5 Thesis Nicole Goedhart.pdf

Research design | 35

Socioeconomic inequalities and intergenerational immobility5 were high on the policy agenda in 2017 in Amsterdam. In the coalition agreement ‘Amsterdam is van iedereen 2014–2018’ [in English: Amsterdam belongs to everyone 2014–2018], the council of aldermen states that all citizens from Amsterdam must be able to participate fully in society, regardless of their financial position. The starting point of the coalition agreement is equal opportunities. To combat inequalities, the City of Amsterdam installed a poverty director in 2016 on the initiative of the councillor responsible for the work, income, participation, and poverty portfolios. This action research was initiated based on a small and practical policy question of the poverty director in Amsterdam in 2017. She stated that children who live in disadvantaged neighbourhoods cannot meet the (online) digital demands of schools and was, therefore, asking for input regarding how to meet the needs of children. A group of professionals who had direct experience with the so-called target population were invited to a meeting to discuss how the City of Amsterdam could accommodate these children. Many questions could not be answered in this meeting, e.g.

what devices do children need in school and how can we support these children in their (online) social life? As a result, a team of professionals was invited to collaborate in a research project to explore the needs of children in the digitizing society. The focus rapidly expanded to not only the needs of children but also to the needs of their parents.

3.2 | Action research approach A unique characteristic of this thesis is that it started from a seemingly minor practical policy-related question. In close cooperation with citizens with a low SEP, volunteers and professionals working with these citizens, and the City of Amsterdam, this practical question evolved to multiple research projects. In these projects, we aimed to collaboratively develop policies and practices that address digital inequalities so that they are more sensitive to the complex daily realities of citizens with a low SEP. To facilitate this, we built upon the principles underlying action research approaches. The term action research was first coined by Kurt Lewin, a German ‘Gestalt psychologist’ who in his work after World War II primarily focused on research concerning practical problems in which he simultaneously worked on theory building and encouraging social progress (Adelman, 1993; Boog, 2003). In 1940, after a few practical experiences, Lewin 5 Intergenerational mobility refers to any – positive or negative – changes in the socioeconomic position of family members between generations. Intergenerational mobility has a negative relationship with income inequality, which means that families persist at the top or bottom in terms of income over (two or more) generations.

Intergenerational immobility is especially high in countries with high educational inequalities and a lack of wage-setting policies in favor of the low- and middle-income groups (Alvaredo et al., 2018).

3

36 | Chapter 3

stated, “No action without research; no research without action” (Adelman, 1993, p. 8). In action research, researchers deliberately not only observe or interpret a complex (social) problem at stake but also provoke and learn from (social) change. At the time of writing, many traditions of action research can be found in several scientific disciplines (Boog, 2003; Bradbury-Huang, 2010; Reason & Bradbury, 2001b). Some traditions point first to the importance of participatory democracy (Freire, 1982; Heron & Reason, 1997). Others mainly focus on achieving a deep understanding of practices through cooperative learning in action and experimenting (Greenwood & Levin, 1998), or intervening in practice to develop knowledge and new normativities (Zuiderent-Jerak, 2015). While these traditions have their own focus and underlying assumptions, in this thesis, we paid particular attention to the following four linked characteristics of action research. First, we approached our research on the normative assumption that citizens with a low SEP have the right to be involved in issues that affect them, like the debate on the digital divide. A mode of action research which particularly emphasizes this point is participatory action research (PAR).

PAR draws on Marxism and emancipatory writings and promotes the engagement of marginalized communities to challenge the dominant knowledge of elites (Abma et al., 2017). In PAR, the engagement of citizens can be seen as an end in itself (International Collaboration for Participatory Health Research (ICPHR), 2013a). This key facet of PAR is made explicit by Baum et al. (2006, p. 854): At its heart is collective, self-reflective inquiry that researchers and participants undertake, so they can understand and improve upon the practices in which they participate and the situations in which they find themselves. [...] The process of PAR should be empowering and lead to people having increased control over their lives. PAR also emphasizes the importance of including those who seldom have a voice in research and policymaking (Abma et al., 2019; Cornwall & Jewkes, 1995; International Collaboration for Participatory Health Research (ICPHR), 2013b; Khanlou & Peter, 2005). To empower these citizens, it is essential to share control between researchers and involved citizens. In other words, it is not the questions of researchers that should be the guide, but the stories of those involved.

To support citizens to reflect on their own stories within PAR, creative and less-conventional methods are often used, such as drawings, photo-collages, or videos (Abma et al., 2019; Aldridge, 2007, 2016; Kara, 2015; Sarti et al., 2017). These creative methods create an opportunity to share stories in a less text-based manner and can provide room for different types of knowledge, i.e. embodied knowledge or emotions (Abma et al., 2019).

Research design | 37

Second, we treated this action research as an iterative research practice (Cook, 2009). In action research, attention is given to problems and questions that occur in practice throughout the research. As emphasized by Lewin (1946), action research should not be done in remote, isolated, experimental settings but within the complex, daily, holistic social setting, focusing on the daily realities of people. This means that action research is done in interaction with the social system in which it takes places (Bradbury-Huang, 2010; Gray, 2014). Creating knowledge in interaction results in new questions and/or problems. To respond to what is happening in practice, action researchers move through cycles of action and reflection to reformulate their research plan (Banks et al., 2017; Burns, 2007; Reason & Bradbury, 2001b). These cycles of action and reflection are operationalized by Kemmis and McTaggart (1988) in the well-known ‘spiral of self-reflection’ (Figure 3.1). This spiral consists of continuous cycles of planning a change, acting, observing what is happening, and reflecting, which often results in a revised plan. In reality, these stages overlap, which makes action research a deeply iterative process. Throughout the cycles of action and reflection, power and decision making are shared between the participants and the researchers.

