Researching co-viewing on social media and instant messaging applications: ethics and challenges

This article discusses the ethical concerns and challenges that should be considered while studying the practice of co-viewing on social media and instant messaging applications. Coviewing practices refer to the intertwined activities that happen while viewers sit together in front of a TV set, watching and making meaning from television content. Connected platforms amplify the possibilities for co-viewing, by allowing people to experience a digital co-presence. Users that engaged in connected co-viewing in unofficial Facebook and WhatsApp groups dedicated to the brazilian telenovela Babilônia (airing in 2015) are used to exemplify the need to establish informed consent and avoid harming the participants when carrying out research online. The research reveals that ethical decision-making with respect to users’ data and viewpoints had to be considered, not only at the beginning stages of research, but assessed and considered throughout each step of the study.


Introduction
Doing research on the internet and its evolving platforms is necessary as digital enviroments are part of our daily lives.Undertaking research about these spaces can involve both advantages and disadvantages.As many authors have already pointed out, it can be less costly and less time consuming than research in non-digital environments.However, while engaging in research on connected platforms, the research design may end up as a continually evolving process as the digital spaces are always changing.
The evolving research design was evident in this case study of two closed Facebook groups and one WhatsApp group dedicated to co-view the Brazilian telenovela Babilônia (that aired March to August 2015 on the Globo Network).The study aimed to understand how co-viewing practices are reshaped by digital platforms and shaped also by users' practices and experiences within these spaces.During this research, it was necessary to continually adapt the design due to the empirical data found in the fieldwork, the issues imposed by the interfaces where the research took place (Facebook and WhatsApp) and by the users' co-viewing activities.Although the groups that were studied followed a television genre that has a stable basis for consumption, the phenomenon of connected coviewing on Social Network Sites (SNS) and instant messaging applications was relatively new, and it had a degree of unexpected events.This is because the studied phenomenon was not part of a controlled environment, the Facebook groups themselves were in the process of formation, and the telenovela replaced a previous telenovela (telenovelas usually only last six months).Another unexpected event during the research was the formation of a WhatsApp group within one of the Facebook groups.It led to investigate co-viewing within this space as well, which generated new insights.
Research, when done online, can involve significant ethical concerns, especially for qualitative researchers where the researcher should be sensitive because the online environment might affect the offline context in which the participants live, and it might harm them (Eynon, Fry & Schroeder, 2008).The internet is also an environment where new phenomena and features are constantly being introduced and it is necessary to be aware of the potential risks to the participants.For instance, it was observed that users collected, shared and archived files and links with the telenovela content and soundtrack without copyright permissions.Some of these activities are illegal and could affect the lives of participants.
The changing nature of the research problem required a case study methodology that followed an ethnographic approach.The general goal of the study required what Stake (2005) has called instrumental cases studies, as particular groups were examined primarily to provide an insight into an issue, in this case, an insight into the connected co-viewing practices of a type of TV genre, the brazilian prime time telenovela.
As demonstrated in the following sections, the outcomes of this adaptive and evolving research design are the products of flexible techniques for interacting with participants and working within the fluid environment of SNS platforms.

Coviewing and the Brazilian telenovela genre within connected spaces
Co-viewing studies originated in the psychology field in the 1970s (Ball, Bogatz, Creech, Ellsworth & Landes, 1970), where studies started to acknowledge the importance of peers at the moment of meaning-meaking while performing viewing activities.Co-viewing studies as Cultural Studies left behind traditional approaches that tended to view the audience as passive or simply not included as a subject of inquiry.From the very beginning, co-viewing studies pointed out that watching television has always been a complex social activity, as it generates various forms of social engagement among peers (McDonald, 1986).Before the widespread usage of connected screen devices and platforms, engagement around a television broadcast was limited to the TV set.Thus, TV engagement had an ephemeral feature because it was not easily traceable and viewer's activities materialized mainly in-person.Holt and Sanson (2014) named the current complex media context connected viewing which "refers specifically to a multiplatform entertainment experience that relates to a larger trend across media industries to integrate digital technology and socially networked communication with traditional screen media practices" (Holt & Sanson 2014, p. 1).Co-viewing in the connected era can happen in a number of ways i.e. physically alone but in the company of others online, both accompanied offline and with other online peers, or in the traditional way in which people sit and co-view audiovisual content together.In addition, co-viewing in connected platforms can leave traces in the format of data, as these digital spaces are based on user-generated content.Now, however, coviewing can be materialized in front of a TV set or using other kinds of screen devices and SNS that were not available in the early years of the Internet.
