5 Thesis Nicole Goedhart.pdf

Social inclusion in digitizing societies Starting from the lifeworld of people with a low socioeconomic position

Nicole Sylvia Goedhart

Members of the thesis committee: · prof.dr. M.J. Broersma · prof.dr. P. Lago · dr. H.C. Ossebaard · dr. B.J. Regeer · prof.dr. S.M.E. Wyatt

ISBN/EAN: 978-94-6423-206-6 Cover Design: Zuzanna Dembiec | zuzannadembiec.com Lay-out: Marian Sloot | www.proefschriftmaken.nl Printing: ProefschriftMaken | www.proefschriftmaken.nl © Nicole Goedhart, 2021. All rights reserved. No part of this work may be reproduced by print, copy, or otherwise without permission of the author.

VRIJE UNIVERSITEIT

Social inclusion in digitizing societies Starting from the lifeworld of people with a low socioeconomic position

ACADEMISCH PROEFSCHRIFT

ter verkrijging van de graad Doctor of Philosophy aan de Vrije Universiteit Amsterdam, op gezag van de rector magnificus prof.dr. V. Subramaniam, in het openbaar te verdedigen ten overstaan van de promotiecommissie van de Faculteit der Bètawetenschappen op donderdag 3 juni 2021 om 11.45 uur in de aula van de universiteit, De Boelelaan 1105

door

Nicole Sylvia Goedhart geboren te Nieuwerkerk aan den IJssel

promotor: copromotoren:

prof.dr. J.E.W. Broerse dr. C. Dedding dr. T. Zuiderent-Jerak

Content Account

7

Introduction: a situated understanding of digital inequalities

9

Theory and concepts: the digital divide, socioeconomic inequalities, and public engagement

15

Research design: action research for change

33

‘Just having a computer doesn’t make sense’: the digital divide from the perspective of mothers with a low socioeconomic position

51

Exploring the boundaries of ‘good’ participatory action research in times of increasing popularity: dealing with constraints in local policy for digital inclusion 73 Fun, engaging, and easily shareable? Exploring the value of co-creating vlogs with citizens from disadvantaged neigbourhoods

93

Meaningful engagement in research of citizens living in vulnerable circumstances: a narrative review using a systematic search

115

Persistent inequitable design and implementation of patient portals for users at the margins

143

Discussion and conclusion: a new geography of responsibility for digital inclusion

161

References

183

Appendices

209

Samenvatting / Summary

233

List of publications

243

Dankwoord / Acknowledgements

247

|7

Account Chapter 4 to 8 are based on co-authored articles that have previously been published, are under review, or have been submitted for publication in peer reviewed articles.

Chapter 4 Goedhart, N. S., Broerse, J. E. W., Kattouw, R., & Dedding, C. (2019). ‘Just having a computer doesn’t make sense’: the digital divide from the perspective of mothers with a low socioeconomic position. New Media & Society, 21(11-12), 2347-2365.

Chapter 5 Dedding, C., Goedhart, N. S., Broerse, J. E. W., & Abma, T. A. (2021). Exploring the boundaries of ‘good’ participatory action research in times of increasing popularity: dealing with constraints in local policy for digital inclusion. Educational Action Research, 29(1), 20-36.

Chapter 6 Goedhart, N. S.*, Lems, E.*, Zuiderent-Jerak, T., Pittens, C. A. M. C., Broerse, J. E. W., & Dedding, C. (accepted). Fun, engaging, and easily shareable? Exploring the value of co-creating vlogs with citizens from disadvantaged neighbourhoods. Action Research.

Chapter 7 Goedhart, N. S., Pittens, C. A. C. M., Tončinić S., Zuiderent-Jerak, T., Dedding, C., & Broerse, J. E. W. (submitted). Meaningful engagement in research of citizens living in vulnerable circumstances: a narrative review using a systematic search. Research Involvement and Engagement.

Chapter 8 Goedhart, N.S., Zuiderent-Jerak, T., Woudstra, J., Betten, A.F., Broerse, J.E.W., & Dedding, C. (2021) Persistent inequitable design and implementation of patient portals for users at the margins. Journal of the American Medical Informatics Association, 28(2), 276-283.

*shared first authorship

Introduction: a situated understanding of digital inequalities

10 | Chapter 1

“My Waternet”, “my energy company”, “my bank”, “my budget”, “my everything and then also the tax authorities!” – Gregory

Today’s society forces Gregory, a father living in Amsterdam, to do everything online: manage his bank account, provide his annual (electric and water) meter readings, contact his health-care insurance provider and debt counsellor, pay his taxes, and so on. It frustrates Gregory that the significant entanglement of information and communication technologies (ICTs) within daily practices and services makes it impracticable not to use them. At times he feels as if he has lost control in an unknown world. It annoys him that he needs to go to a community centre to print out and scan documents because he does not have a printer at home. He also mentions that many websites are too complicated for him and that it feels as if these websites are trying to mislead him. He adds that his neighbour often gets into trouble since she does not have a computer or internet and, therefore, misses essential information for her debt restructuring. The above story is just one of the examples highlighting the fact that without material access to ICTs and/or the necessary skills, citizens are severely constrained in today’s society. Gregory is not alone. In 2020 in the European Union, 9.5% of people have never used the internet, ranging from approximately 3% in Sweden, Denmark, and the Netherlands to 22% in Portugal and Greece and 24% in Bulgaria (Digital Economy and Society Index (DESI), 2020b).

The number of citizens with no ICT skills or a low level of such skills1 is also still substantial. In 2017 in Europe, this applies to 43% of people, ranging from 71% in Bulgaria and Romania to 15% in Luxembourg, 21% in the Netherlands, and 29% in the United Kingdom (Digital Economy and Society Index (DESI), 2019). In other words, in the Europe, more than 2 out of 5 citizens has no ICT skills or a low level of such skills. To refer to the gap between those fully engaged with ICTs and those who are not, terms such as ‘digital immigrants’, ‘digital divide’, and ‘digital gap’ were introduced during the second half of the 1990s. These terms suggest a simple binary division. In practice, this is much more nuanced. There is no singular undifferentiated non-user on one side and a singular undifferentiated user on the other side (Gunkel, 2003; Halford & Savage, 2010; Wyatt, 2003). For example, the so-called ‘rejecters’ or ‘resisters’ purposefully choose not to make use of ICTs, while for the ‘expelled’ or ‘excluded’, not using ICTs is not a deliberate choice (Wyatt, 2003, 2014). There is also differentiation within these groups. 1 If an individual has a low level of ICT skills, this means that they are only able to carry out activities in one of the four digital competence dimensions (information, communication, content creation, and problem solving).