In this research, we used the cycles of action and reflection to revise the design of the studies but also to determine the focus of subsequent studies. Thematic concern

Reconnaissance

Reflection

Observation

Plan

Action Reflection

Observation

Revised plan

Action Revised plan

Figure 3.1: Spiral of self-reflection of Kemmis and McTaggart (1988)

3

38 | Chapter 3

Third, attention is paid to the notion that action researchers not only desire to create social knowledge but also to simultaneously provoke democratic social change (Abma et al., 2019; Bradbury-Huang, 2010; Greenwood & Levin, 1998). The collaborative action research process aims to support action, which leads to more just and satisfying practices for the involved stakeholders. In our research, we tried to provoke social change by also working closely with policymakers and professionals in the field. We simultaneously shared the acquired knowledge in forms other than academic papers or books, since knowledge expressed in an easily accessible, creative, and impactful way seems to have more potential to ‘travel’ outside scientific fields and provoke change (Abma et al., 2019; Downey & Zuiderent-Jerak, 2021). Fourth, we adopted a self-reflective stance towards our role throughout the action research process. With the citizens involved, researchers aim to get a grip on the thorny problem at stake and to develop new actions through cooperative learning (Boog, 2003; Bradbury-Huang, 2010). Scholars become involved in the context of the study and emerging actions.

This personally active role of the researcher requires a self-critical stance towards the researchers’ identities, biases, and (changing) normative assumptions throughout the whole action research process and towards the emerging actions (Bradbury et al., 2020; Zuiderent-Jerak, 2015). These are the four linked characteristics that underlie the studies conducted, while deploying a wide range of research designs. The studies will be discussed in more detail next.

3.3 | Doing (participatory) action research on digital inequalities This action research consists of four sub-studies conducted between January 2017 and June 2020. Although research often starts with a literature review to explore the scientific landscape, we started from the complex daily reality of citizens with a low SEP. We made this choice on the normative assumption that citizens have the right to be involved in policies and practices that affect their lives. This first PAR study, in which the perspectives and needs of citizens with a low SEP were central, was decisive for the design of the subsequent studies. In the following three sub-studies, we further unravelled questions that emerged from this PAR study. To unravel these questions, we involved a wide variety of stakeholders, i.e. from citizens with a low SEP to policymakers to volunteers and professionals working closely with citizens with a low SEP to information and communication technology (ICT) designers.

Research design | 39

In this thesis, the four studies address the two objectives. The results of these studies are described in chapters 4 to 8. Table 3.1 gives an overview of the studies, the research objectives addressed, and the chapters in which the results are given. I discuss the studies and research questions in more detail in the next sections.

Study 1: digital inequality in Amsterdam The PAR study on digital inequality was initiated and commissioned by the City of Amsterdam. This study was conducted between January 2017 and February 2018 in close collaboration with families with a low SEP, volunteers and professionals working closely with these families, and policymakers working for the City of Amsterdam. The aim of this study was 1) to gain insight into the experiences and needs of families with a low SEP in the digitizing society, and 2) to co-create concrete operational (policy) solutions for the City of Amsterdam. This PAR study consisted of two phases of data collection. The first phase focused on understanding the complex daily reality of citizens with a low SEP in the digitizing society. We conducted interviews with professionals (n=42), parents (n=10), and children (n=3), and further increased the understanding through two work sessions with professionals (n=39) as well as participatory observations (n=18) in community centres and formal language classes. Phase two focused on the co-creation of solutions. Three co-creation sessions with volunteers and professionals (n=64) working closely with citizens with a low SEP were hosted, followed by four focus groups with mothers (n=21) and two focus groups with children (n=30). Table 3.1: Overview of the studies Study

Focus of the study

Objective 1 Objective 2 Chapter

1

Perspectives, experiences, and needs of mothers with a low SEP in the digitizing society

4

Reflection on the challenges of the complex reality of PAR for local policy addressing digital inequalities

5

2

Reflection on the use of vlogs in PAR as a way to facilitate shared learning between citizens with a low SEP, peers, researchers, and policymakers

6

3

Reflection on tools and methods, described in the literature, used to meaningfully engage citizens living in vulnerable circumstances in health and social care research

7

4

Perspectives of developers on how digital inequality is perpetuated in the design of patient portals

8

Notes Dark blue = main focus of the study; Light blue = additional focus of the study; SEP = socioeconomic position; PAR = participatory action research

3

40 | Chapter 3

In this PAR study, we worked closely with gatekeepers of community centres to also give a voice to those citizens who are not already known to professional organizations (such as social welfare organizations). To empower these citizens to be engaged, we kept in mind their daily reality when planning and designing the activities. For example, we used creative methods to empower citizens, especially the ones with low literacy levels, so that they could have a meaningful contribution. Also, we often had deliberations with a welfare professional who had worked as a volunteer for more than 10 years to maintain an ongoing awareness of the situational context of the involved families, to be continuously reflective regarding our research practices, and to validate our findings. To foster change throughout this PAR study, we worked closely with policymakers. We had (bi)weekly meetings to discuss our findings, they helped us with the planning of research activities, and were present during some of the research activities. This close collaboration contributed to the translation of the research findings into policy actions. The results of this first PAR study are summarized in an accessible (Dutch) policy report supported by expressive and provocative drawings6 (see Dedding et al., 2017), and in two chapters of this thesis.

In chapter 4 of this thesis, I particularly zoom in on the perspectives, experiences, and needs of mothers with a low SEP. We focused on the perspectives of women because of the long-existing gender inequalities in internet use (Schwanen et al., 2014) and their essential role in supporting their children in their use of ICTs (Eynon & Geniets, 2016). In this chapter I address the following sub-question: Q1: What are the perspectives, experiences, and needs of mothers with a low SEP in the digitizing society? In chapter 5, we zoom in on the PAR process. We reflect on our role and choices as researchers, the collaboration with citizens with a low SEP, professionals and volunteers working closely with these citizens, and with policymakers. Also, we reflect on whether this research contributed to the creation of policies and practices that are more sensitive to the complex daily realities of citizens with a low SEP. The following sub-question guides this chapter: Q2: What can be learned from the challenges of the complex reality of PAR for local policy that addresses digital inequalities? The results of this first PAR study has informed the focus of the subsequent studies. In this study, we learned collaboratively that we should not only deepen our knowledge about the perspectives and needs of those citizens who are digitally excluded, but should 6

The drawings are by Jochem Galama; see: http://new.contour-ids.nl/

Research design | 41

also pay attention to the roles and responsibilities of researchers, policymakers, and technology developers in the digital divide debate. In studies 2 and 3, I therefore focus in more detail on how science and policy collaborate with citizens living in vulnerable circumstances and how these citizens could be meaningfully engaged as knowledgeable actors. In study 4, we broaden our understanding by unravelling, from the perspectives of technology developers, how digital inequality is perpetuated in the design of technology.