In both traditional and connected environments, co-viewing enables meaningmaking by sharing opinions and interactions while viewing.Thus, the characteristics of a co-viewed show play an important role for such practices to occur.The kind of media content that is co-viewed could spark different practices related to the process of reasoning around it, and the companionship co-viewers intend to have.The study of Dezfuli, Khalilbeigi, Mühlhuser, & Geerts (2011), concerning interpersonal relationships around TV viewing, demonstrated that just like in-person co-viewing, users who co-view online can classify their social circles to select with whom to interact, depending on the situation and the TV genre.Harris and Cook (2010) add that media enjoyment may be mediated by different genre preferences, or situations of viewing companionship, which leads a person to greatly enjoy content or to be quite distressed by it.The presence of other co-viewers might be one of the reasons for the disruption of a viewing experience, particularly when the co-viewing revolves around delicate subject matter such as sexuality.
In this study, the focus was on a specific local context of co-viewing practices related to the Brazilian prime time telenovela within unofficial Facebook and Whatsapp groups.This unique television format is central for observing co-viewing practices, as it includes social and political issues in its plots that are extended beyond the television set (Atencio, 2014;Lopes, 2012;Pires de Sá & Roig 2016;Pires de Sá, 2017).For decades, Brazilian telenovelas have been co-viewed among people that ritually reunite themselves around the television, and now with SNS this ritual is being reshaped and made more visible through the use of different kinds of digital materialities.The co-viewed telenovela that formed the basis for this study was called Babilônia.Babilônia presented very delicate social issues such as female homosexuality in the elderly, political corruption and so on.All these problematic subjects formed part of the connected co-viewing practices.Babilônia's controversial plot led to discussions in the political arena and prompted debate in multiple media.Thus, connected co-viewing is a practice that also extends to viewers of the brazilian telenovela, and while studying this phenomenon, researchers have to be careful not to expose people's online co-viewing activity and therefore potentially cause harm to them.

Selection of the Facebook platform
The study was initially designed to be done on Facebook groups specifically dedicated to the telenovela.This was decided after spending several months observing the groups that were created to follow the telenovela Império (that aired from July 2014 to March 2015) -the predecessor of the co-viewed telenovela in this study -and seeing the users' interest in new groups dedicated to Babilônia even before it started to air.
This decision was also inspired by a study on Facebook developed by Lopes (2012) who set the ground to further investigate telenovela related activities inside groups created by users, instead of only looking at the television network's sanctioned or official spaces.All in all, Lopes' works have provided a different perspective to reception studies in Brazil and consolidated the tradition of studying telenovelas.The author demonstrated that the telenovela is part of Brazilian culture and it is a shared repertoire (Lopes, 2009) that has represented the process of industrialization in the country (Lopes, 2000) and continues to be a TV genre that represents a crossing point between popular culture, everyday life, social issues, politics, and media industries.Lopes (2000), based on Martín-Barbero's Mediation Theory (1987) and Orozco's Multiple Mediations (1993) theory, implemented the Mediations framework as a theoretical-analytical tool to overcome the dominant approaches to communication studies: the deterministic approach based on the Lasswell and Shannon model and the Frankfurt School approach.The author overcomes these dualisms by empirically exploring the multidimensional ways of mediation in the reception process of a telenovela.The works of Lopes (2000) and Lopes, Borelli, and Resende (2002) can be considered groundbreaking, since they explored the viewing experiences of four different families from different social classes.The studies looked at where mediation was taking place and by establishing mediation categories of analysis (family daily life, subjectivity, genre and video techniques).This approach allowed them to look at reception through different lenses, including the role of power and discourse.The innovative works by these Brazilian scholars has validated a multidisciplinary methodological approach to study the telenovela reception process by using different qualitative techniques (not only content analysis) and bringing empiricism to the Latin American Mediations framework.Furthermore, this seminal research has fostered an interdisciplinary and empirical approach to the reception phenomenon in other studies.Therefore, it was a step forward in brazilian reception studies, particularly in the way of doing systematic empirical work.Aiming to extend this multidimensional approach, coviewing was studied in its connected format as a way to contribute to what Lopes (2014) has already pointed out: reflecting on the importance of technology in reception studies in order to extend the status of the theoretical and methodological aspects of communication research.Therefore, this study also aimed to contribute to the few recent studies that have tried to understand the viewing activities on connected platforms from users' viewpoints that were developed by the Brazilian team at the Ibero-American Observatory of Television Fiction headed by Lopes, works that are further discussed in a related study (Pires de Sá, 2017).
