The Methodological Approach of the Long-Term Study

This chapter outlines our study’s specific methodological approach as derived from its theoretical foundation. We deal with the questions of collecting, processing and analysing data from a qualitative longitudinal study over 12 years from 2005 to 2017 but also with questions of recruitment and panel maintenance, as well as ethical issues. This chapter, therefore, focuses on the ways in which the methodological design was revised and judiciously complemented, in order to grasp the complexity of the topic. Following the logic of triangulation, we developed and implemented a rich design, where all the components draw on, complement, and monitor each other in the processes of data collection and analysis. This means that we have to discuss the importance of transparency and comprehensibility in terms of intersubjectivity, paying special attention to the qualitative and longitudinal character of the study.


IntroductIon
Lasting from 2005 to 2017, this study's methodological approach to the role of media in the socialisation of socially disadvantaged children and adolescents in Austria illustrates how the complex praxeological approach to researching children's and adolescent's media usage within the process of socialisation can be operationalised. The approach we selected for dealing with this challenge involved a qualitative longitudinal panel and was conducted successfully in six waves of data collection over a period of twelve years. The starting point was the question: How to operationalise the theoretical approach presented in the preceding chapter, in order to provide a fruitful basis for analysing the role of media within socialisation on a methodological level? As growing up in a mediatised world is a complex process, it is abundantly clear that looking merely superficially at how children deal with media will not suffice. Instead, it is essential to investigate how they subjectively make sense of media as a source for coping with factors of which they are not yet aware. The question: "How do children, and also their parents, make sense of media?" is closely connected with the structures of their everyday lives on the micro-level of the child or adolescent, on the meso-level of the family, but also of peers and friends, as well as on the macro-level of the country/society (see Fig. 4.1). Their media-related practices may be understood as subjectively given answers to the challenges of their everyday CHAPTER 4 The Methodological Approach of the Long-Term Study lives. Against this background, we consider media usage a practice within a socially constructed everyday life and thus a form of observable practical ability.
To sum up, the approach implies the following four methodological requirements: first and foremost, since the main analytical concepts refer to entangled and interlinked processes of interaction between children and their parents within their everyday lives, the methodological approach has to be located within a qualitative research paradigm (Clarke, 2005(Clarke, , 2011Flick, 2014;Wilson, 1970). Researching within this paradigm allows us to retrace and reconstruct how individuals subjectively make sense of their social contexts and act in them. Furthermore, it allows us to detect the underlying structures of individuals' utterances and actions. Secondly, in order to be able to reconstruct and to review the processes of doing family, data have to be collected from children as well as from parents. Thirdly, in order to include the broader social contexts that have an impact on a family, particular attention has to be paid to its socio-structural conditions. And fourthly, in