Study 2: reflection on the use of vlogs in PAR to address digital inequalities Throughout the first study, the participating professionals and volunteers often spoke enthusiastically about the idea of co-creating vlogs to engage citizens living in vulnerable circumstances in policy development as a ‘nice, easy, shareable, quick solution’ to peereducate citizens with low ICT skills. Mothers with a low SEP supported this statement by highlighting that YouTube is an often-used source to find information. Additionally, vlogs ‘which can go viral’ seemed to be an attractive policy instrument to use to reach out to a large group of citizens with limited resources. In this PAR study, we therefore took up this challenge to unravel how the co-creation of vlogs could (1) facilitate engagement of citizens with a low SEP, (2) promote (shared) learning between citizens, researchers, and policymakers, and (3) reach out to a wider audience and thereby provoke social change. The sub-question7 that guided this study was: Q3: What can be learned from the use of vlogs in PAR as a way to engage citizens living in vulnerable circumstances and enhance their well-being? This PAR study was commissioned by the City of Amsterdam and conducted in close collaboration with three community centres spread across the city (in Amsterdam NewWest, Amsterdam East and Amsterdam North).

Between November 2018 and April 2019, vlogs were co-created and evaluated with women from disadvantaged neighbourhoods in Amsterdam with a wide range of non-Western nationalities, e.g. Indian, Iraqi, Moroccan, and Chinese. In two projects, the vlogs were co-created over a period of five to seven weeks. Both projects were concluded with a dialogue session in which the vlogs were discussed with a policymaker from the City of Amsterdam. Additionally, the vlogs were evaluated with women who were not involved in the co-creation of the vlogs.

7 In chapter 6, this sub-question is addressed by reflecting on two cases in which vlogs were co-created with citizens, respectively adolescents (case 1) and women (case 2), living in disadvantaged neighborhoods. The aim of case 1 was to co-create vlogs to gain insight into adolescents’ daily reality and perspectives on health and health promotion. Case 2 aimed to gain insight into how women could use vlogs to improve their digital skills. Only the latter case is part of this thesis. Therefore, in this section, we only describe the methods used to co-create vlogs with women living in disadvantaged neighborhoods in Amsterdam.

3

42 | Chapter 3

Gatekeepers of the three community centres involved assisted in planning the events and inviting potential participants. In all these projects, we started with a meeting that aimed to build trust by getting to know each other. There was room for the women to share the difficulties they experienced when using the internet and the researcher(s) provided some help at this point with acute (but minor) problems. In the subsequent meetings, a question or problem of the women involved was central, which we tried to solve together first and then recorded the solution in a vlog. To evaluate the vlogs, we started the meetings by watching and reflecting on one of the vlogs and then started practising what was explained in the vlog. In the meetings, I was supported by a volunteer or an Arabicspeaking colleague. The reflection on the use of vlogs as a way to engage citizens living in vulnerable circumstances and to enhance their well-being can be found in chapter 6. In line with the nature of PAR, the design of this study was iterative. The initial idea was that co-creating vlogs could be a ‘nice, easy, shareable, quick solution’ to engage educated citizens with low ICT skills, but we learned collaboratively that this is a more complex process. In the first two projects, the women involved often emphasized that ICT education should be more integrated into language classes.

Therefore, besides evaluating the vlogs, we also deepened our learning regarding how to integrate ICT into language classes. To provoke change, these substantive lessons are included in a co-created paperbased end product designed especially for professional and voluntary Dutch language teachers (see Appendix 1).

Study 3: narrative review focus on tools and methods that can be used to meaningfully involve citizens with a low SEP In the context of the previous two studies, we often realized that engagement of citizens with a low SEP is not an everyday practice for policymakers. Also, the literature shows that citizens living in vulnerable circumstances are often not involved in engagement practices in health and social research. To broaden our learning concerning the facilitation of meaningful engagement of citizens with a low SEP (chapters 5 and 6), we conducted a narrative review with a systematic search. The search of this narrative review was broad and flexible which was necessary to formulate a precise aim and sounds inclusion and exclusion criteria (Boell & Cecez-Kecmanovic, 2014; Greenhalgh et al., 2019). In this review, we aimed to critically analyse strategies, tools, and methods that could be used to support citizens with a low SEP who want to be engaged in research and policymaking. The research sub-question that guided this narrative review was: Q4: What lessons can be learned regarding the facilitation of meaningful engagement of citizens with a low SEP in research?

Research design | 43

In this narrative review, we systematically searched for articles. Our broad initial search found 6,782 articles. Through a flexible and interpretive screening approach, we ended up with 40 articles for in-depth analyses. Through thematic analysis, we search for stratergies, tools, and methods that could be used to support the inclusion of citizens living in vulnerable circumstances in research. This review did not aim to develop a one-size-fitsall framework since there is no single best approach. The people involved – their needs and competences – the specific topic, the aim, and the situation determine what works best for whom and why (Dedding & Slager, 2013; Greenhalgh et al., 2019). This review provides insights into points of direction based on published empirical data. The results of this narrative review can be found in chapter 7.