The decision to develop this study using Facebook was attractive because Facebook at that time was the most used SNS in Brazil (Secretaria de Comunicação Social da Presidência da República, 2015).I was able to observe how the practice of watching television together was transferred and somehow transformed by the use of SNS and other information technologies.On the other hand, it was also made more difficult because Facebook had an Application Programming Interface (API) with many constraints to extract data.An API is: (…)an interface provided by an application that lets users interact with or respond to data or service requests from another program, other applications, or Web sites.APIs facilitate data exchange between applications, allow the creation of new applications, and form the foundation for the 'Web as a platform' concept... programmers can use screen scraping, a technique for extracting any information, of any type, from any Web page.(Murugesan, 2007, pp. 36-37) Twitter, in comparison, would seemingly have been an easier SNS to facilitate the study, as most data is publicly available, easily retrievable, and this platform is frequently used by researchers (Moe, Poell & Van Dijck, 2016).Nonetheless, it did not provide the possibility to analyze groups' co-viewing activities in the same manner as Facebook.As I aimed to undertake participant observation to better understand users' practices, rather than solely reviewing the content they produced, this study required a case study methodology.Informed consent was obtained from the users and I followed the recommended ethical guidelines of the Association of Internet Researchers' Ethics Working Committee (Markham & Buchanan, 2012).

The case study approach
The justification for the selection of a case study as the applied methodology comes from the phenomenon I decided to study, where the environment could not be controlled.A case study makes it possible to look at a subject from many and varied angles that can help to achieve a better and balanced understanding of the object of study.It is also appropriate to use when the researcher has little to no control over the events that are under investigation (Yin, 2003).Flyvbjerg (2006) points out that a case study provides the researcher with a closeness to real-life situations for understanding a social phenomenon.Multiple circumstances and details arise during the process of investigation, which is context-dependent (Flyvbjerg, 2006), including ethical decision-making.It means that the case study produces experiential knowledge, as the objects under study are complex and can be marked out by their contexts and through circumstances of the occasion when being studied (Stake, 2005;Thomas, 2015).Despite their divergences in terms of paradigm, Yin (2003), Stake (2005), Flyvbjerg (2006), Thomas (2015), and others affirm that one of the main strengths of using this approach is that it allows the use of various sources and multiple techniques to understand the object of study.Therefore, a triangulation of methods allows the case study to obtain more reliability, as during the period of study, descriptions and interpretations are not done in a single step.Instead, many steps are made in various moments of the research.Thus, leading the researcher to undertake a constant process of reflexivity.
Additionally, triangulation allows for a better understanding of the situation in which a case is embedded.In the cases of this research, the methodology and its multiple methods had the influence of ethnography in instrumental and relational terms.The instrumental terms relate to the online methods that Kozinets (2002Kozinets ( , 2010) ) called 'netnography' (adapted from ethnography) that served as research tools to explore the groups that exist in digital platforms like Facebook and WhatsApp.The relational terms helped to comprehend the complexities of users' activities because of their interconnection.
Furthermore, when doing a case study, a close look at the case's activities provides a proximity and better comprehension of the object of study, which allows the researcher to receive constant feedback from those under study (Flyvbjerg, 2006).
By using a case study approach, I was able to undertake this process of reevaluation that allowed me to observe whether my propositions were coherent or if they should be further interrogated and reviewed.As the cases I studied did not follow a deductive model, and my approach was closer to the phenomenological paradigm, it was possible to follow this process of reassessment throughout the distinct phases of the study and adapt accordingly.I began the study with preconceived ideas and objectives, which had to be adjusted during the process of data collection.For example, a working hypotheisis proved to be correct that unofficial spaces within SNS are places where users can escape the direct view of producers and feel comfortable to perform viewing practices that otherwise would not happen.However, my expectations of finding many users creating mashups, videos and memes as part of these activities were not met.Although users shared these types of content during the airing of the telenovela, they did not develop their own content.When talking with users in informal conversation and during the interviews they affirmed that they do create content.Nonetheless, the way they understood content production was different from my initial conception in that the act of providing an opinion or sharing any telenovela content was seen as producing content.But content do not need to be sophisticated like mashups or videos.This required adapting the research questions and the objectives.