recruItment of the famIlIes
We conducted a panel study of twenty (reduced to 18 after the second wave due to drop-outs) socially disadvantaged families with children (boys and girls), who were, respectively, almost five or six years old in 2005 (see Paus -Hasebrink & Bichler, 2008, pp. 132-141;Paus-Hasebrink & Kulterer, 2014;Paus-Hasebrink, Sinner, Kulterer, & Oberlinner, 2017, pp. 48-49). They were not yet in school and still attending kindergarten. So, from the preparatory theoretical work, we deduced the following selection criteria as being central to researching social inequality and then recruited the 20 families accordingly.
• Apparent markers of social disadvantage following Hradil (1987Hradil ( , 1999: for example, unemployment, low income (defined as relatively poor, less than 50% of the national median income, and as at risk of poverty, less than 60% of the national median income), lower formal education, residential neighbourhood, bad housing conditions-singly or in combination. • Family structure: families were so chosen that the sample would contain different family configurations, like single-parent families, large families (more than five children) and nuclear families. • Migrant background: households with at least one foreign member (understood as a non EU-or EFTA-citizen) belong to the groups at risk of poverty or being marginalised (see Statistik Austria, 2017, pp. 21, 25). • Area of living: urban and rural areas/areas with a poor infrastructure (for example, mountain areas, areas with bad public transport connections, no accessible railway stations) (see Table 4.1).
It was not our aim to depict statistical distribution throughout the population but to select typical and "information-rich cases" (Patton, 2002, p. 230, emphasis in original) in our field of research, a procedure that is called "purposeful sampling" (Rapley, 2014, p. 56): "Informationrich cases are those from which one can learn a great deal about issues of central importance to the purpose of inquiry, thus the term purposeful sampling" (Patton, 2002, p. 230, emphasis in original). Based on a literature review (Paus-Hasebrink & Bichler, 2008, pp. 135-136) conducted to acquire "knowledge of the phenomenon" (Rapley, 2014, p. 50), we selected markers that define the characteristic living conditions of socially disadvantaged families, in order to make a "thoughtful and rigorous" (Rapley, 2014, p. 49) choice. Lack of money is the most important marker of social disadvantage. On the one hand, family income depends on the specific living conditions of a family's wage earners, on the other hand, the availability of material resources defines the living conditions of a family as a whole. In order to include families in our sample, we had to be able to easily survey suitable markers, but with the provision that they would be capable of representing a wide range of socially disadvantaged families. Accordingly, for our sampling we decided to consider those factors that increase the risk of poverty (see Chapter 2): low income, unemployment, lower formal education of the parents, single-parent families, large families and families with migrant background. In addition, we decided to select children between five and six years old, who were not yet attending school, in order to capture the important transition period from kindergarten to primary school in our second wave of surveying our panel. We decided in advance on a total of 20 cases and based our decision on the qualitative orientation of the study, the complex methodological approach and the available resources (temporal, personal and financial). Both the distribution of families across rural and urban areas and the distribution between the sexes were to be equal. For practical reasons, we recruited families in the Federal States of Salzburg and Upper Austria. The process of recruitment was supported by a number of social organisations and social facilities like the Organisation of Single Mothers and Single Fathers, Caritas, the Youth Welfare Office and numerous kindergartens, as well as by public authorities like the Department of Families and the relevant ministers from the two states. We informed potential participants about the project, invited them to contact the principal investigator and offered an expense allowance (50€ per data collection) to the families, an amount which remained unchanged until 2016. The children, and later the adolescents, received small non-monetary gifts (for example, sweets, pens or handicraft kits). However, socially disadvantaged people do have a record of rejecting government institutions and participation in research projects, so that the process of recruitment was quite hard and time-consuming (see also Patton, 2002, pp. 310-311). In order to achieve our goals, we were forced to take further action: we started to cooperate with additional social organisations and social facilities, used personal contacts and political support, and we placed advertisements in local and regional newspapers and journals. Table 4.1 provides an overview of the sample's main characteristics at the beginning of the study (see Chapter 5 for a description of the families). Due to the fact that low income in the beginning was a major selection criterion to become a part of the panel this criterion is not listed in the table-at that time all families had a low income.
The focus of the research was on one child at the right age from each family. However, the siblings figured in the interviews, too. Since two families did not participate throughout, by the end of the project the sample consisted of ten boys and eight girls. Over the study period, the families' changing circumstances altered other characteristics as well, so the sample looked different at the end of the project, a point extensively discussed below (see Chapter 8). The aim of choosing children between the ages of five and six was to be able to investigate the relevant phases of a child's development from kindergarten, through mid-childhood to youth over six waves of data collection (2005, 2007, 2010, 2012, 2014, and 2016). To complete it, we conducted a telephone survey in winter 2016/2017, in order to get the latest information about the development of the adolescents and their families.

the challenges of managIng a long-term study
The whole project consists of three separate phases (waves 1 and 2, waves 3 and 4, waves 5 and 6 and call-back interviews). All three project phases had to be applied for singly. Not only did they have to pass the blind peer-review process, but they also had to fulfil the requirements and restrictions of the OeNB Anniversary Fund (2015) (Fund of the Austrian National Bank). In addition, the proposals needed the approval of Department of Communications of the University of Salzburg, the approval of the Research Service of the University of Salzburg (2018) and the approval of the Rectorate of the University of Salzburg. Due to the limited time frames, we had to plan the three project waves separately, not as a single twelve-year project. However, we did manage to acquire three consecutive grants, but there was always the risk of terminating the study after its first or second phase. Financing over a period of years is one of the most challenging tasks in conducting long-term research, due to reduced overall funding for research, but also given the existing structures (in Austria) for research funding. As a consequence, we had to work hard on maintaining our panel. On the one hand, we had to keep contact-data up to date, as our clients moved house and changed telephone numbers, email addresses and other contact details quite often. Furthermore, they often neglected official letters, sometimes not even opening them. Consequently, we had to visit some of the families over the time, just in order to stay in touch and to re-recruit them for our next wave of data collection. On the other hand, we had to deal with the uncertainties over the project funding. Every time we had to invite the families to participate again in future, even though we could not guarantee if, and when, the next wave of data collection would take place. This required great commitment from both sides, families and researchers, but in the end, we were able to achieve a very low drop-out quota.