Study 4: perspectives of developers on equitable design We increasingly became aware that policies and practices addressing digital inequalities are often focused only on the demand side, that is, on the training of users living in vulnerable circumstances; but people who also have to deal with poverty and/or illiteracy do not necessarily have enough energy to learn new things. But what about the supply side, that is, the developers of ICT? We learned collaboratively that to gain insight into the complex problem of digital inequalities, we should also pay attention to how technology is developed. To explore in more detail the relation between technology design and digital inequalities and how design decisions might systematically and unfairly discriminate against specific individuals or groups, we addressed the following sub-question: Q5: How do the (conceptual) assumptions of designers and other professionals responsible for the implementation of patient portals and/or electronic Patient Health Records (ePHRs) strengthen or otherwise reconfigure the dynamics of inclusion and exclusion of citizens with a low SEP? This sub-question was studied in the context of patient portal development and implementation in the Dutch health-care system, in which access to health-care is increasingly digital.

We chose to explore these questions in the context of the Dutch health-care system since citizens living in vulnerable circumstances have more difficulties using patient portals (Patiëntenfederatie Nederland, 2017; Vreugdenhil et al., 2019). Citizens with a low SEP not only often experience difficulties using online services, but also live seven years less and enjoy 19 years less in good health than those with a high SEP (VTV, 2018). This qualitative study was conducted between March and November 2018. Data collection was performed using interviews to understand the experiences and needs of portal designers, health-care professionals, and policymakers. We used a semi-structured interview guide that included questions about the following: (1) the need to make patient

3

44 | Chapter 3

portals and/or ePHRs accessible for everyone; (2) mechanisms that would support or prohibit the inclusion of citizens living in vulnerable circumstances; and (3) awareness of people with low digital skills. We conducted 24 interviews. These were followed by an open-ended questionnaire for chief medical information officers (CMIOs) – medical professionals who play a significant role in digitalization in hospitals. In total, 14 CMIOs completed the questionnaire. The results of this qualitative study can be found in chapter 8.

3.4 | Validity Originally, the concept of validity was used in quantitative research. Research with a participatory (action) approach requires different criteria for validity since it involves human understandings rather than instruments with technical characteristics (Edwards et al., 2008; International Collaboration for Participatory Health Research (ICPHR), 2013a). In addition, collaborations with citizens living in vulnerable circumstances can be complicated by cultural differences, fear of stigmatization, community research fatigue, and other conditions which should be taken into account by reflecting on the validity of a participatory research process. Therefore, I chose to use the six validity criteria for PAR as defined by International Collaboration for Participatory Health Research (2013) for our studies (see Table 3.2). I discuss the six validity criteria next, except for ethical validity, which is addressed in section 3.5, Ethical considerations. Attention was paid to participatory validity. To support the engagement of (often busy) professionals, a wide range of methodologies – (informal) interviews, focus groups, co-creation sessions, and a questionnaire – were included. To support the inclusion of citizens with a low SEP, we considered their daily reality when planning and designing the activities.

First, the meetings were always scheduled at times that suited their daily lives. Also, mothers were welcome to bring their toddlers to activities that took place during school hours. Furthermore, we often joined existing activities, such as language classes, to minimize the time investment of the citizens involved. If we joined existing activities, we were sympathetic to the main activity while collecting data. Second, the venues were carefully selected. For the citizens involved, it is not an everyday business to travel from one side of the city to the other. Therefore, meetings were organized in their neighbourhood, most often in community centres. An additional benefit of community centres is that they are familiar to those involved and seen as safe locations. Familiarity and a sense of security gave citizens enough confidence to express their views (Redwood et al., 2012).

Research design | 45

Table 3.2: Specific validity criteria for participatory health research as defined by International Collaboration for Participatory Health Research (2013) (p. 20) Underpinning value

Explanation

Mutual respect

Developing research relationships based on mutual respect.

Equality and inclusion

Encouraging and enabling people from a range of backgrounds and identities (e.g. ethnicity, faith, class, education, sex, sexual orientation, (dis) ability, age) to lead, design and take part in the research,

Democratic participation

Encouraging and enabling all participants to contribute meaningfully to decision-making and other aspects of the research process according to skills, interests and collective needs.

Active learning

Seeing research collaboration and the process of research as providing opportunities to learn from each other

Making a difference

Promoting research that creates positive change for communities of place, interest or identity.

Collective action

Individuals and groups working together to achieve change.

Personal integrity

Participants behaving reliably, honestly and in a transparent and trustworthy fashion.

Third, we did not adopt any exclusion criteria regarding language, but deliberately chose methods in which the Dutch language was less critical (Kara, 2015), i.e. we used drawings and pictures in the meetings to support the spoken word. Fourth, we were flexible regarding attendance of the meetings. The challenges the women face in daily life meant that some of them could not be present at multiple sessions or needed to cancel at the last minute if, for example, their child was sick. In some instances, the women who had been invited spontaneously brought along other women for fun or because these women were also experiencing difficulties while using the internet. To be inclusive, we accommodated these changes in the composition of the group. Last, in collaborations with the citizens, we made time to build trust and to get to know each other. Building trust was essential to empower these citizens, especially those who had a lack of trust in public authorities (such as the City of Amsterdam), to realize that they had something to add and were willing to share their personal experiences (Isler et al., 2015; Kaiser et al., 2017). The significant diversity of the stakeholders involved contributed to the intersubjective validity. Many of the stakeholders thought that addressing digital inequalities was an urgent need.

For example, the gatekeepers of community centres were very willing to help us. They recognized the problem of digital inequality in their (daily) work, but many had paid no attention to this yet. The gatekeepers were, therefore, eager to learn how they could support their members in the digitized world. To increase the meaningfulness of this research for citizens living in vulnerable circumstances, we actively explored and

3

46 | Chapter 3

discussed topics that they were interested in at the meetings. For example, in study 2, vlogs were co-created about digital problems the women involved faced in their daily life. To realize contextual validity in this thesis, we took into account the experiential knowledge of multiple actors and collected data in real-life situations (Chalmers et al., 2014; Council for Public Health and Society, 2017). For example, in study 1, we discussed with citizens living in vulnerable circumstances existing interventions addressing digital inequalities, such as an app developed by the municipality called ‘understand the letter’. We tried using this app together and explored the technical difficulties the women were experiencing in their daily life, e.g. if you don’t speak Dutch, it is scary to leave your number to be called back The generalizability of this research may be low because the contextual validity level is high (Gray, 2014; Thomas & Magilvy, 2011). We tried to increase the generalizability, however, by giving detailed descriptions of the daily reality and socio-demographics of the citizens involved (‘vicarious learning’). In addition, we paid attention to the mechanism underlying digital inequality, which might apply in other metropolitan post-welfare contexts. These mechanisms were compared and validated using the relevant literature.

