4. Methodology

4. Methodology

Chapter Contents

This chapter will explore the methodological approach of the research. The first part of the chapter will explore my ontological position, both as a practitioner researcher and as a feminist. Furthermore, the section on reflexivity in this chapter will explain and offer more insight into how being honest about one’s ontological position as a feminist researcher can be beneficial to a topic (Wild, 2019). The chapter will then assess the rationale for the research design, including exploring the theoretical framework of contextual constructionism, which was used as a basis for initiating the thought process encompassing the research topic. Subsequently I will explore the epistemological position of my research, leading on to the data collection methods and the advantages and limitations within them. The chapter will then explore the ethical considerations that were faced in the research.

The Role of a Practitioner Researcher

Practitioner researchers are a growing phenomenon, their presence has been increasing since the 1970s (Jarvis, 1999, p. 4). There are schools of thought that are opposed to the notion that one can perform both roles adequately, particularly as traditionally scientific research has endeavoured to remain objective and detached (Jarvis, 1999, pp. 4,23). But there are many reasons why practitioners choose to research their own professions, and there has been an acceptance within academia of the emerging growth of practitioner researchers (Jarvis, 1993, p. 27, 179).

Whilst a full discussion on the role of practitioner researchers is outside of the scope of this study, it is important to acknowledge the benefits and disadvantages of my position. My career at the point of undertaking the research project spanned twenty-five years in the MVAW sector. Having worked in, managed, and set up, single sex spaces for women experiencing male violence, I had the advantage of understanding the theoretical basis and importance of women-only provision, resulting in me having an ‘insider’ researcher position (Aiello & Nero, 2019, p. 252). Whilst the ‘outsider’ researcher benefits from the curiosity of entering the topic with no knowledge and asks more provocative questions, the insider researcher gains from already understanding the intricacies and intimacies of a subject, thereby producing richer data descriptions (Aiello & Nero, 2019, p. 253). The active use of reflexive and reflective practices in research methods enables the practitioner researcher to acknowledge their position and for an ‘insider’ researcher it is possibly the most important element of the research (Aiello & Nero, 2019, p. 252), for me it was an enjoyable and enriching part of the process (Jarvis, 1999, p. 27-28).

Rogers (1961) asserts researchers need to take control of their study to “target their knowledge and take responsibility for what they hit” (Robson, 2009, p. 67). It was therefore essential that I was fully aware of my position and bias in the research and rather than identify information that only supports my viewpoint, which Rogers (1961) defines as an “ammunition wagon” (Robson, 2009, p. 67), I consciously sought out opinions opposite to my own. To this end, the combination of being a practitioner researcher and using qualitative research methods meant I had to actively take responsibility for my positionality throughout the research. Qualitative research is not an innocent practice, it is full of the researcher’s moral and ideological assumptions (Denzin, 2003; Gubrium & Holstein, 2001); positivist researchers argue strongly that bias has no place in the research endeavour, moreover, they believe that bias deems the whole project “nonsense” (Mantzoukas, 2005, p. 282). Therefore, the reflective nature of ‘insider’ research in this sense is an essential element to ‘owning’ the study. Feminist research is inherently reflective, with the feminist element being found in the relationship between the researcher and the research, as well as in the consciousness raising experiences and realities of being a woman (Stanley & Wise, 1993, p. 32).

Feminist Research

No one method of research is inherently feminist, but “feminist researchers do feminist research” (Clippingdale, 1996, p. 170); the way findings and studies are conceptualised, and the position of the researcher gives the research its feminist perspective (Skinner, Hester & Malos, 2005, p. 38). Feminists have long understood the need to place ourselves at the centre of our own research, the premise being, the “personal is political” and moreover, the personal is not inferior to science (Ginsberg, 2009; Hanisch, 1969; Stanley & Wise, 1993, p. 21). It has been suggested that feminist research is ultimately conducted “in the service of women’s liberation” and should be designed and carried out by women who are in the movement (Stanley & Wise, 1993, p. 25). Therefore, the desire to be honest, and transparent about my position as a feminist researcher was essential, particularly as this research provides an original contribution to the topic. In addition to this, the research design was congruent with the feminist ethos as qualitative methods are noted to be an excellent way to explore human behaviour and were instrumental in the initial academic work on MVAW (Skinner, Hester & Malos, 2005, p. 38).