data collectIon
In order to reconstruct, describe and explain how children make sense of media against the background of their everyday lives, we have applied a broad repertoire of qualitative methods (Morse & Maddox, 2014, pp. 524-525; see also Paus-Hasebrink et al., 2017, pp. 51-58;see Figs. 4.2 and 4.3). However, we have gone one step further, using a rich design which is characterised "as one that is not restricted to one theory and method, or one set of categories or instruments, but which embraces diverse and multiple perspectives brought together with coherence and harmony. It is more than a multi-method design per se" (Paus- Hasebrink, Prochazka, & Sinner, 2013, p. 23): we conducted guided in-depth interviews (semi-structured) with the children. Commencing with the fifth wave (beginning of adolescence, 2014), we added three more methods of data collection: (1) Thinking aloud about a social networking tool (for example Facebook or WhatsApp) selected by each child, (2) In addition, we asked them to produce their own network maps of media and relationship structures, and (3) to take a set of photos (favourite spot, place of work, preferred spot for media usage). We also conducted guided in-depth interviews (semi-structured) with the parents or one single parent. In addition, we asked them to independently complete a standardised questionnaire on their living conditions. In order to get an overall impression of the respective families, we used observation protocols to provide a report on the visit by the interviewers.
These components of our rich design draw on, complement, and monitor each other during data collection and analysis, a process which simultaneously makes the research transparent and intersubjectively traceable: "The goal of such rich design approaches may be to eliminate weaknesses and blind spots perceived in one method by using complementary approaches that have specific strengths in such areas. A sensitive combination of methods can therefore shed light on aspects that cannot be covered adequately by only one method. Such rich designs feature a high density of data and a high level of reflection on the research process itself" (Paus- Hasebrink et al., 2013, p. 23). This is in accordance with Denzin's (1989, p. 307) concept of triangulation: "By combining multiple observers, theories, methods, and data sources, [researchers] can hope to overcome the intrinsic bias that comes from single-methods, single-observer, and single-theory studies" (see also Denzin & Lincoln,Fig. 4.3 Methods of data collection 2011, p. 2). The capacity for "comparing and cross-checking the consistency of information" (Patton, 2002, p. 559) allows researchers to generate greater added value.
The strategies of research were guided by the analytical concepts presented in Chapter 3: options for action, outlines for action and competences for action. In order to operationalise these concepts, we used several reactive and non-reactive methods. The standardised questionnaire for the parents was used in order to examine the options for actions. Observational methods served as an additional tool to investigate how the child and the parents conduct their everyday life and how they deal with the conditions of their specific social situation in doing family (for example, conflicts, proximity and so on). In order to define the outlines for action, we combined these questionnaires and the named observations with guided in-depth interviews conducted separately with the child and the parents-mostly the mother, in some cases both parentson different aspects of their everyday lives and their usage of media. In order to gather evidence regarding competences for action, we used indications from interviews and observations of a wide range of aspects, for instance, a child's cognitive and motivational resources, their attitudes towards kindergarten, as well as school and education in general, parents' strategies for bringing up their children, together with their ability to implement them, and for mediating their media usage, and both children's and parents' media-related skills. All these aspects were investigated from both the children's and the parents' perspective. This made it possible to discover discrepancies or commonalities in how children and parents mutually perceive their competences. Differences between the answers of parents and children suggested conclusions regarding family climate and doing family (relationships between parents and children and within the family as a whole). Furthermore, this approach allowed conclusions concerning the applied mediation strategies and their acceptance, as well as their relevance for the process of socialisation. In addition, it became possible to track the process of growing up and the process of distancing from parents as the children became older and subsequently adolescent (at the beginning of puberty and youth) (see Paus-Hasebrink et al., 2017, p. 51). All the data collected from these different methods were used to create and to update a global characteristic (Charlton & Neumann, 1986;Schütze, 1977; see also Paus-Hasebrink & Bichler, 2008, pp. 142-145;Paus-Hasebrink & Kulterer, 2014, pp. 57-59;Paus-Hasebrink et al., 2017, p. 51) for each family investigated.

Standardised Questionnaire
At each visit, the parents were asked to complete a standardised questionnaire on their living conditions. These questionnaires covered central aspects about living together, the constellation within individual families, housing and socio-economic circumstances. It consisted of the following main categories: • family constellation and family characteristics: marital status, number of family members living together, number of children, overall and in residence, ages of children and parents, nationality, religious affiliation; • professional career and activity of the parents: school education and apprenticeship, (un)employment and self-employment, training courses; • economic situation of the family: household income (single earner or not), social benefits, further sources of income (for example, rental income, interest, inheritances), assets and/or estates, amount of pocket money (children), later on also, income of the adolescents; • housing situation: residential area and neighbourhood, size and state of flat or house, ownership or renting a flat or house, duration of residence, plans for relocation.