Catalytic validity was central throughout this research. In general, we tried to show in all of the studies that citizens living in vulnerable circumstances are knowledgeable actors who should be involved in research and policymaking concerning ICTs. These citzens should not only be seen as problem bearers but also as part of the solutions. Through close collaboration with policymakers working for the City of Amsterdam, we tried to support policy development that was more in line with the daily reality of citizens with a low SEP. Also, we translated our results into two proactive (Dutch) end products that were supported by expressive drawings. These products are easily accessible and thereby have the potential to provoke action (Abma et al., 2019). In addition, by talking about digital inequality with professionals in the field of ICT design and/or implementation in interviews, we hoped to increase the awareness of digital inequality among designers, health-care professionals, and policymakers and thereby to trigger a process of re-evaluation of their work processes (Banks et al., 2017). Attention was paid to emphatic validity by sharing the detailed stories of citizens with a low SEP in easily accessible products and in vlogs. These detailed and tangible stories helped policymakers to understand the daily reality of citizens with a low SEP.

In study 2, the vlogs showed in detail the daily struggles of the women in the digital world. These vlogs allowed the practice of looking ‘alongside’ rather than ‘at’ the citizens involved (Kindon, 2003) and thereby helped to provoke empathy with what is expressed.

Research design | 47

3.5 | Ethical considerations Ethics are often operationalized in guidelines or codes to ensure that research does not harm participants, respects the participants’ privacy and human dignity, and that everyone involved is completely informed and gave consent to participate. The studies in this thesis adhered to the Code of Ethics for the Social and Behavioural Sciences involving human participants as accepted by the deans of social sciences in the Netherlands in January 2016. The studies did not need formal approval of a medical ethics committee, according to the Dutch Act on Medical Research Involving Human Participants (Wet medisch-wetenschappelijk onderzoek met mensen [WMO]). Over and above these general research guidelines or codes, in (participatory) AR, researchers should pay attention to ethical issues that arise throughout the process (Abma et al., 2019; International Collaboration for Participatory Health Research (ICPHR), 2013b). The iterative process of PAR means that it is often unknown how the research will evolve (Banks et al., 2013). In addition, the changing role between researchers and participants and power differences between the stakeholders involved can generate ethical challenges throughout the process (International Collaboration for Participatory Health Research (ICPHR), 2013b). Ethical reflection should therefore be an integral part of doing PAR (cf.

Banks & Brydon-Miller, 2019; International Collaboration for Participatory Health Research (ICPHR), 2013b). To address these challenges in all of the research projects, we continuously paid attention to the seven underlying values and ethical principles of PAR: (1) mutual respect, (2) equality and inclusion, (3) democratic participation, (4) active learning, (5) making a difference, (6) collective action, and (7) personal integrity (International Collaboration for Participatory Health Research (ICPHR), 2013b). In chapter 5, these ethical values and principles are explained in more detail. Since these ethical principles are closely related to validity criteria, we will only describe the actions we undertook to do justice to these ethical principles, which have not been described previously. First, the citizens and professionals involved were informed beforehand – verbally or in writing – about the goal of the study, the voluntary basis of participation, and that they could withdraw at any time without giving a reason. Verbal or written consent was obtained for the recording of the interviews. We used consent forms in simplified language that included pictures. We used Google Translate when needed to make sure that everyone understood the goal of the study and the procedure that would be implemented.

Second, to ensure privacy, pseudonyms were used in all documentation and the participants’ coded names were saved on a secure drive.

3

48 | Chapter 3

Third, the recorded interviews were summarized and a member check was done after each interview to ensure that the interpretations of the interviews were sound (Gray, 2014). The co-creation sessions and focus groups with citizens with a low SEP were not audio-recorded since suggesting this could have created distrust or limited the discussion. Instead, extensive field notes were made. Collaboratively, we came to the understanding that lack of trust in public authorities (such as the City of Amsterdam) and having limited skills in the Dutch language hindered some citizens from talking freely and honestly during the participatory observations. Therefore, we started to write and draw our observations on a tablecloth rather than making private notes. These public notes increased trust and reflectivity. They helped to create trust because everyone could get an overview of what was being said and learned during the meeting. In addition, breaking ‘the rule’ of not writing on a tablecloth helped to create a relaxed atmosphere. It also increased reflection and thereby increased validity, since we could use the words and drawings to check whether we understood the participants correctly. Fourth, I reflected on my role as an action researcher. I define myself as an ‘outsider’ (Waterman et al., 2001) because I have no formal role in the study setting other than doing research.

I knew that my ‘outsider’ relationship and my own status and background (young woman, middle class, white, Dutch) would undoubtedly somehow influence the data collection and analysis. To be reflective about positionality, I discussed the interim findings with the research team, which included senior researchers with extensive experience in participatory research and/or research with citizens living in vulnerable circumstances, junior researchers, one of whom is an Arabic-speaking bilingual researcher, an expert in the field of social welfare who had worked as a volunteer in her spare time for more than 10 years, and several interns (Shenton, 2004). Keeping a diary supported the discussion of (interim) findings. Last, throughout this research, we critically reflected on whether the outcomes were just or, in other words, equitable. For example, solutions and actions were not only sought in relation to the already complex daily reality of citizens with low digital skills, but also in relation to the design of technology. To make sure there was a fair distribution of benefits and burdens in the studies, the question ‘What is it in for them?’ was always central, especially, in meetings with citizens living in vulnerable circumstances.

