao, P. et al. (2017), “Web-based medical appointment systems: A systematic review”, The Journal of Medical Internet Research, Vol. 19, No. 4.

## Database references

Eurostat Digital Economy and Society (database), http://ec.europa.eu/eurostat/web/digital-economy-andsociety/data/comprehensive-database.

European Quality of Life Survey (database), Eurofound, www.eurofound.europa.eu/surveys/europeanquality-of-life-surveys.

European Survey on Income and Living Conditions (EU SILC), Eurostat, http://ec.europa.eu/eurostat/web/income-and-living-conditions/overview.

European Working Conditions Survey Integrated Data File, 1991-2015 (database), Eurofound, http://doi.org/10.5255/UKDA-SN-7363-7.

Gallup World Poll, www.gallup.com/services/170945/world-poll.aspx.

Health Behaviour in School-Aged Children Study (2014), www.hbsc.org/news/index.aspx?ni=3473.

OECD ICT Access and Usage by Households and Individuals database, http://oe.cd/hhind.

OECD OURdata Index on Open Government Data, www.oecd.org/gov/digital-government/open-government-data.htm.

OECD Programme for International Student Assessment (PISA) database, www.oecd.org/pisa/data/.

OECD Structural Analysis (STAN) Databases, https://stats.oecd.org/Index.aspx?DataSetCode=STANI4_2016.

OECD Survey of Adult Skills (PIAAC) database, www.oecd.org/skills/piaac/publicdataandanalysis/.

OECD Teaching and Learning International Survey (TALIS) database, https://stats.oecd.org/index.aspx?datasetcode=talis_2013%20.

United States School Crime Supplement of the National Crime Victimization Survey (2013), www.icpsr.umich.edu/icpsrweb/NACJD/studies/34980.

Annex 2.A. Internet access and life satisfaction

The European Statistics on Income and Living Conditions (EU-SILC) instrument contains detailed data on living conditions for participating European countries. The 2013 Well-being Module includes a set of self-reported well-being questions alongside a question on Internet access. Similar to other large survey vehicles, the Internet-related question is not very detailed and in this case does not specify whether the individual uses the Internet, let alone the frequency of use. Internet access is therefore considered a proxy for Internet use, where the potential causal pathways of Internet use on subjective well-being run through any of the dimensions discussed in this paper, from changes in social connections to increased transparency of government to access to education, etc. The overall direction of the effect of Internet access is therefore a product of the relative weight of different negative and positive effects of Internet use on life satisfaction.

To estimate the effects of Internet access on life satisfaction, a standard model of the determinants of life satisfaction is used (Frey and Stutzer, 2005; Helliwell, 2008; Dolan, Peasgood and White, 2008), where Internet access is used as the explanatory variable of interest alongside a set of demographic characteristics as well as country-fixed effects to control for country-level variance in terms of living standards as well as potential cultural determinants of life satisfaction responses (Boarini et al., 2012). This approach does not differ substantially from other studies into the relationship between life satisfaction and Internet access or use (Graham and Nikolova, 2012; Lohmann, 2015). Conceptually, the nature of the life satisfaction variable lends itself best to an ordered probit model. However, Ferrer-i-Carbonell and Frijters (2004) show that coefficients estimated with ordinary least squares (OLS) are very similar and the following model builds upon standard practice that utilises OLS for life satisfaction regressions in order to support the ease of interpretation (Boarini et al. 2012). Therefore, the satisfaction of life scale from 0 (Not at all satisfied) to 10 (Completely satisfied) is used as explanatory variable.

+ ${X}_{i}$ +

where the index ic denotes a respondent i in country c, Internet connection is a dummy variable that denotes whether the respondent has access to the Internet at home, ${X}_{i}$ is a set of individual characteristics including age, gender, marital status, employment status, income and education. Finally ${\mu }_{c}$ captures country-fixed effects for the countries included.

The risk of overestimating the effect of Internet use on life satisfaction in this model stems from the possibility that having an Internet connection is strongly correlated with other material life conditions that are not captured by household income, such as individual or household assets that facilitate the capability of getting Internet access. For this reason, ${X}_{i}$ also includes a measure of financial satisfaction in order to further account for individual differences in material well-being. This way, any effect of Internet access is closer to the actual effect of having and using the access, rather than the ability to acquire it.

Results from the regression are shown in Annex Table 2.A.1. The coefficient of Internet access is positive and significant, indicating that the overall effect of being able to access the Internet on life satisfaction is in fact positive. For all other variables, results are in line with formerly found relationships between life satisfaction and demographic characteristics (Dolan, Peasgood and White, 2008). Column (2) shows that the inclusion of financial satisfaction indeed lowers the estimated effect of Internet access, and this second estimate is used to calculate the country-specific effects. The share of life evaluation that is explained by the model is in line with general outcomes of happiness or life evaluation regressions. Senik (2014) notes that the typical share of happiness explained by observable variables in terms of the R2 of an OLS estimate is around 10%. The model that includes financial satisfaction has a substantially higher R2, which is likely partially a result of a shared method variance bias resulting from the similarity in the two questions.