Guided In-Depth Interview with the Parents
Since its inception, we tried to keep changes or modifications to the guideline at a minimum, in order to sustain a maximum of comparability throughout the investigation period. The guidelines for both parents and children were closely coordinated. However, we questioned the parents more intensively about their children than the other way round.
Our interviews included questions about how, in the first instance, the parents-but then, to a lesser extent, their children and other family members-dealt with and experienced possible challenges, such as unemployment or poorly-paid jobs, shortage of money, bad housing conditions, living in a single parent family and/or in the context of a large family, and so on. We also asked them how they interacted with their family and if they felt integrated in it, in their neighbourhood, peer-groups, friendships, kindergarten or class, and social life in general, and what kinds of media offerings they used, to what extent, for what purposes, and with whom. Additionally, we asked children and parents in each wave of data collection for their ideas, preferences, goals, plans and motives for action and what they were planning for their future. However, in the interviews with parents some topics were examined more intensively than they were with the children. There were, for example, questions about their mediation strategies, questions not only about the child of particular interest but also about the siblings, and questions concerning developments in different aspects of individual families' lives.
In order to provide an overview, the guidelines for parents followed four main themes: • how the family felt situated socially-the everyday life of the familythe participation of parents and children within society, that is, the social conditions of growing up, the changes that have happened in living conditions, the climate and relationships within the family, daily routines and aspects of everyday life, the preferred leisure activities of parents, and of children, their shared interests, social activities and involvement in associations, cultural life and active citizenship; • their attitudes towards the media-the media usage of parents, child and siblings-the behaviour within the family as regards the media, meaning the importance of different media for the parents, the individual patterns of media usage of the family members, the relevance of joint media practices, the sources of information, the expertise within the family when dealing with the media; • media repertoires: the media ownership of the family and of single members of the family, the financing of access to the media, the media usage of parents and children, the individual media repertoires of the family members, the importance of media for the children in general/in their everyday lives, the roles and functions of the media for the children, but also for the parents, the modes of media usage (for example, where, when, how long, with whom, on which device), the existence of mediation strategies, the importance of media education for parents, the regulation of media usage by the parents; • the communication and transferring of values-the extent, importance and credibility of the family-the extent, importance and credibility of other contexts of socialisation (for example, media, kindergarten, school, friends, peers) and the changes over time, the sources of knowledge transfer, the sources of value transfer, role models and idols, the importance of the family along with other players, the changes due to the development of the child.

Guided In-Depth Interview with the Children and Adolescents
In order to sustain maximum comparability, changes were reduced to a minimum, with the main themes remaining the same over time.
Although the main aim of the interviews with the children was to investigate their media usage within a mediatised world, we also asked them about the circumstances of their lives (for example, family life, kindergarten and school, friends and peers, financial worries, housing, their own room, wishes and dreams and so on). Special attention was paid to the growing importance of online media and internet-based (communication) services. We did not enquire about the usage of single devices, but rather more about individual media repertoires, based on functions and interests (see Hasebrink, 2014;Hasebrink & Domeyer, 2012;Hasebrink & Popp, 2006). Thus, some aspects of our guidelines were restructured over time. This meant specifically adding new questions, removing others, or rewording some, in order to cater for new technological developments. In the interests of a pleasant and more relaxed atmosphere, boys were always interviewed by male members of the team and girls by female members. Whenever possible, we tried to send the same interview teams to the same families over several years and waves of investigation. On the one hand, the interviewer gained special knowledge about the respective family and its members, on the other hand, this approach created a particular relationship of trust. For the same reasons, we tried to send more experienced interviewers to more challenging families (see Chapter 8).
• The situation of the child's life in society-the everyday life of the family-the participation of the child within societyleisure activities-the social conditions of growing up and living conditions, the climate and relationships within families, the daily routines and aspects of everyday life, the preferred leisure activities of the children, their social activities, involvement in associations, cultural life and active citizenship; • Media ownership-their media repertoires-their media usage and media behaviour-the importance of media-the roles and functions of media for the child-the structure of media repertoires, the accessibility of media for the children, the personal ownership of media, the media used by the children and the role of media in the children's everyday lives; • The behaviour of parents and siblings as regards the media-their attitudes towards media-the media usage within the family-media education and the regulation of media usage-the role of different media for individual members of the family, the behaviour of individual family members as regards the media, the media practices shared within the family, the mediation strategies, media education and the media expertise within the family; • The ways of adopting values-the role of the family concerning communication and transfer of knowledge and values-the role of other contexts of socialisation (for example, media, kindergarten, school, friends, peers)-the contexts in which children adopt values and codes of practice-the role models and idols for children.
In contrast to the guidelines for parents, the guidelines for children contained so-called role-playing questions, in order to better understand the importance of media and media personalities for the children. Different layers of media processing-cognitive, emotional and socialformed our analytic distinctions. However, in real life they are closely linked and intertwined. Therefore, it is quite difficult for the adolescents to articulate themselves concerning these fields (Cakici & Bayir, 2012, pp. 1076-1077Cohen, Manion, & Morrison, 2011, pp. 513-522;Jörg, 1994;Paus-Haase, 1998, p. 104;Stahlke, 2005, pp. 496-507;Tilemann, 2017, p. 393). To answer role-playing questions may be helpful in expressing complex contexts of experience and subjective understanding (Paus-Haase, 1998, p. 165;Sader, 1995, p. 194). We used three different types of role-playing questions. In an "island-question", the adolescents were asked to explain which persons and which objects they would take with themselves to a desert island and why. In the "100 Euro-question" (later "500 Euro-question") they were asked to tell us what they would do with 100 (or 500) Euro. In the "wishing question", they were asked at the end of the interview what they would do if they had one wish. We found the answers to these questions were particularly important sources of information.