More details at: https://ec.europa.eu/jrc/digcomp

Introduction | 11

For some, the internet is an unknown, somewhat vague, and complicated world which limits ICT use; others do not have the financial resources to get access to ICTs. Therefore, digital inequality is better depicted as a complex problem rather than a binary division (cf. Gunkel, 2003; Halford & Savage, 2010). Digital inequality is interwoven into everyday life and there is not a single easy solution. Solely providing citizens with initial access to ICT devices will not solve the problem of digital inequality and often reveals other problems, such as not having the time or energy to attend a course to learn ICT skills or having low literacy levels which prohibit them from using the internet. Socio-demographics such as gender, ethnicity, and poverty correlates with how individuals approach and use ICTs (e.g. Henwood & Wyatt, 2009; Laz & Berenson, 2013; Morey, 2007; Van Deursen & Van Dijk, 2015). There is often a widely held assumption that digital inequality is only a problem among elderly people. Research, however, shows that someone’s income level or ethnic background, for example, also influence online health-related information-seeking behaviour (Laz & Berenson, 2013; Morey, 2007). Moreover, the socioeconomic environment can have an effect on how young people develop digital skills, gain experiences online, and can benefit from the online world (e.g.

Livingstone & Helsper, 2007). Research into digital inequality traditionally focuses on the conceptualization of instruments for measuring key concepts of the digital divide debate, such as ‘digital skills’ or ‘ICT or internet usage’ (e.g. Schreeder et al., 2017; Van Deursen et al., 2016; Van Laar et al., 2017). For example, Van Deursen et al. (2016) developed a theoretically informed and empirically tested instrument for measuring different forms of internet skills. These types of instruments are used to quantitatively demonstrate the interrelatedness of social and digital inequality (e.g. Helsper et al., 2015; Helsper & Reisdorf, 2017). They show, for instance, that citizens who already have fewer resources, such as money or educational attainment, use ICTs to a lesser extent and for less diverse purposes. This means that those who lack these resources often benefit less from the opportunities that ICTs can offer (e.g. Helsper et al., 2015; Van Deursen & Helsper, 2015). In other words, socioeconomic and digital inequalities are not only interrelated but can also reinforce each other. While these primarily quantitative studies are helpful in mapping the digital divide landscape and show that socioeconomic and digital inequalities interact, these studies are less useful for understanding how these factors interact in the daily lives of people.

The numbers do not show what it means to them, and what is needed to diminish the digital inequalities. To understand digital inequalities, a ‘situated’ is needed (cf. Henwood & Wyatt, 2009). A situated understanding refers to the understanding of the complex dynamics underlying digital inequalities and the relation that these inequalities have with long-standing inequalities, such as those related to gender, people with an ethnic minority

1

12 | Chapter 1

background, or poverty. To the time of writing, only limited information is known about the stories underlying the numbers (cf. Gilbert, 2010). To create a situated understanding of digital inequalities, it is necessary to study digital inequalities in the broader social context and with all citizens. Not to mention citizens with a low socioeconomic position (SEP). Experiential knowledge of citizens who have to deal with social and digital inequalities is crucial to define and understand the problem and to search for solutions. We can then move beyond simply reporting barriers to ICT use, such as affordability (Alam & Imran, 2015; Powell, 2010), language (Alam & Imran, 2015; Guttman et al., 2018), or ICT illiteracy (Alam & Imran, 2015). Citizens have a crucial role to play in defining the problem of digital inequality, which may differ from the one set out by policymakers (Zuiderent-Jerak, 2015). Also, the idea is that co-creating solutions with citizens will result in policies and practices that are more in line with people’s needs and daily reality (Chalmers et al., 2014; Ioannidis et al., 2014). The importance of engaging citizens (or ‘the public’) in science, technology, and policymaking is increasingly acknowledged. The upsurge in public engagement practices is driven by three interrelated imperatives: (1) normative, (2) instrumental, and (3) substantive (e.g.

Broerse & Buning de Cock, 2012; Stirling, 2008). The normative imperative – also sometimes called the democratic imperative – states that members of the lay public have the right to be involved in matters that affect them. This is based on the assumption that involving the public, especially groups that are often ignored, is in itself a good thing. As a result of this involvement, power relations might shift (Anderson, 2017; Broerse & Buning de Cock, 2012; Stirling, 2008). The instrumental imperative is primarily focused on the outcomes of public engagement. The assumption is that if you involve the public, this can result in a higher level of acceptance of the outcomes by the public. In other words, public engagement is a tool for building legitimacy (Broerse & Buning de Cock, 2012). This argument fits well with the increasing concern of funders and governments that research should have a social and economic impact and that research waste should be reduced (Gillard et al., 2012; Ioannidis et al., 2014). Lastly, the substantive imperative reveals that knowledge gained through experiences in a specific context is unique, highly relevant, and complementary to professional knowledge. It is increasingly acknowledged that research without the involvement of the lay public does not always do justice to citizens’ needs or their daily reality.