The meetings with these citizens (participatory observations, co-creation sessions, focus groups) were designed in such a way that there was room for (digital) questions and opportunities for them to improve their (digital) skills. Citizens who took part and/or the community centre that hosted the meetings received a token representing our appreciation of their involvement (i.e. gift card worth between 7.50 euros and 20 euros).

Research design | 49

3.6 | Outline of the thesis In the following five chapters (chapters 4–8), I will discuss the results of the studies described above. Chapter 4 starts from the perspectives, experiences, and needs of women living in disadvantaged neighbourhoods in Amsterdam. In chapter 5, I reflect on the challenges of the complex reality of PAR for local policy that addresses digital inequalities. Chapter 6 focuses on the use of vlogs as a way to facilitate shared learning between citizens, peers, researchers, and policymakers. In chapter 7, I zoom in on how other scholars support the engagement of citizens living in vulnerable circumstances. This thesis ends with the perspective of technology developers, health-care professionals, and policymakers who are involved in the implementation of patient portals. In that chapter I reflect on how digital inequality is perpetuated in the design of patient portals. In chapter 9, I discuss the findings of our work and connect these to the broader digital divide debate. Also, I reflect on the validity of the studies as well as possible future steps.

3

52 | Chapter 4

Abstract The introduction of information and communication technology (ICT) has been shown to play a role in reinforcing existing social inequalities. This study aims to gain insights into the perspectives, experiences and needs of disadvantaged groups with respect to ICT. In the Netherlands, a relatively large and important group of non-users are mothers with a low socioeconomic position (SEP). Little attention has been paid to understanding why mothers with a low SEP scarcely use ICTs. In an explorative participatory study in Amsterdam, using observations, interviews and focus groups, our results show that reinforcing factors related to poverty, motherhood, the complexity of ICT and being first-generation immigrant influence access. The mothers’ needs are multifaceted and go beyond the simple distribution of ICT devices. ICT education should be tailored and integrated into existing social services and the daily lives of the mothers. Moreover, policymakers should pay attention to adapting the online information services to the mothers’ strengths.

‘Just having a computer doesn’t make sense’ | 53

4.1 | The digital divide in welfare states People increasingly depend on ICT to search for information and entertainment and to get in touch with health providers, government agencies, schools and social networks. ICT refers to electronic or digital devices, that is, smartphones, laptops, computers or tablets, that facilitate the transfer of information needed to participate in society. Although ICT enables many people to lead more productive and rewarding lives, this is not the case for everyone (Helsper & Van Deursen, 2017). The number of ICT non-users, even in advanced welfare states, is still high. In the Netherlands, 17% of citizens lack or have low ICT skills and another 6% have never used the Internet (CBS StatLine, n.d.). Differences in access to, and use of, ICT has resulted in a so-called ‘digital divide’. The concept seems to refer to a binary problem, but in real-life the divide is complex, multifaceted and dynamic, and often related to already existing social inequalities (Van Deursen & Helsper, 2015). This study faces the challenges of the digital divide by focusing on the perspectives, experiences and needs of socially disadvantaged citizens with respect to ICT.

Interrelatedness of social and digital inequalities In the literature, three levels of the digital divide are distinguished. The first level is a simple distinction between those who have material access and those who do not. This level of the divide is decreasing; in the last decade, an increasing proportion of citizens in welfare states gained access to ICT, whether at home, work, school, library or community centre. The second-level digital divide emphasizes that access to technology does not automatically lead to the use of ICT. Research into the second-level digital divide also explains the unequal use of ICT as a consequence of underlying social inequalities (Helsper & Reisdorf, 2017; Helsper & Van Deursen, 2017; Van Deursen & Van Dijk, 2015). A person’s income, education, literacy level and gender are significantly (directly or indirectly) associated with material access, ICT skills and usage diversity (Morey, 2007; Van Deursen & Van Dijk, 2015). The third level approaches the digital divide more comprehensively by looking not only at the causes but also at the consequences of ICT (e.g. Bruno et al., 2010; Helsper, 2012; Schreeder et al., 2017; Van Deursen & Helsper, 2015). It highlights the concern that the introduction of ICT produces ‘winners’ and ‘losers’ (Bruno et al., 2010).

Research shows that the greater an individual’s offline resources, the more ICT leads to beneficial economic, social and educational outcomes (Van Deursen & Helsper, 2015). In other words, rather than ICT leading to social inclusion, it is the other way around; ICT plays a major role in reinforcing existing social inequalities. This is also confirmed by insights from feminist studies of science, technology and society (STS) (Haraway, 1991; Oudshoorn et al., 2016; Wajcman, 2010).

4

54 | Chapter 4

Figure 4.1: Percentages of non-Internet users in 2018 (CBS Stateline)

Feminist STS analyses the mutual shaping relationship between technologies and existing gender inequalities. Offline gender division of household task and childcare responsibilities shape women’s Internet use (Helsper, 2010, p. e.g.; Schwanen et al., 2014). Moreover, feminist STS gives insight into the masculine engineering of ICT and how gender is inscribed into ICT and their intended use (Haraway, 1991; Oudshoorn et al., 2016; Wajcman, 2010). Scholars theorizing about intersections add the importance of interrelatedness of socio-demographics such as race, class, age, sexuality and disability in understanding how technologies are constructed, experienced and used (Crenshaw, 1989). Inequalities concerning ICT use are not solely related to singular socio-demographics but should be approached entangled (Crenshaw, 1989; Leurs & Ponzanesi, 2014; Oudshoorn et al., 2016). Therefore, the intersectionality approach highlights the importance of a better understanding of the everyday Internet use in the local context to inform micropolitics concerning social and digital inequalities (Kennedy, 2005; Leurs & Ponzanesi, 2014). So far, however, most research into the digital divide offers snapshots of how social and digital exclusion are related (Helsper et al., 2015; Helsper & Reisdorf, 2017; Van Deursen & Van Dijk, 2015).