Observation Protocol
Finally, we will now outline our observation protocol (Mason, 2002, pp. 96-98). This completed our survey of the families from the interviewers' point of view by basically fulfilling two purposes: on the one hand, it systematically recorded the respective families' living and housing conditions and their media hardware. On the other hand, it served as a guideline for the participatory observation, in order to grasp how all family members were doing family. We kept the criteria of the protocol the same over time (see also • the state of the house or flat and rooms: cleanliness, neatness, furniture, differences between the children's rooms and the rest of the rest of the houses or flats; • the media hardware: devices, number of them, their positioning, media-related items, media hardware in the children's bedrooms; • the pets: species, number, location, behaviour, cleanliness; • the family members: garments, smoking, alcohol, behaviour and manner of children, parents and siblings; • any further particular features.
As criteria like cleanliness, neatness or the conditions of furniture and clothing may be viewed subjectively, it was essential to reach a common understanding. Therefore, we applied a two-step procedure: firstly, both interviewers had to fill out the protocol together, immediately after finishing the visit in the family. They were instructed to compare their personal impressions with their experiences in other families they visited. Secondly, in order to consolidate a common understanding of a given sample, the research team discussed all recent observation protocols and compared them to the data from the previous waves.

Complementary Methods for Adolescents: Thinking Aloud, Network Maps, Photos
With the fifth wave of data collection in 2014, the children we were observing entered on a new stage of life, adolescence. So, we decided to add three more qualitative methods of data collection to our already existing range. On the one hand, we wanted to do justice to our adolescents by acknowledging the changes in their everyday lives, their new interests, behavioural patterns and options. On the other hand, we had to adjust to the technological and medial changes in a mediatised world between our waves of interviews. These newly added methods, then, allowed a better understanding of the media behaviour of the adolescents and their social inclusion in family and non-family relationships. This approach was successful in revealing hidden information, not only when it was difficult for the adolescents to express their thoughts and ideas, but also when they were not fully aware of certain facts and structures (see Paus-Hasebrink, 2005, p. 224;2017, pp. 277-278;Paus-Hasebrink et al., 2017, pp. 55-56).

Thinking Aloud About Favourite Social Media Tools
With the first supplementary method, we could stay abreast of changes caused by the increasing importance of social networking tools. Based on the method of thinking aloud (Bilandzic, 2005, pp. 362-364;2017, pp. 407-408), we asked the adolescent to talk about their favourite social networking tools (for example, Facebook, WhatsApp, YouTube). They were asked to show us their profiles and settings (for example, privacy settings, friends list, photos, groups, single chats, followed pages and channels and so on) and to talk about how they make use of the applications and features (see Rose, 2016, pp. 301-302). Our aim was to better understand what is really important to the adolescents (see Marotzki, Holze, & Verständig, 2014, p. 457), so the guideline for this method contained only a few questions and topics (Toerien, 2014, pp. 330-331), but did include questions about the structures of the respective social networking tools, for example core or advanced features but also restrictions (see also Trappel, 2007, pp. 156-158). However, we likewise used the method of thinking aloud in order to reveal how competent and knowledgeable the adolescents were about the social networking tools, although they had not been willing or able to talk about them during the formal interview.