Bringing in the experiences of the public can result in research that is more contextualized and needs oriented (Abma et al., 2019; Broerse & Buning de Cock, 2012; Chalmers et al., 2014; Wynne, 1996). Despite the recent upsurge in public engagement in research and policymaking, the involvement of citizens living in vulnerable circumstances, such as citizens who live in poverty or have a low educational or literacy level, is often minimal. Mainstream

Introduction | 13

research methods, such as questionnaires, are designed to fit best with professionals and well-educated citizens than citizens with a low SEP (Redwood et al., 2012). In addition, existing power relations influence the opportunities citizens with a low SEP have to be engaged. Power inequalities in collaborations are sometimes masked if more knowledgeable actors smoothly outvote the less knowledgeable, or less literate, actors (Broerse & Buning de Cock, 2012; Stirling, 2008). This can result in tokenistic collaborations which could cause harm, create mistrust, or result in participation fatigue (Banks et al., 2013; Dedding & Slager, 2013) To support citizens living in more vulnerable circumstances, there is a particular need for good practices or information about strategies that can guide researchers to involve citizens who are often hard to reach (Decamp et al., 2015; O’Reilly-De Brún et al., 2016) To endorse social inclusion in the digitizing society in this thesis, I take up the challenge of meaningfully engaging citizens with a low SEP, policymakers, community-based organizations, technology developers, and others in the debate about digital inequalities. Involving a diverse group of citizens makes it possible not only to explore the multilayered, complex problem of digital inequalities but also to support collective action.

Importantly, the perspectives and needs of citizens with a low SEP will be central throughout the thesis. As previously described, doing research with citizens with a low SEP is experienced as difficult and there is a lack of information about strategies that can guide researchers. Therefore, in this thesis I also critically reflect on the inclusion and exclusion mechanisms concerning the engagement of citizens with a low SEP in the research process itself. To summarize, the aim of this thesis is to facilitate the development of policies and practices that address digital inequalities so that they are more sensitive to the complex daily realities of citizens with a low SEP by (1) gaining insight into the complex daily realities of citizens with a low SEP and understanding their needs in the digitizing society, and (2) facilitating and reflecting on the meaningful engagement of citizens, researchers, technology developers, and policymakers in digital inclusion.

1

Theory and concepts: the digital divide, socioeconomic inequalities, and public engagement

16 | Chapter 2

In the introductory chapter I aimed to show the complexity of the digitizing society for those with limited access to information and communication technologies (ICTs) and for those who do not have the required skills. In this chapter I will explore the digital divide concept in more detail and provide an insight into the interrelatedness of digital and social inequalities in the context of digitizing post-welfare societies. In addition, I zoom in on the conceptualization of public engagement and the complexities of involving citizens who are living in vulnerable circumstances in research and policymaking.

2.1 | The concept of the digital divide and well-known critiques The term ‘digital divide’ has been used since the second half of the 1990s as a result of the fast diffusion of digital technologies and the often-unquestioned favourable presumptions about ICTs. The origin of the term digital divide is uncertain, but in publications it is most often credited to the authors of a series of reports published by the United States Department of Commerce’s National Telecommunication and Information Administration (NTIA) (Gunkel, 2003; Van Dijk, 2005). The popular press, scholarly studies, and politics quickly took up the term as a means to communicate about the gap between those who are engaged with digital technologies and those who are not. The use of the term appeared to be successful in putting the issue high on scholarly and political agendas, especially between 2000 and 2004. During this period, hundreds of (scientific) conferences and policy-related gatherings were dedicated to this topic, because unequal physical access to technology seemed to be the last hurdle that was obstructing the digital revolution and the internet hype (Van Dijk, 2005, 2006). Meanwhile, scholars started to publish work about the ambiguity and potential pitfalls of the digital divide concept. Van Dijk (2002, 2006) and colleagues (i.e.

Gunkel, 2003; Halford & Savage, 2010; Hargittai, 2003) warned against several potential misunderstandings. These scholars particularly emphasized that the digital divide metaphor seems to suggest and is often visualized as a sharp and simple dichotomy (see, for example, Figure 2.1). In practice, there are, however, complex social, economic, and cultural differentiations in ICT use. Therefore, these scholars argued that a reference to a gradation of ICT use would be more appropriate. In response to these critiques, researchers started to use the concept ‘digital inequality’ rather than ‘digital divide’ (Dimaggio & Hargittai, 2001; Stiakakis et al., 2009). This concept creates less of an impression that there is a binary division between the ‘haves’ and ‘have-nots’ which is primarily caused by lack of access to technologies (Stiakakis et al., 2009). These scholars attempted to emphasize that elimination of the digital divide

Theory and concepts | 17

2

Figure 2.1: Metaphorical visualization of the digital divide drawn by sketch artist Bas Kohler (www.baskohler. nl) during the Digi-Challenge, a conference about the digital divide in Amsterdam (March 2018) http:// www.pact-amsterdam.nl/wp-content/uploads/2018/03/Digi-Challenge-6-3-2018-A3.pdf

through increasing access to ICTs might actually produce further digital inequalities. In this respect, digital inequalities refer to inequalities concerning elements such as skills, social support, or being able to use ICTs autonomously’ (Dimaggio & Hargittai, 2001; Stiakakis et al., 2009). At the same time, scholars started to publish about the changing conceptualization of the digital divide concept. Gunkel (2003) emphasized that the concept is linked to continuously changing technologies, which makes the concept equivocal, plural, and flexible. Over the years, ICTs have changed in appearance and function and will be continually developing technologies. The definition of the digital divide is similarly transient. In the next section, I will explain in more detail how the conceptualization of the digital divide concept has changed over time and thereby has become more similar to the meaning of digital inequality as defined by, for example, Dimaggio and Hargittai (2001). Throughout this thesis, I use the terms ‘digital divide’ and ‘digital inequality’ interchangeably. While I acknowledge the potential pitfalls and possible misunderstanding of the digital divide concept, I purposefully choose to still use this concept from time to time throughout this action-oriented thesis.

The main reason for this choice is that the digital divide concept has appeared to thrive and is still thriving regarding getting the issue on scholarly and political agendas (see Figure 2.1) and therefore can be used as a lever to create action that does not take the notion at face value.

18 | Chapter 2

2.2 | First-, second-, and third-level digital divide Digital divide research is grounded in multiple scientific fields, such as sociology, psychology, communication science, and education science, and is often studied at the intersection of these fields. In sociology, the digital divide is seen as a form of stratification and, therefore, is often linked to the concept of social inequalities. Conceptual and analytical tools used by well-known sociologists such as Marx, Weber, Simmel, and Bourdieu have been applied to study the digital divide (Ragnedda & Muschert, 2013; Van Dijk, 2005; Witte & Mannon, 2010). Psychological studies focus more on attitudes towards the internet and motivation for using it (e.g. Chua et al., 1999; Durndell & Haag, 2002). Other scholars are more concerned with why a certain technology is used or not used. For example, the theory of the diffusion of innovation of Rogers (1995) or the technology acceptance model (Davis, 1989), originally developed within, respectively, communication science and information system research, are often adopted. Lastly, education scientists focus on information and digital literacy. These scholars often focuses on the integration of ICTs in the educational curriculum (Chen & Price, 2006; Lawless & Pellegrino, 2007). Within these different scientific fields, the conceptualization of the digital divide changed over time.