The primarily quantitative research is helpful in mapping the landscape of access and skills, and the way social, economic and digital inequality reinforce each other. However, it is less helpful for understanding the interrelatedness of social and digital inequalities. There is empirical research that explores the issue of digital inclusion from the perspective of the elderly (Lüders & Brandtzæg, 2017, p. e.g.;

‘Just having a computer doesn’t make sense’ | 55

Quan-Haase et al., 2016). To a lesser extent, literature also reports on the perspectives of immigrants (Alam & Imran, 2015; Guttman et al., 2018; Hallbeg et al., 2016) and low-income communities (Powell, 2010). Those studies are mostly focused on the barriers to ICT use such as affordability (Alam & Imran, 2015; Powell, 2010), language (Alam & Imran, 2015; Guttman et al., 2018), ICT illiteracy (Alam & Imran, 2015), lack of familiarity (Hallbeg et al., 2016), lack of time (Hallbeg et al., 2016) and cultural differences (Guttman et al., 2018). Very little is known about how already disadvantaged groups, in terms of gender, class or race, experience the digitalizing society, and use ICT in daily life, and what they need to participate in our digital society. This study, therefore, aims to gain insight into the perspectives, experiences and needs of socially disadvantaged groups with respect to ICT to inform micro-politics concerning digital inequalities. This study addresses the issue of the digital divide from the perspectives of mothers with a low SEP, since this is a large group of non-Internet users in the Netherlands. As shown in Figure 4.1, women, non-Western immigrants and citizens with a low SEP are more often offline than other groups.

SEP is a multidimensional construct referring to income, educational level and position in the job market (Dutch National Institute for Public Health and the Environment (RIVM), n.d.). In the Netherlands, citizens with a low SEP have more often a non-Western than a Western or Dutch background8 (CBS StatLine, n.d.). This study focuses specifically on women because of the long-existing offline gender inequalities that affect Internet use (Helsper, 2010; Schwanen et al., 2014; Wajcman, 2010). Moreover, the focus on mothers is highly relevant because they play an essential role in supporting their children’s ICT use; limited or no computer knowledge of parents creates a unique challenge for children in a digitalized society (Tripp, 2010). If mothers can support their children in ICT use, this will have important implications for breaking the cycle of inter-generational poverty (Eynon & Geniets, 2016; Hallbeg et al., 2016).

Analytical framework Following the intersectional approach, this study avoids generalizing or homogenizing conclusions about gender, SEP and technology (Crenshaw, 1989). This study will, by means of an explorative participatory study, look into the perspectives, experiences and needs of mothers with a low SEP. To study the perspectives, experiences and needs of mothers, we use the well-known multiple access model of Van Dijk (Van Dijk, 2005), 8 Citizens with Western or Non-Western background have at least one parent who was born outside the Netherlands. A Western background refers to Europe (excluding Turkey), North America, Oceania, Indonesia or Japan. Non-Western background refers to Asia (excluding Indonesia and Japan), Latin America, Africa and Turkey. This division is based on the social- economic differences. Indonesia is seen as a Western country because of the high percentage of originally Dutch re-immigrants (CBS StatLine, n.d.)

4

56 | Chapter 4

since this model explains the total process of appropriation of ICT, which is a product of several factors such as motivation, material, skills and usage access. Motivational access. Motivational access is the condition for the other three kinds of access and concerns the appreciation of the technology and intention to purchase and use ICT. Emotional reasons, for example, anxiety about using computers, as well as more rational reasons, for example, malfunctioning of digital technology, can discourage people from purchasing or using ICT (Van Dijk, 2005). Material access. Having physical access to ICT and the required software is a necessary condition to acquire skills and use ICT. Someone can have physical access at home, work, school, public places or with family or friends. Material access is related to ongoing economic costs such as payment of the monthly Internet bill and replacement of cheap, second-hand, and frequently broken hardware (Gonzales, 2016). Skills access. The changing and complex digital world requires continuously new and different skills. Internationally, there is no consensus about the definition of digital skills. However, for full participation in the digital society, it is important to incorporate the full range of skills: from the technical to the more social communication and critical information-navigation skills. Usage access.

Quantitative studies repeatedly show that socioeconomic status influences for what purposes ICT is used. For example, poorly educated citizens tend to use ICT more often for games and entertainment, while highly educated citizens more frequently use the computer for capital-enhancing activities like communication, news and information (e.g. Livingstone & Helsper, 2007; Van Deursen & Van Dijk, 2013).

4.2 | Methodology This explorative participatory study was conducted in Amsterdam – as with over 850,000 citizens it is the most populous city in the Netherlands. The city has a fairly large group of people who risk falling behind in the digital society; 25% have a low level of education, 22.1% live in poverty and 35% have a non-Western background (Onderzoek Informatie en Statistiek[OIS], 2018). We conducted this research especially in socioeconomically deprived areas of Amsterdam: Amsterdam East, North, West and New-West, where those numbers are even higher. This study was conducted in the context of a larger study on

‘Just having a computer doesn’t make sense’ | 57

the digital divide in Amsterdam in which mothers, fathers, children, and volunteers and professionals who work closely with families with a low SEP were involved. Formal and informal gatekeepers of language schools and community centres assisted in reaching potential participants. In total, we involved 66 socioeconomically disadvantaged mothers living in New-West (n = 31), Amsterdam West (n=19), East (n=11), and North (n=5). The participants were between 20 and 60 years old. Nearly all mothers were married and the number of children varied between one and seven. The age of the children ranged from neonate to 20 years. Most mothers were unemployed and a few were volunteering in a community centre. Most mothers had a non-Western background (92%), of whom 27% were from Morocco and 13% from Turkey. The non-Western mothers were mostly first-generation immigrants, and they often had poor Dutch language and ICT skills. Only 8% had a Dutch background. While we know that mothers with a Dutch background probably experience different barriers in ICT use (Alam & Imran, 2015; Guttman et al., 2018), we did not exclude these women since this study focused on mothers with a low SEP.