These changes could only be made because I had close contact with the users, as well as with their practices and the environment in which co-viewing practices occurred.Therefore, the case study helped me to learn more about the phenomenon that was being studied and to contrast it with prior conceptions.
This study used a triangulation of methods to better understand the co-viewing phenomena: participant observation, semi-structured interviews, and a qualitative questionnaire.This triangulation enabled an in-depth analysis of the viewing practices of the groups that were selected to be the cases (two unofficial Facebook groups and one What-sApp group).The in-depth analysis was possible because of the applicability to real-life and contemporary issues that the case study approach allows when looking at an ongoing phenomenon, such as co-viewing through connected platforms.

Selecting the cases
As co-viewing happens when media content is being viewed, a methodological decision was made to scrutinize these practices at the same time they were going on.Therefore, co-viewing of the telenovela Babilônia was selected, as it was the next prime time telenovela airing on the Globo Network after the project started.
For the objectives of this study, I chose to focus the attention on unofficial Facebook groups and the activities performed within them instead of researching "official" spaces mediated by the network, where it would be difficult to see practices that might be considered illegal such as file exchange.Furthermore, the aim was to observe people's activities, and not a constructed idea of an audience or fan predetermined by institutions (Ang, 1991) or shaped by the network.I also preferred to focus on unofficial spaces, by not looking at them as the most representative place to record co-viewing experiences in general, but as ideal places to observe and participate in the practices of connected co-viewing without possible interference or surveillance by the Globo Network.It must be noted, however, that these groups were not ideal open spaces for free debate as there were forms of moderation and even some censorship by the group owners and administrators, as well as by filtering algorithms that Facebook uses to highlight some posts (Araújo & Pires de Sá, 2016), and other direct and indirect restrictions posed by the platforms (Pires de Sá, Araújo & Roig, 2017).
Although the groups used the platforms of a private corporation like Facebook, which also owns Whatsapp, users perceived them to be more appropriate places for developing particular practices.Therefore, it was possible to observe the practices of file exchange and content sharing, debates about the function of the telenovela, its technical features and the way it represents reality.
The three studied groups were selected differently.Nonetheless, for this research, I followed the logic of instrumental collective case design defined by Stake (2005).I used instrumental collective cases to achieve a better understanding of the connected co-viewing phenomenon.Stake (2005) promotes the use of an instrumental case study when "(…) a particular case is examined mainly to provide insight into an issue or to redraw a generalization.The case is of secondary interest, it plays a supportive role, and it facilitates our understanding of something else" (Stake, 2005, p. 445).In this research, I was interested in understanding connected co-viewing and User-Generated Content (UGC) practices related to a particular type of TV show (the Brazilian prime time telenovela).In order to achieve this goal, the Babilônia groups were studied in depth, as Stake (2005) affirms that "the case still is looked at in depth, its contexts scrutinized and its ordinary activities detailed, but all because this helps us pursue the external interest" (Stake, 2005, p. 445).
Many Facebook groups dedicated to following Babilônia were created even before Babilônia began to air.They started to be formed more than a month before the previous telenovela (Império) was over, when the Globo Network released the synopsis and aired some teasers during the breaks in its programming.Within existing Império groups, creators and administrators of the new Babilônia groups invited users to participate, which assured that there would be groups to study the co-viewing and UGC practices vis-à-vis this television genre.They were around ten active groups one month before the release date of Babilônia (its first episode aired on the 16 th of March 2015).