The literature distinguishes between three levels of the digital divide. To explain the difference between these three levels, I use the causal and sequential model of digital technology access by individuals in contemporary societies (Figure 2.2) of Van Dijk (2005). This model is based on Van Dijk’s resources and appropriation theory that was inspired by the structuration theory of Giddens, the capital theory of Bourdieu, the relational perspective of Tilly, and several theories about technological acceptance. The nature of the model is broad and is therefore used by many scientists from different disciplines. Research into the first-level (blue areas in Figure 2.2) digital divide (1995–2005) is dominated by studies focusing on computer ownership, internet access, and broadband adoption rates (e.g. Hargittai, 1999; Ono & Zavodny, 2003). Literature reporting on the first-level digital divide conceptualizes the digital divide as a binary division. During this period, computers were considered to be just another emerging technology, like televisions or radios, that was ready to adopt. Differences in adoption rates were most often linked to the natural diffusion process of the innovation (Rogers, 1995) or technological acceptance (Davis, 1989). Moreover, studies were concentrated on why certain groups were motivated to use ICTs and others were not.

Terms such as technophobia and computer anxiety were introduced (Brosnan, 1998; Chua et al., 1999; Durndell & Haag, 2002) to refer to psychological factors such as fear or anxiety about using a computer or even just thinking about the possibility of using a computer.

Theory and concepts | 19

Positional categories - Labour - Education - Household - Nation

Positional categories - Age/generation - Sex/gender - Race/ethnicity - Intelligence - Personality - Health/ability

Resources - Temporal - Material - Mental - Social - Cultural

Participation in Society - Economy - Social networks - Space/geography - Culture - Politics - Institutions Access

Usage access

Material access

Skills access - Strategic - Informational - Instrumental - Digital skills

sequential part of the model primary causal or sequential relation secondary causal or sequential relation

Motivational access

Next innovation

Technologica properties of ICT (hardware, software, content)

Figure 2.2: The causal and sequential model of digital technology access by individuals in contemporary societies of Van Dijk (2005) with colours added by the author (blue = first-level digital divide, orange = second-level digital divide, green = third-level digital divide)

Since 2002, an increasing number of scholars have studied the digital divide ‘beyond access’ (Orange areas in Figure 2.2) (Van Dijk, 2006). The second-level digital divide, as introduced by Hargittai (2002), emphasis that having material access is not the only factor that closes the digital divide. It is important to understand the differences between people even after they get initial physical access. Van Dijk (2005, pp. 20–21) conceptualized the general term ‘access’ into four successive kinds of access: (1) motivational access, (2) material access, (3) skills access, and (4) usage access. These four types of access are supposed to be cumulative and recursive if a new technology is introduced (Van Dijk, 2005). This implies that before individuals can use a technology, they need to have the motivation to use it, then buy it, and develop new skills. This process recurs as new ICTs are introduced or upgraded. Research into the second-level digital divide focuses on the conceptualization of these four successive kinds of access and how they can be measured. Particular attention is paid to the conceptualization of digital skills (cf. Van Laar et al., 2017). For example, van Deursen and van Dijk (2009) make a distinction between (1) medium-related skills, including software skills (i.e. making spreadsheets, using a browser, email, and Word)

2

20 | Chapter 2

and operational skills (i.e. instrumental skills), and (2) content-related skills, such as information-related skills (e.g. understanding online information about health), creative skills, or social skills (i.e. social media). Over the years, the digital skills concept was often substituted by the concept of internet skills, because a person needs more skills to use the internet than for using a computer alone. To measure internet skills, Van Deursen et al. (2016) developed the Internet Skills Scale (ISS), a validated instrument consisting of five types of internet skills: operational, information navigation, social, creative, and mobile. Moreover, there was an increased scholarly focus on online risks, security, and ethics (cf. Schreeder et al., 2017), such as fraudulent online scams and identity theft. In addition, research into the second-level digital divide paid attention to the relation between social and digital inequalities. Inequalities in personal and positional resources can cause unequal access to, and use of, ICTs (Van Dijk, 2005). In other words, the lack of access is seen as a consequence of existing social inequalities. Quantitative research shows, for example, that someone’s ICT skills and usage diversity, e.g.

the different ways in which they use technology, are significantly (directly or indirectly) associated with income, education, literacy level, and gender (Morey, 2007; Van Deursen & Van Dijk, 2015). Attention has also been paid to the fact that children who are growing up in families with a low level of resources often lack opportunities to get material access to ICTs or develop ICT skills. This reduces their chances of flourishing in a digitized society and coping with the risks of the internet, such as cyberbullying, exposure to violence or pornography, sexting, or making contact with strangers (Livingstone & Helsper, 2007; Paus-Hasebrink et al., 2014; Rodríguez-de-Dios et al., 2018). Research on the third-level digital divide (green area, Figure 2.2) focuses not only on the causes but also on the consequences of it (Helsper, 2012; Helsper & Van Deursen, 2017; Schreeder et al., 2017). The introduction of ICTs produces so-called ‘winners’ and ‘losers’ (Bruno et al., 2010). Research on the third-level digital divide reports in detail on the interrelationship between social and digital inequalities. These studies highlight that social exclusion leads to digital exclusion, but also show how digital engagement might, or might not, change social exclusion (Helsper, 2012; Helsper et al., 2015).

Studies reporting on the third-level digital divide often base their work on Bourdieu’s (1986) understanding of capital (Carlson & Isaacs, 2018; Gilbert, 2010; Helsper, 2012). Bourdieu emphasizes that someone’s position or possibilities in a certain society not only depend on economic capital but also on cultural, social, and personal categories (Bourdieu, 1986). These different forms can be seen as different spheres of inequality and outcomes of the social (technical) space we live in. Individuals often experience the different forms of capital together due to underlying power structures that concentrate (dis)advantages in certain groups.