Study design This explorative participatory study, based on qualitative data gathered through participant observations (POs), semi-structured interviews (SSIs) and focus group discussions (FGDs), reports on the perspectives of mothers with a low SEP. A participatory approach was chosen because by fostering dialogue between stakeholders and explicitly including the voices of those who are often underrepresented or have less power or possibilities to make themselves heard, this approach enlarges the social relevance (Abma et al., 2017). Data collection took place from April to October 2017 and comprised two phases. In the first phase, seven language courses and two digital skills courses were observed (n=38, duration 1–3 hours), and seven SSIs were conducted (duration 20–70 minutes). The aim of the POs was to gain more insight into the complex daily reality of mothers and how mothers act in our digitalized society. The observed meetings were not especially initiated for this study and therefore the researchers were sympathetic to the main goal of the activity or event while gathering information about the mothers’ lives, as well as their views on today’s digital society. However, in all cases, the topic of the digital society was discussed for at least 30 minutes.

The observations list of the researchers was based on the four types of access as described by Van Dijk (Van Dijk, 2005). In all cases, extensive field notes were written immediately afterwards. Also, the observations were written or drawn on a tablecloth together with the mothers (Figure 4.2). This creative method was introduced after PO6 because we observed and obtained feedback that a verbal discussion does not result in deep insights due to language barriers and trust issues. The introduced method is more inclusive to people with low levels of literacy. The visualizations reduce

4

58 | Chapter 4

Figure 4.2: Examples of the use of a tablecloth during the POs and FGDs.

the language barrier and making a shared product creates an avenue for the researcher to build trust and to show an understanding of what is considered important to the mothers. Moreover, using drawings and easy words on a tablecloth increases the reflectivity of the mothers and increases validity since the researcher can check whether the mothers have been well understood. The aim of the SSIs was to gain a deeper under- standing of the needs of the mothers in our digitalized society. To ensure validity, the questions of the SSIs were concentrated on the four kinds of access described by Van Dijk (Van Dijk, 2005) as well as the research objective (Table 4.1). When consent was obtained, interviews were audio-taped and transcribed verbatim. Four out of seven interviews were audio-recorded. During the other three interviews, for which no consent was obtained, extensive fields notes were written. Low language level was the main reason for mothers to not give consent for recording. In the second phase, four FGDs (n=21, duration 1–3 hours) were organized. The main aim was to gain insight into the ideas and advices of mothers with a low SEP to increase Table 4.1: Interview prompts Research topics

Example question(s)

Motivation acess

What is your main motivation to use (or not-use) ICT?

Material access

Where do you use ICT most often? Do you have home access to a computer/laptop/tablet?

Skills access

Do you experience difficulties using a computer/mobile phone/tablet? If so, what are these difficulties?

Usage access

How often do you use ICT? For what purpose do you use ICT? Do you use ICT for communication/find information about your health/find information on governmental websites/etc.

Barriers

Which barriers hamper you in ICT use?

Needs

What or who can help you to use the computer more often?

‘Just having a computer doesn’t make sense’ | 59

meaningful ICT use. In order to work in co-creation towards opportunities that can increase ICT use, three topics were discussed (Table 4.2). To empower all mothers, with varying Dutch language skills in a verbal discussion, the activities include pictures and visualizations. The moderator made sure that each mother had the opportunity to share her opinion and, if needed, clarified what was said. The topics were determined on the basis of the data analysis from the first phase, POs and SSIs, and complemented with insights from professionals (e.g. debt counsellors, and language, ICT and high school teachers) and volunteers (e.g. language and ICT teachers and community service workers) who are closely working with citizens with a low SEP. During the FGDs, field notes were written on the tablecloth, an A3 poster, and in the researchers’ notebook. Immediately after the sessions, the field notes were complemented, clarified and summarized in one document.

Data analysis Full transcripts and field notes were analysed through ethnographic content analysis using the software programme Atlas.ti. One researcher (NG) coded the data. From the start in weekly meetings with the project team, comprising the authors (NG, RK and CD) and three policymakers of the City of Amsterdam, interim findings and coding were discussed. These discussions were used as input in the next data-collection sessions to deepen or validate the findings. With respect to the digital divide, we distinguish four themes based on the multiple access model of Van Dijk (2005): motivational, material, skills and usage access. Moreover, one additional theme was included: needs and ideas. In the final stage of this research, the project team agreed upon the coding scheme and findings. Data collection and analysis was done in Dutch. To prevent loss of meaning, the codes and quotes were translated into English only in the final phase of preparing this manuscript. Table 4.2: FGDs activities and prompts Research topic

Activity

Example question(s)

Where do you want Drawing ideal ‘learn locations’ on a tableto learn more about cloth and group discussion ICT?

What do you need to learn more about ICT? Why do you prefer to learn alone or with peers?

Where can your child learn more about ICT?

What is the most ideal place for your child to learn more about ICT? Which activities is your child allowed to do during an ICT class?

Group discussion using A3 poster with pictures covering three topics: teachers, place/location, activities.

Can vlogs help you Vlog making in or around the community How can vlogs help you to learn to learn more about centre. In three out of four FGDs a short more about ICT? ICT? video was made. Topic of the vlog was chosen by the mothers.

4

60 | Chapter 4

Ethics This study adhered to the Code of Ethics for the Social and Behavioural Sciences involving human participants as accepted by the deans of Social Sciences in the Netherlands, January 2016. The study did not need formal approval of a Medical Ethical Committee according to the Dutch Act on Medical Research Involving Human Participants (Wet medischwetenschappelijk onderzoek met mensen [WMO]). The study aimed to respect the privacy, human dignity and fair distribution of benefits and burden. Mothers were informed beforehand about the goal of the study, the voluntary basis of participation and that they could withdraw at any moment without reason. To develop a relationship of trust and establish high levels of rapport, the researchers took time to get to know the mothers. To guarantee privacy and anonymity, participants’ names were coded and fictitious names are used.