Network Maps Drawn by Our Subjects
In order to cater for the increased relevance of social relationships within the everyday lives of the adolescents (Cotterell, 2007, p. 91), we decided to integrate network maps into our repertoire of methods (Coffey, 2014, p. 367;Samuelsson, Thernlund, & Ringström, 1996, pp. 327-330) as these can visualise connections and communication paths and routines (Hepp & Düvel, 2010, p. 271). Hence, they are suitable for researching not only media repertoires but also information repertoires and relationship structures. We decided to apply a two-step approach by firstly asking the adolescents to write down the names of all relevant persons on a sheet of paper with the word "me" already printed in its centre to symbolise the position of the respective adolescents, around which they then grouped other relevant persons. The number of persons, their position in relation to the adolescents and the links between them were up to the adolescents (Crosnoe, 2000, p. 379). When they had finished, we asked them to elaborate on the spatial distances, on the choice of people (maybe on people who had been left out or forgotten and so on) and on the connections between themselves and the people nominated. In a second step, we asked the adolescents to add the media services they used and their media devices to the map. If possible, they might indicate connections between certain media (devices and services) and (groups of) persons (for example, PlayStation with friend XY, WhatsApp group for best friends, TV-series with brother and sister). All this resulted in network maps about everyday communication, the media used (devices and services) and connections among the elements cited. This method shed light on the importance of (groups of) persons and media thus selected, in order to visualise their status (see Hepp & Düvel, 2010, pp. 271-273;Hepp, Berg, & Roitsch, 2011, pp. 10-12). However, it was also very informative to note the persons and media not present, or no more than distant from the respective adolescents.

Photos by Our Subjects
Inspired by the concept of "bedroom culture" (Bovill & Livingstone, 2001, p. 2;Lincoln, 2013, p. 315), we asked the adolescents to take a set of photos in their own room, in order to better understand their personalities and their media behaviour. In practical terms, they were asked to take three pictures (see Harper, 2005, pp. 231, 233;Marshall & Rossman, 2016, pp. 184-187;Rose, 2016, pp. 310-312): a photo of their favourite spot, a photo of their place of work and a photo of their preferred spot for media usage. However, they could also indicate where one photo could cover two or even three spots (Mason, 2002, pp. 112-113). Such photos do provide deep insights into the private spheres of the adolescents (Rose, 2016, pp. 360-362). They reveal rooms where they spend a lot of time, and ones which are today strongly affected by media (see Bovill & Livingstone, 2001, p. 3). In their own realm within the house or flat they can realise their own ideas and express themselves (see Lincoln, 2013, pp. 318-320). Examples of their self-expression are individual schemes for painting walls and the positioning of media devices, but also the use of posters or merchandising products and selected furniture and how these are set out in the room.

Final Call-Back Interview
The sixth and last planned wave of surveys had already taken place in January and February 2016. Since it was our aim to bring the data up to date before completing the study, we shortened our guidelines and conducted an additional call-back interview in January 2017. To integrate the results with preceding waves, we talked to the adolescents and also to at least one of the guardians, in most cases the mothers. We kept to our protocol of girls being interviewed by a female team member and boys by a male team member. The focus in these call-back interviews was on the current living situations, life-changing experiences and our subjects' perception of the changes. However, against the backdrop of the sixth wave, questions about future prospects and developments, working life and unemployment were particularly important. At this time, some of the adolescents had to face a period of uncertainty. They had to decide about their plans for the future, whether to continue schooling or start an apprenticeship or even working life itself. Finding a (appropriate) training place was not easy for all of them, so they had to deal with unemployment and with training activities at job centres. The following topics were part of our guidelines: • current living situation; • school education or apprenticeship; • future prospects and unemployment; • family life and satisfaction; • friends, peers and (new) relationships; • favourite media; • role of the media in everyday life.
In the data processing, we processed the call-back interviews just like the guided interviews from the previous six waves of data collection. This included, in particular, transcription, anonymisation of data and subsequent analysis. However, we did not intend the call-back interviews as a substitute for a seventh wave of data collection, due to temporal, financial and personal restrictions. They were only used to record the latest developments in the topics mentioned above to wrap up before finishing the study. Compared to the sixth wave of data collection, we had to accept two adolescents dropping out: one boy was not reachable in any way, one girl missed all appointments for a telephone call.