Theory and concepts | 21

Bourdieu’s understanding of capital is, for example, used in the corresponding field model for digital exclusion of Helsper (2012) to unpack the offline field of inequalities – in the form of economic, cultural, social, and personal resources – and describes how these are linked to and interact with the corresponding digital fields. For example, the offline economic field (e.g. someone’s income, employment status, or educational level) often interacts with online economic achievements (e.g. whether someone was able to buy (cheaper) products online or their ability to participate in online learning) (Helsper, 2012; Van Deursen & Helsper, 2015). Also, Robinson (2009) uses Bourdieu’s ideas to show that disparities in outcomes of internet use are not only intermediated by access to ICTs but also by their social context and field position. In her ethnographic work in California, she showed that children from low- and middle-income families internalized ICT as a ‘taste for the necessary’. In contrast, the children of high-income parents experienced ICT as ‘serious play’. The latter understanding gave children more opportunities to develop skills and they benefited more from ICTs than their peers from low- and middle-income families. To summarize, the literature on the digital divide identifies three levels.

The second level arose from the criticism that access is more than just having a device, and the third level shows that not only the causes but also the consequences of ICTs must be considered. Quantitative studies reporting on the digital divide clearly show that social and digital inequalities are linked and can even reinforce each other. This shows that the digital divide is an entrenched social problem that transcends the material access problem. While the development of these three different levels is often described as a historical development, it is not the case that only the third level is still important. All three levels still dominate to varying degrees. The evolving levels made the conceptualization of the digital divide more complete but also more complex. It is interesting that although technology is seen as evolving, the digital divide literature is often rather deterministic (Gunkel, 2003; Mariën et al., 2016; Wyatt et al., 2005). It is much more focused on the significant social ramifications of technology than on the technology itself (Gunkel, 2003). For example, the corresponding field model for digital exclusion of Helsper (2012) is based on the normative assumption that ICT usage will result in socioeconomic development and is always something we should strive for (Mariën et al., 2016).

Whether the digitizing of the society is a good thing, whether it is (always) desirable or not, and how technology should be used are often not questioned. Also, in the digital divide literature, ICTs are (almost always) treated as neutral goods. Understanding how technologies evolve and their implications for shaping the problematics of digital inequality seem to be left out of these often-used models in the digital divide debate (Gunkel, 2003).

2

22 | Chapter 2

2.3 | Interplay between digital technologies and society The development of the world wide web and the increasing upsurge of ICTs since the mid-1990s have undoubtedly been a significant force in improving economic and social conditions and contributing to human progress. On the other hand, technology has also introduced risks and unintended consequences. Think about cybercrime, invasion of privacy, bullying, and increased social inequalities. These claims about the social and/ or political implications of the rapid diffusion of technology, either positive or negative, view technology as the primary catalyst for social change. This deterministic view of technology is still used in the justification for and analyses of new technologies by many different actors (Wyatt, 2008). Technological determinism assumes that technological developments take place outside society: ICTs are shaped without any interference from the social and develop through their own technological logic. These socially neutral technologies then cause and determine social changes. In the simplest form of this explanation, technologies that prove to be valuable for society succeed, while unsuccessful technology does not break through because it is not successful in making a social impact.

Technological determinism is blind to the notion that technologies are not only developed though their own technological logic but are also constructed and used in a society in which there is space for ambivalence and multiple meanings. As summarized in the work of Wyatt (2008), empirical, most often ethnographic, studies in the field of science, technology, and society (STS) have demonstrated that the problem with technological determinism is that it absolves all of us of any responsibility for the technological choices we – individually or collectively – make. If we treat technologies as neutral and as having mysterious origins, there is no room for a discussion about the effect of human choice on design and there is nothing to be gained with questioning the development of specific technologies. In STS, various scholars do question the implications of the development of technology. For example, in the field of feminist STS, Donna Haraway (1985) uses the metaphor of a cyborg, a hybrid figure with human and machine parts, to show that humans and machines are inextricable linked and that technology is a strong field of normative action. She focuses in her writings on the power of technologies to make new worlds by creating new meanings or entities.

She thereby encourages critical feminist analyses of the politics of technologies and how technologies are intertwined with everyday lives. Wajcman (2007, 2010) introduced technofeminism, in which the relations between the social and technical determine how ICTs are developed and shaped, but technologies also affect whether social power relations are neutralized, weakened, or just reinforced. She argues that gender and ICTs are mutual shaping which implies that “technology is both a source and a consequence of gender relations” (Wajcman, 2004, p. 7). This applies

Theory and concepts | 23

not only for gender but also for other sociodemographic factors, such as race, class, or age (cf. Wajcman, 2007, 2010). STS scholars who follow the mutual shaping approach show how design and innovation can impact social, political, and economic arrangements and thereby highlight the problem of the exclusion of certain groups from the technological domain (Van Doorn & Van Zoonen, 2008; Wajcman, 2010). Els Rommes, for example, introduced the concept of ‘I-methodology’, in which designers (most often male) take themselves as model users, which has resulted in technologies that favour high-tech male users (Rommes, 2002; and described in Van Doorn & Van Zoonen, 2008). And Diane Forsythe (1993) showed that a system that was designed to empower patients ended up being a system that only reinforced power structures. She thereby highlighted that a wide set of subjectivities that go beyond the proto-typical ‘user’ should be included in the development of technologies. The notion of the mutual shaping of society and technology has three implications for studying digital inequality and the development of policies and practices. First, it emphasizes that we should explore and, thereby, understand the intersecting demographics that influence internet use and non-use in everyday life (Halford & Savage, 2010; Leurs & Ponzanesi, 2014; Wyatt et al., 2005).