One of the criticisms of the inherent sexism within social science from feminist researchers is that research focused on women should be done by them and for them (Corry, 2018), and far too often research about women is undertaken by men. I wanted to mirror the ‘by women for women’ premise (Charlton, 1972; Dobash & Dobash 1983, p. 2) in this research design, and as such I made a conscious choice to interview women only, under the cohort of participants opposed to gender reform and transgender ideologies. In doing this I hoped to honour the historical way in which research was designed for the MVAW sector, where silenced women were encouraged to speak about their experiences (Skinner, Hester & Malos, 2005, p. 38). However, there are those who believe that female-only services should be opened to self-declared transwomen and it was essential to include their voices; I did this by recruiting a mix of adults of both sexes, including those who identify as transgender, in the cohort I named ‘Pro Self-ID’. I understood that without reaching out to all parties the research could be criticised as biased, and if feminist research is to make a difference “it must be concerned with all aspects of social reality and all participants in it” (Stanley & Wise, 1993, p. 31).


I have spent over half of my life working in the MVAW sector. My position on female-only spaces has been something that I grew up with in my career and on reflection, I took this for granted. I first started working on the frontline of the MVAW movement in 1995 when I was just nineteen years old. I had the advantage of working in services that just a few decades earlier had been hard fought for by my feminist fore sisters. For a short period in my career, I accepted male workers on some one-to-one community projects in the sector, but I have never accepted men in female-only spaces and I now disagree with them being part of any provision. My strongly held view is that men have no place in any frontline MVAW services. Women repeatedly ask for women to support them and I have used the experience of working with men, alongside my research, to return to my feminist roots and to remind myself of the precious baton that was passed to me by second wave women. 

During the debate raging around gender reform, I watched as women were targeted and smeared for their opposition to transgender ideology alongside the policy creep into institutions, without any apparent scrutiny of the impact on women. I was frightened about speaking up because of the threat of loss of funding for the charity I lead. I attended some grassroots feminist meetings and met many other frightened women, as well as many brave, vocal and tenacious women, some of whom I am now lucky to call my friends. In early July of 2018 I did get ‘brave’ and tentatively sent a tweet stating that I supported a discussion around the impact of gender reform. My fears were justified as within 30 minutes of that tweet I was then subjected to targeted harassment from a lone, transgender rights activist over the period of approximately four days. Part of their online harassment involved contacting my employers, funders, and the Charities Commission, and I had to contact those organisations and individuals to refute the accusations of transphobia that were levelled against me. It was a particularly sinister experience and I reached out to some people working in the second tier umbrella bodies of my sector for support, but to no avail; they remained publicly silent and I viewed that as further evidence that I was wrong to want to discuss the implications of self-ID. 

I am aware my position and experience in this debate means the research is personal to me and there is no escaping the fact that I had never felt so silenced. The reality is that I, and other colleagues, felt frightened and this experience of feeling silenced in a movement that is notoriously loud led me to research this topic. I did not want to sit passively whilst witnessing the silencing of women, and the fundamental changes approaching female spaces. If I cannot protect spaces for women who have been abused and tortured by men, then what would be the point in my career? I will always believe that men cannot become women and I will not change my position on that, but I did not start this research to change my mind, more to understand why other people think differently to me. I appreciate there are swathes of people who do disagree with me, and that the wider ideological transgender movement would label me a bigot for my views but understanding ‘why’ was more important than the threat from trans activists. Above all I wanted to understand if and how women felt silenced and I wanted to give them a voice; if I am honest, I wanted to give myself a voice. I have grown up in a movement that has taught me to defend women and to rise up in unity, even if our voices shake, and I will always fiercely protect a woman’s right to speak. Above all I wanted a discussion about a sensitive topic that I know online to be incredibly polarised and a topic that my sector has left untouched. Ultimately, I challenged myself to remain open minded to those on the other side of the debate and I understood that they would be challenging themselves in doing the same with me. 