data ProcessIng and data analysIs
The wide range of data collection methods required a common strategy to analyse the huge amount of heterogeneous data from different sources. The guided interviews and their analysis were at the heart of the process, but all other sources were taken into account at every stage of the analysis as well (Banks, 2014, pp. 394, 402;Bilandzic, 2005, pp. 362-364;2017, pp. 407-408;Coffey, 2014, pp. 367, 377-373;Crosnoe, 2000, pp. 379, 387;Marotzki, Holze, & Verständig, 2014, p. 457;Marvasti, 2014, pp. 359-361;Samuelsson et al., 1996, pp. 327-330). We divided the analysis of the collected data into a number of steps (see Fig. 4.4), which are closely linked to each other, so Fig. 4.4 Overview of data processing and data analysis that the team is able to go back and forth between original data and, for example, the coded material, in order to clarify issues of interpretation (see Paus-Hasebrink, Sinner, Prochazka, & Kulterer, 2018 for a more detailed description of the process of analysing qualitative data in a longterm study).
(1) The transcription of interviews, including consistent and thorough anonymisation of names and places throughout all successive waves of interviews with panel members: The literary transcription of the audio files of the interviews (including the call-back interviews) was the starting point in each wave. Due to the sensitive character of the data and the possibility of unmasking our interview partners within small-scale communities, we rigorously practiced a strict and consistent anonymisation, including the separate storage of the original audio files, the anonymised transcripts and the anonymisation protocol with proper names and aliases (see Marshall & Rossman, 2016, pp. 212-213;Mertens, 2014, p. 510). However, we made use of anonymisation not only for proper names (family, friends, peers, teachers and so on) but also for pets' names, locations (stores, schools, facilities) and locations of leisure activities. Strict rules were essential for us to coordinate our subjective impressions from all stages of the project. However, over time we had to pass material from one staff member to another. To maintain transparent processes, we, therefore, conducted employee training and meetings of the coders at regular intervals. In order to support the process of transcription, we used the advanced transcription software f4transkript (Windows) and f5transkript (MacOS) (for example, automated switching of speakers, time-markers, slowed playback speed (Dresing, Pehl, & Schmieder, 2015;Marshall & Rossman, 2016, pp. 209-210).
(2) The development of a comprehensive coding scheme for the analyses of all the forms of data, and computer-assisted coding of data using qualitative data analysis software (MAXQDA): We applied thematic coding (Flick, 2013;Kuckartz, 2010, p. 10) combining deductive and inductive steps. Based on theoretical considerations of grounded theory research (Strauss & Corbin, 1996, 1998, we made use of open coding, axial coding and selective coding. Categories were defined partly deductively, based on the above-mentioned considerations, and inductively, based on the actual material. The coding schemes (one for parents, one for children) were continuously enhanced throughout the successive waves of interviews with panel members, in order to grasp new phenomena (for example, new media services). However, we tried to keep it consistent over time, and coders were not authorised to make changes on their own. Instead, we used memos and meetings of the coders to discuss any changes within our schemes and any later additions. (3) The data analysis followed an approach developed by Paus-Haase and Keuneke (Paus-Haase, Hasebrink, Mattusch, Keuneke, & Krotz, 1999, pp. 143, 147) As a first step, we conducted a "focused analysis" of all data across all families: This was conducted in alignment with the main categories of the coding schemes: the role of family, of friends and peers, school, media usage, leisure activities, financial aspects and personal ambitions. Comparability across the waves of data collection was ensured by thematically structured matrices for each family, which included data from all waves and all methods of data collection. These matrices are organised by year (2005, 2007, 2010, 2012, 2014, and 2016) and category. Due to the long-term character of the study, the knowledge we generated was not always linear and chronological.
In some cases, we were able to clarify situations in later years, or we were able to find out more about past events. In such cases, we updated not only the matrix of the family but also family profiles and cases. In order to support the processing, we also made use of the summary-function and the summary-grid-function included in MAXQDA. This meant we were able to portray developments over the years systematically. (4) For the next step, we conducted a "contextual in-depth analysis" (Paus-Haase et al., 1999, pp. 143, 191) using the three analytical concepts, options for action, outlines for action, and competences for action: Based on a qualitative analysis of the entire data set, we created a characterisation for each of our children within its respective family. By accounting for the contexts, we were able to understand how the child developed its cognitive and motivational resources and gained competences for action, which, against the backdrop of the options for action, are reflected in the realisation of the child's and the parents' outlines for actions (see Paus-Hasebrink et al., 2017, pp. 64-66;2018, pp. 218-221;Paus-Hasebrink & Kulterer, 2014, p. 70). (5) Case studies: On completion of the second wave, we went on to select families of particular significance. We examined these case studies even more intensively as regards options for action, outlines for action, and competences for action. In order to assure that this qualitative analysis was comprehensible to all our researchers, we applied two safety mechanisms: on the one hand, we used strict guidelines for composing these particular cases and describing them. On the other hand, two team members always dealt with these case examples together, in order to check each other's approach, but also in order to support each other's understanding. Later on, we compiled fact sheets for all the other families too and developed them into full family profiles. Consequently, there is no difference in our treatment of the individual families in this study as we applied identical methodology to all of them (see Chapter 5). (6) The family characterisations, as determined in the two previous steps, served as the basis for identifying family types (Kluge, 2000): Using our three analytical concepts, we defined three criteria, which, in turn, indicated meaningful differences between the families: socio-economic circumstances as indicator of options and for action; socio-emotional climate as indicator of outlines for action; coping resources as indicator of competences for action. Each family was characterised by these criteria indicating a specific pattern for each, and based on them, we were able to define similar cases as types (see Chapter 8 and Paus- Hasebrink et al., 2018, pp. 221-222).