Based on the work of Crenshaw (1989), writings on intersectionality emphasize that it is not only gender but other demographics – such as age, ethnicity, or generation – that influence how ICTs are used and experienced. For example, by studying the digital practices of Moroccan Dutch youth, Leurs and Ponzanesi (2007) uncovered the intersecting demographics, such as age, race, or sexuality, that influence online experiences. Moroccan Dutch youth create their online digital identities by making safe digital spaces in which to connect with their peers, but also by connecting to social networking sites which are in line with the dominant culture in which they grew up. Although quantitative studies emphasize the interrelatedness of socio-demographics and digital inequality (Schreeder et al., 2017), relatively little qualitative research has been conducted on understanding this relationship. Second, we should not only study how individuals’ experience ICT use but also how technologies are shaped and how power arrangements influence the function and meaning of ICTs (Van Doorn & Van Zoonen, 2008; Wajcman, 2010). For example, as explained above, cultural assumptions are embedded in design and influence the meaning of technologies and by whom they can easily be used (Forsythe et al., 1993; Rommes, 2002).

Last, forming these two implications, the specific use of technology and the consequences of design, also give rise to the central STS insight that: ‘it could be otherwise’. Socioeconomic inequalities need not result in digital inequalities (Mariën et al., 2016). This implies that we should open up the debate about how technology shapes or reshapes values in society and how development can be affected in such a way that it

2

24 | Chapter 2

results in more just outcomes. All of these three implications highlight the importance of public engagement in the governance of ICTs and the digital divide debate and, not least, the engagement of citizens who are dealing with digital inequalities.

2.4 | Public engagement Increasing scholarly attention is being paid worldwide to the possibility of public engagement opening up science and technology for a broader range of actors (e.g. Broerse & Buning de Cock, 2012; Dewey, 1927; Funtowicz & Ravetz, 1993; Jasanoff, 2003; Nowotny et al., 2001; Stirling, 2008; Thompson Klein, 2001). These scholars highlight that members of the lay public have experiential knowledge2 which might be different from scientific understanding but is not less valuable than scientific knowledge. The opening up of science to the public is conceptualized by many scholars by using terms such as participatory action research (PAR) (Freire, 1982), post-normal science (Funtowicz & Ravetz, 1993), mode-2 knowledge production (Jasanoff, 2003; Nowotny et al., 2001), or transdisciplinary research (Thompson Klein, 2001),

Conceptualization of ‘public engagement’ To conceptualize public engagement, I first need to define ‘the public’. The public, sometimes described as citizens, consists of people who have no scientific or professional affiliation with the (scientific) issue at hand. The public is not as homogeneous as the term suggests. Which subdivision is relevant depends on the issue at hand (cf. Broerse & Buning de Cock, 2012; Dewey, 1927). With respect to the digital divide, we can distinguish between multiple non-academic stakeholder groups: citizens who live very separately from the online world, their family and/or friends, professionals and volunteers who work closely with those who are excluded, ICT developers, policymakers, and so on. The term ‘engagement’, or ‘involvement’3, refers to research practices in which the lay public have a say. In other words, it is research with publics rather than about them. Publics can be involved in several ways and in different phases of the research. The framework of Smit et al. (2011) distinguishes between six modes of involvement: being the research object, information provider, advisor, assessor, co-researcher, or driving force (Figure 2.3). The differences between these modes are based on the level of control that the public has in the engagement practice and whether the engagement practice started from a public or a scientific interest.

2 Defined by Broerse and de Cock Buning as follows: “knowledge gained through experience as opposed to scientific knowledge that is acquired through scientific inquiry.”. (p. 674). 3 Only subtle semantic difference can be identified (Fumagalli et al., 2015), and therefore throughout this thesis, I use the terms interchangeably.

Theory and concepts | 25

2

Figure 2.3: Horizontal participation Ladder of Smit, van der Valk, and Wever (2011)

The ladder of Smit et al. (2011) is based on the well-known Arnstein’s ladder of participation (1969) but is, however, horizontal. With this change, they indicate that a higher level of control or whether the engagement practice was initiated from public or not is not by definition more ‘meaningful’ or ‘better’. The core idea is that the public can have valuable input in varying roles throughout the whole research cycle – from problem definition to dissemination of the findings.

Approaches underpinning public engagement Public engagement has a rich socio-intellectual history, and a variety of methodological approaches underpinning it can be identified. For example, patient and public involvement (PPI) draws upon work which argues for greater recognition for the significance of ‘lay expertise’ in research and industry (e.g. Jasanoff, 2003). Because it is rooted in the multidisciplinary field of STS, much research within this field has focused on studying practices of involvement carried out with policy actors or other researchers (Hagendijk & Irwin, 2006). Other PPI scholars are more actively involved in experiments in participation (Lezaun et al., 2016) and draw upon well-known methods such as interviews, surveys, advisory councils, focus groups, and on health-organizational experiments as well. One example of such an experiment was carried out by the Australian collective Xperiment!, in which elderly patients visualized their daily lives in drawings to create a deeper understanding of their everyday lives (Kraeftner & Kroell, 2005; see also: Lezaun et al., 2016). There are also methodological approaches which take a more radical view by taking the experience of citizens as the starting point of the research and involving citizens throughout the whole process. As summarized by Abma et al.

(2017), these approaches build on the rich history of traditions such as action research (AR), participatory research (PR), PAR, or community-based participatory research (CBPR). These traditions gained attention in different contexts and times. Kurt Lewin (1946), one of the champions of AR, emphasizes in his work on organizational development the importance of knowledge

26 | Chapter 2

gained from experience of those involved. His work particularly concerns the involvement of minority groups and strives for embedded action in order to improve the situation of these groups (Adelman, 1993). PR and PAR gained much of their popularity in the 1940s when academics used them to research communities and countries that were confronted with structural injustice, such as Africa, Latin America, and Asia (Freire, 1982). CBPR, in contrast, is grounded in the United States and is most often used in public health studies (Israel et al., 2008). The starting point of this thesis is the experience of citizen confronted with digital inequalities. In this thesis, I pay attention to power differences and strives for embedded action. Therefore, this thesis builds mainly upon the history of AR and PAR4.