In addition to estimating the effect of Internet access on life satisfaction, interaction variables are included to consider the effect for key demographic groups. These regressions are presented in columns (3-6). Internet access appears to be particularly beneficial for the more vulnerable social groups. The higher people’s income, the less benefit they draw from Internet access and the same counts for young people and the highly educated. These are important findings, because they suggest that Internet access may be inequality-reducing than inequality-inducing in a variety of ways. Conversely, women do benefit less from Internet access than men.

Annex Table 2.A.1. Regression results: Internet access and life satisfaction

(1)

(2)

(3)

(4)

(5)

(6)

Life satisfaction

Life satisfaction

Life satisfaction

Life satisfaction

Life satisfaction

Life satisfaction

Internet access

0.551***

0.277***

1.108***

0.318***

0.247***

0.279***

(0.06)

(0.03)

(0.36)

(0.02)

(0.04)

(0.03)

-0.089**

(0.04)

Internet access*female

-0.074*

(0.04)

Internet access*old

0.063

(0.04)

Internet access*young

-0.049**

(0.02)

Internet access*low education

0.007

(0.02)

Internet access*high education

-0.060*

(0.03)

0.417***

-0.066***

-0.003

-0.067***

-0.066***

-0.066***

(0.03)

(0.02)

(0.03)

(0.02)

(0.02)

(0.02)

Age

-0.102***

-0.049***

-0.048***

-0.049***

-0.054***

-0.049***

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

Age2

0.001***

0.000***

0.000***

0.000***

0.000***

0.000***

(0.00)

(0.00)

(0.00)

(0.00)

(0.00)

(0.00)

Female

0.122***

0.089***

0.088***

0.144***

0.090***

0.089***

(0.03)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

Married

0.452***

0.319***

0.319***

0.319***

0.318***

0.319***

(0.05)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

Separated, divorced or widowed

-0.176***

-0.012

-0.012

-0.015

-0.010

-0.011

(0.04)

(0.03)

(0.03)

(0.03)

(0.03)

(0.03)

Unemployed

-0.560***

-0.113***

-0.111***

-0.110***

-0.114***

-0.113***

(0.08)

(0.04)

(0.04)

(0.04)

(0.04)

(0.04)

Retired

0.407***

0.225***

0.224***

0.231***

0.220***

0.224***

(0.11)

(0.06)

(0.06)

(0.05)

(0.05)

(0.06)

Employed

0.369***

0.226***

0.226***

0.227***

0.232***

0.226***

(0.09)

(0.05)

(0.05)

(0.05)

(0.05)

(0.05)

Low education

-0.095**

-0.029

-0.030

-0.030

-0.029

-0.032

(0.04)

(0.02)

(0.02)

(0.02)

(0.02)

(0.02)

High education

0.146***

0.008

0.014

0.008

0.009

0.063*

(0.02)

(0.01)

(0.01)

(0.01)

(0.01)

(0.04)

Financial satisfaction

0.481***

0.481***

0.481***

0.480***

0.481***

(0.01)

(0.01)

(0.01)

(0.01)

(0.01)

Country fixed effects

(yes)

(yes)

(yes)

(yes)

(yes)

(yes)

N

242 530

242 530

242 530

242 530

242 530

242 530

R2

0.159

0.400

0.400

0.400

0.400

0.400

Note: Standard errors in parentheses. *=p<0.10, **=p<0.05, ***=p<0.01. Results obtained using an ordinary least squares regression with robust standard errors clustered by country and survey weights (but not population weights).

Source: OECD calculations based on EU-SILC (2013), http://ec.europa.eu/eurostat/web/income-and-living-conditions/overview.

Finally, the coefficient on Internet access is used to estimate the life satisfaction benefit of Internet access in each country based on the number of people with Internet access as reported in EU-SILC, using the following calculation:

This estimation of added life satisfaction due to Internet access by country represents the gains in life satisfaction associated with people having access to the Internet (Annex Table 2.A.2). This estimate is somewhat artificial as it compares the life satisfaction benefit of having Internet access between two different groups, rather than between the same group before and after having access. However, it does point to the potential net positive effects that may result from having Internet access and to the importance of policies that bridge the digital divide and ensure that everyone has the possibility to access the Internet.