ethIcal challenges
As we have shown above, conducting a qualitative long-term study over twelve years means intensive and close contact to the families studied, because our data collection was conducted at their homes and not in a laboratory or at a neutral location (see Marshall & Rossman, 2016, pp. 146-147). Conducting (qualitative) research with children and subsequently with adolescents does, then, require particular care (see also Marshall & Rossman, 2016, p. 162). Qualitative researchers "situate themselves in consciously value-laden territory in which human relationships and critical self-reflection loom prominently. This positioning leads to the emergence of ethical dilemmas throughout the conduct of research that go beyond legal requirements or many professional standards for ethically responsible research" (Mertens, 2014, p. 510). How do researchers interact with their subjects without being either patronising or didactic? How do they retain their role and not become a form of consultant, who might influence the subjects of research? How do they deal with severe problems and conflicts within the families, for example, psychiatric disorders, physical violence, experiences of abuse or suspicion about (sexual) abuse? At what point are previously defined limits exceeded, so that researchers cannot remain in their role and are obligated to interfere or to provide support (see also Mertens, 2014, pp. 516-518;Patton, 2002, pp. 326-327, 405-415)? Furthermore, we had to deal with the small-scale structures of Austrian communities, where individuals might be recognised easily. Given the sensitive topic of our study, we had to find ways to guarantee absolute anonymity to our participants while going deeper and deeper into their personal spheres (see data processing; see Marshall & Rossman, 2016, p. 126).
All the difficulties outlined in the paragraph above occurred during the twelve years of research. Each time, we had to deal with a particular case based on individual circumstances. We discussed each situation before carefully deciding whether or not to interfere. Any intervention then resulted from further discussion of the correct procedure. In some cases, we involved a psychologist at the university, not only to request support in deciding our own action, but also to possibly offer professional help to affected families (see also Lobe, Livingstone, Ólafsson, & Simões, 2008, pp. 29-31, 47-49, 52-53, 57-59;Ólafsson, Livingstone, & Haddon, 2013, pp. 39-42, 75-76;Paus-Hasebrink, 2008, p. 57).

conclusIon
In this chapter, we have explicated the methodological design behind operationalising our theoretical approach to researching the role of media within children's and adolescents' socialisation in socially disadvantaged families. We conducted a longitudinal study to understand the complex interplay of subjective and structural factors which shape the lives of children as they grow up. We applied a complex design, convinced that "triangulation (…) increases credibility and quality by countering the concern (or accusation) that a study's findings are simply an artefact of a single method, a single source or a single investigator's blinders" (Patton, 2002, p. 563). However, we stress that "it is important to avoid the use of triangulation simply as an end in itself. Rich design or triangulation is only valid if it is applicable to the research question" (Paus- Hasebrink et al., 2013, p. 23). Over the years of our study, this challenging approach proved suitable for researching complex processes. Our experience shows, however, how important it is, when conducting a qualitative long-term study, to follow strict rules and to work intersubjectively traceable, ensuring that all aspects remain comprehensible to the researchers involved. At the same time, we found it essential to react to the new challenges presented not only by medial and technological changes but also by social changes and by the development of the children themselves.
Conducting a complex research project is always a sophisticated process. Methodical procedures and ethical guidelines need to be followed and fulfilled. Moreover, the research team has to deal with restriction in relation to financial, temporal and personnel resources. But every kind of project has specific peculiarities. To conclude, we want to highlight two aspects that turned out to be of particular relevance conducting our longitudinal study over twelve years. At the same time, to meet these aspects was extremely challenging: The first point concerns the maintenance of the panel. Not only we had to convince our families to remain part of our project over and over again, but also, we had to engage in keeping the contacts alive. Therefore, we had to learn to gather as much contact details as possible (for example, different e-mail addresses and telephone numbers of mothers and fathers, not only a single one), because the contact details were changing quite quickly. Furthermore, when the children were older, it turned out to be purposeful to ask them for contact details and the permission to contact them, too. The second point concerns the handling but also the storage of the data. As said before, the strict compliance of rules and intersubjective working methods are indispensable in order to generate valid data in a qualitative longitudinal study. In this context, a major challenge was to handle our sensitive data with care, this included not only the processing of data, but also the safe storage of data and the administration of access rights. Our solution included a mirrored database with backups on different hard drives in spatially separate rooms and strict access rules to the entire data. As mentioned above, this included as well the separate storage of anonymised data and the anonymisation protocol. Furthermore, we had to deal with rapidly increasing volumes of data. This development was caused by additional methods of data collection on the one hand, but also by more extensive answers of our subjects (due to more complex media repertoires and living situations) on the other hand. Having these experiences in mind, it seems to be essential to plan data handling, data storage and panel maintenance very carefully from the beginning, when it comes to developing a longitudinal research project comparable to our study.