2.5 | Public engagement in the debate on digital inequality In the digital inequality debate, multiple relevant non-academic stakeholder groups can be identified, such as citizens who are separated from the online world, their family and/or friends, professionals and volunteers who work closely with these citizens, ICT developers, policymakers, and so on. The inequality, in terms of knowledge, power, and resources, between these participating stakeholders poses a challenge to give everyone an equal voice and to manage potentially conflicting views (Abma et al., 2019; Broerse & Buning de Cock, 2012). In this thesis special attention is paid to the empowerment of those citizens with a distance from the online world. Citizens who are separate from the online world make up a diverse group, ranging from those who purposefully avoid the internet to those who face socioeconomic barriers to using it. The focus of the thesis is understanding the interrelatedness of social and digital inequalities. Therefore, when I refer to the public I am mainly referring to citizens with a low socioeconomic position (SEP), their family and friends, and those who work closely with these citizens. SEP refers to someone’s individual economic and social position in relation to others within a society (Galobardes et al., 2006).

Therefore, someone’s position depends on the context in which they are located. There is no unambiguous measure. A measure often encompasses a combination of (household) income, housing conditions, education level, and occupation (Galobardes et al., 2006). Throughout this thesis, I use the term ‘citizens with a low SEP’ interchangeably with ‘citizens living in vulnerable circumstances’. In the literature, the concept ‘vulnerable’ is often rather vaguely defined and open for discussion. The term is commonly used in social sciences, as well as in policymaking and health research, to refer to a large 4

See more on AR and PAR in chapter 3.

Theory and concepts | 27

variety of groups of individuals who are vulnerable in a specific study context (Larkin, 2009; Virokannas et al., 2018; Walker & Fox, 2018). For example, Smith (2008) defines ‘vulnerable groups’ as those who cannot look after their own interests and therefore risk being harmed in health services research. In contrast, Dixon-Woods (2005) considers citizens in vulnerable cirucumstances to be those with a minority ethnic background, children, and elderly people as vulnerable. Also, in policy documents, the definition of vulnerability or vulnerable groups differs according to the primary goal of the organization and the context (Box 2.1). In this thesis, I purposefully choose, however, to use the term ‘citizens living in vulnerable circumstances’ rather than ‘vulnerable groups’. Citizens are not themselves intrinsically ‘vulnerable’, but the context places these citizens in vulnerable circumstances (Walker & Fox, 2018). Second, using ‘vulnerable groups’ risks these citzens only being seen as vulnerable rather than as knowledgeable social actors who can be actively involved in engagement practices and can influence the world they live in. In the context of this thesis (the Netherlands – a post-welfare state), I define citizens living in vulnerable

Box 2.1: Context-dependent definition of vulnerability or vulnerable groups in policy documents World Health Organization: Vulnerability is the degree to which a population, individual, or organization is unable to anticipate, cope with, resist, and recover from the impacts of disasters. Children, pregnant women, elderly people, malnourished people, and people who are ill or immunocompromised are particularly vulnerable when a disaster strikes (Wisner & Adams, 2002). American Public Health Association and Public Health Institute: The degree to which an individual or group of people are susceptible to harm. In public health, this term is commonly used in the fields of disaster response and climate change adaptation, but can also apply to the broader social determinants of health. Vulnerability of individuals and populations to injury or chronic disease can vary as a result of socioeconomic status, social cohesion, gender, health, age, and other factors (Ben-Moshe et al., 2013, p. 137). Bureau for Economic Policy Analysis [In Dutch: Centraal Plan Bureau]: Dutch citizens are vulnerable if they lack the resources which can help them to prevent problems and achieve a good quality of life. These primarily concern income, education, health, and labour (Pommer & Boelhouwer, 2017).

2

28 | Chapter 2

circumstances as those citizens who have low socioeconomic resources and, therefore, cannot achieve a good quality of life (cf. Pommer and Boelhouwer, 2017). In this thesis, I do research with citizens living in vulnerable circumstances or citizens with a low SEP. In other words, citizens who have to deal with socioeconomic inequalities. But what does this imply for an individuals in post-welfare states? What does it mean? And what are the mechanisms behind socioeconomic inequalities in post-welfare states?

Socioeconomic inequalities in post-welfare states Socioeconomic inequality is a global issue: it exists everywhere to some degree, regardless of how wealthy a country is. In many Western countries, economic prosperity has steadily and significantly increased since the period of economic recovery after World War II. This, however, has not necessarily resulted in a decreased number of people living in poverty (Alvaredo et al., 2018; Piketty, 2015). Piketty and his associates make it clear that uncovering the distribution of top incomes is more potent than studying overall measures of inequality (such as the Gini-index, a inequality measure based on a frequency distribution of income) to make the public aware of income inequalities (Galvin & Sunikka-Blank, 2018; Milanovic, 2014; Piketty, 2015). Based on their work, the World Inequality Report 2018 shows that, since 1980, income inequality within developed countries, particularly extreme in the United States (US), has significantly increased (Alvaredo et al., 2018).

Neoliberal policies – first implemented in the United Kingdom (UK) and the US after the election of, respectively, Margaret Thatcher (1979) and Ronald Reagan (1981), and progressively, in varying degrees, implemented in other AngeloSaxon and Western European countries – had a significant influence on the increasing income inequalities between the top-earning 10% and the share of the lowest-earning 10%. But other trends, such as the financial crisis in 2007 and 2008 and the increasing sophistication of tax havens, have also increased income inequalities in developed countries (Galvin & Sunikka-Blank, 2018). In the Netherlands, the country where the various studies of this thesis took place, income inequality has also increased (Alvaredo et al., 2018). In 2017, the number of Dutch citizens who were living in poverty was 6.6%, higher than in 1990 (when 5.7% were poor) (Sociaal en Cultureel Planbureau, 2016a). Poverty is defined as a household that earns less money than the ‘basic needs budget’, which reflects the amount of money needed to run an independent household and to pay for unavoidable basic expenditure on items such as food, clothing, and living expenses. The ‘basic needs budget is based on the costs for one household and in 2014, for example, was 1,830 euros for a two-parent family with two children (Sociaal en Cultureel Planbureau, 2016a).