To select the cases from this population of Babilônia groups, I followed purposive sampling by selecting the two most active groups, taking into account the groups where people posted and commented more often.This is because there would be fewer examples of co-viewing practices in groups with fewer interactions.As Rubin and Babbie (2016) point out, a purposive sample is a technique in which you handpick your cases because they have certain attributes or perform practices that might represent a segment group that is relevant to the study.Hence, this purposive sampling was done to participate in spaces where co-viewing practices would materialize because of the users' activity.The WhatsApp group, on the other hand, was not intentionally selected.In one of the studied groups, a Whatsapp subgroup was created, in which users were invited to participate by leaving their phone numbers in a comment.As it was in my own interest to observe co-viewing in a connected way, I also provided my number.Thus, the WhatsApp group selection conformed more to the snowball type of sampling that is often employed to try to "(…) obtain information on and access to 'hidden populations'…" (Noy, 2008, p. 330).

Informed consent and ethical concerns
This research was done within the micro-ecologies of Facebook (two unofficial and closed groups) and WhatsApp (one group), two platforms that are always adding or changing features.These are digital spaces that can have grey areas at the moment of ethical desion-making.In order to overcome these grey areas during the research process, the AOIR suggested guidelines were followed, which indicate that an inductive approach is one of the best-suited ways to avoid harm to research participants and overcome their possible vunerability (Markham & Buchanan, 2012).
In this study, the inductive approach was based on several aspects, but it was always considered that harm can be defined contextually, so the ethical decisions were taken according to each situation during the process, rather than applied in an universal manner (Markham & Buchanan, 2012;Nissenbaum, 2010).Within WhatsApp, the phone numbers of participants are required to be able to use the services and to connect with peers, which can be perceived by users as a very private channel of communication.On Facebook, users have to register their metadata (name, gender, etc.) in order to have access to the services.In both cases, the perception of privacy was not straightforward.Instead, as already pointed by Nissenbaum (2010), participant expectations and consensus regarding what is meant by privacy play an important role.Researchers should not ignore the participants' perceptions of privacy.Especially in cases where users may not be aware of data privacy settings, as raised during this study.Users perform many online activities per day, and yet do not know much about data regulations (Madejski, Johnson, & Bellovin, 2011).When engaging with the participants on this topic, many affirmed to just pressing click-to-agree and not reading the terms and conditions nor the user license agreements that are usually shown when creating a profile in SNS or when updating applications.
This means that depending on the circumstances, research outputs might outrage participants (Reilly & Trevisan, 2016) and researchers need to go beyond only adhering to the platform's conceptualization of privacy.Thus, informed consent should always be an issue to keep in mind before starting a research study on connected platforms, as online and offline contexts are often interwoven.Therefore, during the different phases of this research, it was noted that any decision-making while working within digital spaces involves data that could reveal sensitive information from individual persons (Markham & Buchanan, 2012) and in some cases cause anxiety (Pink, Lanzeni & Horst 2018).
There are debates as to whether Facebook is a public space.Following these debates, Reilly and Trevisan (2016) looked at the Terms and Conditions and pointed out that Facebook might be both public and private, or even a semi-public space, based on the privacy settings.In the case of spaces like Facebook groups, settings are either public, closed, or secret.The studied groups had the status of closed, whereby only members can post and see each other's posts.Therefore, it was considered a semi-public space, and informed consent was a major issue that was taken into account.
After selecting the case studies, and before starting the process of data collection, a Facebook page was created.This was done to provide information about the different stages and news related to the research project, as well as to inform potential participants about the general aims of the study, and act as an open channel for communication.Furthermore, as Hine (2008) points out, it is important for online researchers to have informative web pages, as informants are able check the researcher's identity and affiliations and see us as legitimate scholars when undertaking a virtual ethnography study.Hence, a link was included on this Facebook page to the website of the university I am affiliated with.This link included a description of my research group and its members.
The choice to use Facebook as the platform to host this page was based on the idea that the studied users would be using the same SNS they use to access the groups.Therefore, facilitating easy access to the page, as they would already be logged in to Facebook.
This page was also a tool that helped to acquire informed consent from the group owners, administrators and studied users.It helped me to be clear about the objectives of the study and to explain that the collaboration of the users was voluntary and not obligatory (Estadella & Ardevol, 2007;Hine, 2008;Miller & Boulton, 2007).
First, the group owners and administrators were contacted via my personal Facebook profile to receive pertinent information about how I intended to develop this research project.The decision to use my real Facebook profile while participating in the groups or talking to users was done with the intention to transmit trust.Therefore, this enabled the creation of solid bonds with the users while involved in the group practices, as they would see the researcher as a real person.The Facebook page was fundamental for establishing this contact and presenting the study.