Annex Table 2.A.2. Life satisfaction gains associated with Internet access for selected countries

Country

Self-reported Internet access

Life satisfaction gains associated

with Internet access

Austria

0.82

0.23

Belgium

0.83

0.23

Estonia

0.77

0.21

Finland

0.82

0.23

France

0.77

0.21

Greece

0.58

0.16

Hungary

0.61

0.17

Ireland

0.78

0.22

Iceland

0.92

0.25

Italy

0.52

0.15

Luxembourg

0.84

0.23

Latvia

0.66

0.18

Netherlands

0.90

0.25

Norway

0.90

0.25

Poland

0.73

0.20

Portugal

0.59

0.16

Spain

0.70

0.19

Sweden

0.88

0.24

Switzerland

0.89

0.25

United Kingdom

0.82

0.23

Note: Calculations are made using survey weights.

Source: OECD calculations based on EU-SILC (2013), http://ec.europa.eu/eurostat/web/income-and-living-conditions/overview.

## Notes

← 1. Data is from 2017 or latest available year. Average does not include Australia, New Zealand and the United States. The OECD average is population weighted. Source: OECD Information and Communication Technology database, 2017.

← 2. These figures are for 2017. Source: OECD Information and Communication Technology database.

← 3. Scores for the problem-solving proficiency in technology-rich environments task are classified in four levels: Below Level 1 through Level 3. In addition to these four proficiency levels, there are three additional categories (no computer experience, failed ICT core and opted out) for those adults who were unable to demonstrate their proficiency in this area due to a lack of basic computer skills needed to complete the assessment.

← 4. The consumer surplus refers to the difference between the price that consumers are willing to pay for a specific product and the actual price they pay for the product.

← 5. The results for task discretion are in line with those of Salvatori, Menon and Zwysen (2018).

← 6. The extended job strain index (OECD, 2017i) is a composite measures of the quality of the working environment that considers a larger number of job resources and job demands (6) compared to the index included in other OECD reports (3).

← 7. Importantly, this means that the indicator of job stress associated with computer-based jobs does not reflect any potential cross-country variation in the extent to which computer-based jobs increase job stress. It is conceivable that in some countries, computer-based jobs have a higher impact on job stress than others due to workplace policies and cultural factors. These differences are not taken into consideration in this indicator.

← 8. Similar to the indicator on job stress, this means that the indicator of worries about work when not working associated with computer-based jobs does not reflect any potential cross-country variation in the extent to which computer-based jobs increase worries outside work hours. It is conceivable that in some countries, computer-based jobs have a higher impact on worries about work when not working than others due to workplace policies and cultural factors. These differences are not taken into consideration in this indicator.

← 9. In the case of the television, a large body of research also supports the displacement hypothesis that television crowds out social interactions (Kraut et al., 1998; Putnam, 2000; Frey, Benesch and Stutzer, 2007; Bruni and Stanca, 2008). This effect is particularly strong for individuals with low levels of self-control over their own behaviour and for people who watch excessive amounts of television (Frey, Benesch and Stutzer, 2007). Not only does the television depress the frequency of social interactions, but it also has significantly negative effect on life satisfaction (Bruni and Stanca, 2008).

← 10. For example, McKenna and Bargh (1998) found that finding support online increases the possibility of coming out in real life for homosexuals.

← 11. The KidsOnline survey is another survey that focuses on children’s online behaviour; it includes a confidential section that is filled out at the parental home, which may be a safer space to self-report bullying. However, this survey is currently only available for EU countries and therefore has limited comparability.

← 12. Change.org, www.change.org (accessed on 31 January 2019).

← 13. Avaaz: the world in action, https://avaaz.org/page/en/ (accessed on 31 January 2019).

← 14. DemocracyOS: change the tool, Democracia en Red, http://democracyos.org/ (accessed on 31 January 2019).

← 15. Disinformation is defined as all forms of false, inaccurate, or misleading information designed, presented and promoted to intentionally cause public harm or for profit (European Commission, 2018).

← 16. This figure therefore does not show cross-country variation in the strength of the association between life satisfaction and Internet use. The reported variation across countries only reflects differences in Internet access. Since Internet is consistently found to be associated with higher levels of life satisfaction, this figure reflects an illustration of the life satisfaction gains associated with Internet access.

← 17. Several authors have suggested that Internet access may provide the greatest benefits in life satisfaction at the lower end of the income distribution (Graham and Nikolova, 2012; Penard, Poussing and Suire, 2013). An interaction term between the logarithm of household income and the access to Internet variable in the model confirms that the relationship is stronger at the lower end of the income distribution and diminishes with increasing income. This finding has different interpretations. Thanks to the Internet, people with lower incomes may benefit from services that were previously inaccessible to them, which is not the case to the same extent for people in higher incomes, who had access to such services even without Internet. However, it is also possible, as mentioned above, that Internet access does reflect an uncaptured income or wealth effect.

End of the section – Back to iLibrary publication page