Theory and concepts | 29

The Dutch neoliberal austerity policies, introduced in response to the 2007–2008 financial crisis, resulted in the shrinking role of the state and increased the emphasis on engaging civil society. The Dutch welfare state reformed into a post-welfare society (term was already introduced by Nick Ellison in 1997); in the Netherlands this is often called a ‘participation society’. Care-related tasks and social support (i.e. elderly care, youth protection) were decentralized from central to local governments. Individuals were increasingly held responsible for their own lives and invited to provide informal care to friends or family. Consequently, citizens living in vulnerable circumstance first needed to work things out on their own or with help from family and friends, with professional support becoming the last resort. This is often portrayed as the ‘de-professionalization’ of care and social work (Trappenburg & van Beek, 2019; Verhoeven & Tonkens, 2013). According to several Dutch research institutes, the significant switch in tasks and responsibilities seems to create a wider gap between rich, highly educated civilians and poor civilians with a low educational level.

They emphasize that citizens living in vulnerable circumstances, such as citizens with a low literacy level, a low educational level, mental health issues, and/or those who need to make difficult (financial) tradeoffs, are often not able to meet the high demands of the post-welfare state (National ombudsman, 2016; Sociaal en Cultureel Planbureau, 2016b; Wetenschappelijke Raad voor Regeringsbeleid, 2017). Also, the de-professionalization of social work means that citizens living in vulnerable circumstance are increasingly dependent on their social network. Their network is, however, often small and not adequate for helping them with the high demands of the (digitized) participation society (Trappenburg & van Beek, 2019).

Complexities of the engagement of citizens with a low SEP To have meaningful deliberations on the governance of digital inequality, it is essential to involve a variety of actors, not least those who are dealing with social and digital inequalities. In practice, the engagement of these citizens is a complex endeavour (Read & Maslin-Prothereo, 2011; Snow et al., 2018). Just giving these citizens living in vulnerable circumstances ‘a seat at the table’ is not enough to empower them to be meaningfully and effectively engaged in research or policymaking. The challenge is to meaningfully embed their experiential knowledge in decision making and not consider them just as ‘another input’. Exclusion, the process whereby certain relevant stakeholder groups are not part of the decision-making progress or their perspectives are not considered, can be seen in several processes. First, the exclusion of citizens living in vulnerable circumstances will not hamper the progress of research or innovations. Mander (2018) talks about ‘imposed invisibility’; this term indicates that there is almost a denial of the existence of these

2

30 | Chapter 2

citizens and their need to participate. For example, sending online invitations to citizens asking them whether they want to participate in a public debate will exclude those without access to ICT and those without ICT skills or any understanding of the questions. In addition, people who are unfamiliar with research, perceive themselves not to be experts, or mistrust research or official institutes will not be engaged (Isler et al., 2015; Lionis et al., 2016). Although, their perspectives will not be considered, the public debate will still proceed. Second, it is often easier for researchers or policymakers to involve the more knowledgeable citizens rather than the less knowledgeable ones (Broerse & Buning de Cock, 2012). Researchers consider it to be ‘too difficult’ to involve citizens living in vulnerable circumstances (Bonevski et al., 2014; Ellard-Gray et al., 2015; Whitehurst, 2007). For example, by excluding citizens living in vulnerable circumstances, a researcher can avoid lengthy and challenging ethical procedures. Sometimes, citizens living in vulnerable circumstances are even considered as not able to participate in scientific or technological developments. Only ‘scientific citizens’, that is, citizens with certain skills and knowledge, are considered educated enough to participate in scientific research and technological developments.

Third, there is a lack of awareness that research methods are not suitable for trying to include citizens living in vulnerable circumstances (Aldridge, 2014). Frequently used engagement formats such as advisory councils and questionnaires are not inclusive regarding those who are less literate or not confident enough to speak up during a group meeting (Kaiser et al., 2017; Snow et al., 2018) Also, intimidating settings or those that make someone feel uncomfortable might result in the self-exclusion of groups that are not familiar with research or policymaking (Broerse & Buning de Cock, 2012; Elberse et al., 2011). To summarize, the aim of this thesis is to facilitate the development of policies and practices that address digital inequalities so that they are more sensitive to the complex daily realities of citizens living in vulnerable circumstances. Even though this is often experienced as difficult, in this thesis, the views, perspectives, and needs of these citizens are central. In the next chapter, the aim is operationalized and the action research approach is described. I especially zoom in on how we engaged those citizens who are often considered to be hard to reach.

Theory and concepts | 31

2

Research design: action research for change

34 | Chapter 3

As described in chapter 1, the central objective of this thesis is to facilitate the development of policies and practices that address digital inequalities so that they are more sensitive to the complex daily realities of citizens with a low socioeconomic position (SEP) by (1) gaining insight into the complex daily realities of citizens with a low SEP and understanding their needs in the digitizing society, and (2) facilitating and reflecting on the meaningful engagement of citizens, researchers, technology developers, and policymakers in digital inclusion. To reach these objectives, the following main research question will be central: How can we build policies and practices that address digital inequalities so that they are more sensitive to the complex daily realities of citizens with a low SEP? In this chapter I first describe the research context. Thereafter, I zoom in on the chosen research approach and the methods used in the studies conducted to answer the main research question.

3.1 | Research setting This thesis builds mainly on action research projects conducted in Amsterdam. Amsterdam has over 850,000 citizens and is the most populous city in the Netherlands. As in many other cities in post-welfare societies, in Amsterdam the number of people living in poverty has increased during the last 30 years. Compared to other parts of the Netherlands, the city has a relatively high number of households living in vulnerable circumstances (Sociaal en Cultureel Planbureau, 2016a); they have a substantially lower income, live in poverty for a relatively long time, or suffer from generational poverty. In 2017, 24% of households had a low income (defined in Amsterdam as 120% of the statutory social minimum). Also, a quarter of young people were growing up in poverty (Michon et al., 2017). Importantly, living in poverty means more than just having a low income. These are households that usually lead an isolated existence, with a focus on the here and now. In addition, citizens living in poverty are usually poorly educated (26%), have a low level of literacy (18%) and have problems with housing, health, work, or their income (Bijlsma et al., 2016; Michon et al., 2017).

The action research was conducted in socioeconomically disadvantaged neighbourhoods of Amsterdam – Amsterdam East, North, West, and NewWest – where the percentages for the factors just mentioned are even higher.