Second, after the agreement with owners and administrators, several posts were made within the groups to engage with the participants.These posts explained that I would be participating in the groups while studying the users' activities and gathering their data.Where necessary, a user's doubts were resolved through comments and during conversations via Facebook Messenger or WhatsApp.Users were also encouraged to speak out in case they did not want their information to be used, something that happened with two participants.
Dealing with concerns regarding data anonymization also helped to acquire the informed consent.It was explicitly stated to informants that their personal information would not be revealed.It was acknowledged during informal conversations with users, during the interviews, within the presentation of the questionnaire and in publications.Although I used the real name of the telenovela they were co-viewing in the research outputs (presentations and publications), I never used the actual names of the groups, nor provided any links to their posts.Moreover, both groups underwent name changes several times, thus disabling the recognition of the studied groups, even more with the large flux of information that they had.

Challenges of studying co-viewing in a connected context: the need to move between data collection and data analysis
As a participant observer, it was crucial to co-view the show together with the group members.By performing the activities with the users, it was possible to understand the ways in which these practices were carried out and their causes, which were usually related to the features of the interface and its constraints.For instance, the WhatsApp group had far fewer users than the Facebook groups, and yet the activity was much greater.This can be attributed to the WhatsApp interface.It is an instant messaging application that allows messages to be delivered and accessed quickly on the mobile phone.Whereas on Facebook, it is necessary to be connected to this SNS to see the posts within the groups of interest.Additionally, while participating in co-viewing it was possible to engage in the discussed topics and the activities in the groups and then compare with the datasets.This comparissom of participant observation with mined datasets was essential, as it followed the understanding that co-viewing data as other digital data go beyond the idea of an archive.These activities were seen as dynamic, real-time human practices that take place in digital spaces (Kitchin, 2014;Pink et al., 2018).If I was not part of the groups and did not participate in the co-viewing activities, it would be almost impossible to read or have a deeper understanding of the content and data generated during pre-viewing, co-viewing, and post-viewing activities.Therefore, the process of studying co-viewing in connected platforms was based on moving between data collection and data analysis that proceeded in a constant process of evolution and adaptability.
During the first month of the airing of Babilônia, after co-viewing the telenovela within the studied groups, the N Capture plugin in NVIVO 10 was used to mine the datasets of each day from Facebook to compare with the field notes and lived experiences.
Every day, I used the coding features of NVIVO to establish and store relevant data in categories and concepts through what are called nodes in this software.It was necessary to do it every day because of the high volume of data that was produced.During co-viewing, these data appeared to be less than at other times, as I was seeing and participating in the activities happening almost synchronously.Nonetheless, when looking at the daily datasets it was actually more, as each post or comment represents a discrete row in the dataset in NVIVO, even when it only includes emoticons or single words.For instance, during the first episode, the dataset contained nine hundred and forty-two rows (including posts and comments) from both Facebook groups.Therefore, it required a constant process of selection and categorization of the relevant content.This selection was based on my experience of co-viewing and the notes taken during this practice.
At the end of April 2015, when Facebook changed its API, it was necessary to adapt the data collection, as I could no longer continue to extract the data from the Facebook groups.The solution found was to create an Excel spreadsheet with the most relevant activities.This relevance was based on the topics and on the engagement of the viewers.Usually, each saved interaction had at least ten comments or more.This decision decreased the amount of data I had to store, and changed the process of coding.In the spreadsheet, I included my notes, links and saved the trace of the entire interaction among the users in one discrete row.Those excel sheets were imported to NVIVO and coded once a week.While in the WhatsApp group, I made notes from the activities and at the end of the fieldwork I downloaded a text file and imported it into NVIVO.
The interviews were also conducted and accessed for an initial analysis during the data collection period as a way to dialogue with the users' viewing practices.They were coded (categorized) in two different ways.The first way was to create nodes of each question to be able to see all the interviewees' responses together.Second, by including relevant information in the nodes already created during the participant observation.
The questionnaire was designed based on the interviews and the obeserved coviewing practices.It was developed on Google Forms and made available during the last week of the airing of the telenovela.NVIVO was used for running queries and tests that were also applied to the data from participant observation and interviews, such as searching for the main used words to contrast with the existent categories, to see a word in a particular context, and so on.
In the beginning, the classification process of the analysis was intuitive, and a large number of nodes were created.After some weeks of airing, patterns started to emerge and nodes began to acquire a better shape, consequently decreasing the number of nodes during the process of analysis.After the end of the fieldwork, the coded data was further analyzed, and common properties and patterns of social practices were found among nodes.

Concluding remarks
This article demonstrates that the studied cases led me, the researcher, to "learn by doing", as some of my intuitions and preconceptions were changed while doing the fieldwork.This is because connected co-viewing is a practice that is alive and is always evolving.Connected co-viewing depends on the viewed genre, the co-viewers, the places where it occurs, and the technology and materiality available for carrying out this activity.It was found that the interfaces of the platforms (Facebook and WhatsApp) in which the research occurs are always evolving and have features that shape this experience.Thus, studying a "live" phenomenon such as connected co-viewing, in the context of SNS and instant messaging applications, has proven to be possible only with a flexible methodology that considered ethical decision-making as contextual in every step of the research, and constantly moved between data collection and analysis.
The explained methodological aspects of the study, the steps and the ethical decisions taken during the research were important because they contributed to the relatively recent field of inquiry regarding user activities on SNS platforms within reception studies, more particularly within telenovela studies.The methodological approach used in this study followed an inductive ethical procedure for doing participant observation and collecting data from SNS with the informed consent of users.Therefore, this created a space for debating the importance of the ethical treatment of the subjects under study who use SNS and instant messaging applications for viewing purposes.Furthermore, the need to address the research problem consciously is made explicit, because the way the data is treated might affect the daily lives of users, regarding issues such as data privacy, discussed social themes, illicit activities and so on.Also, this ethical decisionmaking formed part of the adaptive research process, as the object of study was under development.This confirms that conducting ethical research in the connected environment requires the researcher to make sensitive decisions during the entire process, as it is not possible to control the research setting.
Once again, it is valuable to highlight the informed consent procedures undertaken while doing this research and when participating in user activities.If I had not followed these methodological guidelines and did not engage in the group practices, I would only Researching co-viewing on social media and instant messaging applications: ethics and challenges .Fernanda Pires de Sá have developed a superficial understanding of the connected co-viewing practices and the intertwined activities that occurred in these kinds of spaces, thus, potentially maintaining my initial preconceptions.It must be emphasized that most of these practices could be studied only because they occurred within a space that was not established by the producers.Although official spaces can be convenient for researchers to delve into viewing habits, this study indicates that sometimes selecting unofficial spaces for observation can be fruitful for discovering practices that would never be seen in official ones.Furthermore, Facebook and WhatsApp were revealed to be often used platforms for the follow-up of live TV activities.
The procedural and contextual aspects of the research process may have limited the findings and contributions of this study.However, they also created opportunities for a better understanding of the context and the path-place array in which the study occurred.As explained, one of the constraints of this study was the technological restraints of the Facebook platform API that stopped working for mining closed group datasets.This required me to manually record the data for future analysis.However, this limitation was also an opportunity for developing a deeper understanding of the digital-material setting in which the practices of connected co-viewing took place.It was possible to understand how Facebook, SNS and instant messaging applications are in a constant cycle of being updated.This affects the research procedures, which requires researchers to be flexible and continually review their research processes.
Because of the high numbers of members in the Facebook groups and the many users in the Whatsapp group, it was not possible to get to know all the co-viewers involved in these practices, or to be completely sure that all participants were aware of the study.This is because the groups' membership grew over the months of the study.However, this was an opportunity to understand that ethical decision-making and informed consent when facing a vast volume of people and information, comes from the posture and decisions that the researcher takes according to the encountered context.It depends on the position that the researcher assumes, by striving to protect participants from harm and not looking at the SNS content as only text, but as part and property of the human participants.
Future research would extend the themes and methods explored in this study.The research approach and applied methods should be explored in similar and different contexts of viewing.For example, it would be interesting to research co-viewing in other television genres or within other types of viewing that are not marked by the broadcasting time.
In conclusion, this study has applied fundamental developments in methodological aspects for doing a case study on a live phenomenon such as co-viewing in social media and instant messaging applications while using a qualitative approach that followed the ethical protocols of informed consent.The original and significant contributions of this study will be further explored to better understand our expanding connected sociality.