The process overall was incredibly rewarding, I hope I have contributed and supported the many feminist sisters who feel silenced. Equally, in speaking to participants who disagreed with me and sharing a space for respectful dialogue on differing political issues, I learnt a great deal and I hope this was as worthwhile to them. By speaking to people on both sides of a heavily contested debate and using the forum of academic research, I wanted to represent a fair balance to opposing viewpoints and provide a catalyst in fostering trust in the process of respectful disagreement. The aim was to initiate a wider discussion in my sector, victims watch what we as professionals do, they deserve better than our silence. The issue is not going away and the sand we have buried our heads in only goes so deep, when we come up for air the women we serve could have lost trust in our ability to model nuanced discussion and defend their spaces. As so many of us in the movement are survivors ourselves, we would do well to remember that a silent passive position is not one that reflects back the smashing of patriarchal power and control from those who wish to shut us up.

Contextual Constructionism

The theoretical approach for my doctoral research is influenced by contextual constructionism, which began through the 1970s as a way of analysing how and why social problems emerge and evolve (Best, 2001, p. 1). Constructionists examine the claims or rhetoric used to define social problems and the promotion of policy solutions, looking to the people who make claims (claimsmakers), and the subsequent response of press, public and policymakers (Best, 2001, p. 1). The basis for utilising this model is to define the social problem, and the arguments were researched with a basis of the ‘claimsmaking’ activities and policy solutions proposed by transgender ideologists in relation to single sex MVAW spaces (Orcutt & Best, 2001, p. 339).

Some claimsmakers greatly expand their grounds, aiming to appeal to the emotion of their audiences (Loseke & Best, 2003, p. 109), and they construct full blown typifications, based on false narratives, of the types of people harmed (Loseke & Best, 2003, p. 110). For claimsmakers, statistics are not necessary, as ‘audiences tend to be innumerate’ and numbers are meaningless without interpretation (Loseke & Best, 2003, p. 40). In its most basic form, trans ideologists claim that biological sex is on a spectrum and they use this as a launching pad for the gender identity debate, which is not founded in science but based in the philosophical framework of queer theory, as discussed in chapter three (Butler, 1990; Fausto-Sterling, 1993). Trans rights claimsmakers want their audiences to feel justified in ignoring all the scientific details of biological sex in order to understand the plight of transgender people’s oppression (Stein, 2017). It is undeniable that discrimination against transgender people exists, and it’s essential that they are free from this, but transgender rights activists make claims that are based not in reality, but in the culture of fear and emotion (Williams, 2020, pp. 78 – 85).

It matters not that transgender people are not all in agreement, and some have bravely spoken publicly against the lobbyists position (Hayton, 2018). But claimsmakers represent the loudest voices and achieve the biggest gains in relation to policy influence (Loseke & Best, 2003, pp. 4-5), this caused clashes with feminists as the claims lead to a redefinition of womanhood resulting in a move to declare open access for transwomen in MVAW spaces (Stonewall & NfpSynergy, 2018). By simultaneously ignoring the statistical data that evidences women’s desire for their own spaces (FOVAS, 2018; Women and Girls Scotland, 2019, p. 12), trans lobbyists also deny that men cannot become women, grounding their claims in a philosophical ideology (Harvey, 2019). In constructing transwomen as women, the claimsmaking activities of trans ideologists ignore the sex of most victims (Ingala Smith, 2019a). Of course not all claimsmakers have a need to ignore or manipulate statistics, feminists claimsmakers aiming to end MVAW are backed up by facts that evidence the global reality of male violence as a sex-based oppression (Ingala Smith, 2019a; Ingala Smith, 2019b; World Health Organisation, 2013). For example, the UK statistics evidence that 11 transgender people have been murdered between 2008 and 2020 in the United Kingdom (“TMM Absolute numbers – TvT”, 2020), compared to at least 1,425 women being murdered by men in the UK in last ten years (Long et al., 2020 p. 1; Moore, 2020). Despite this evidence,trans activist claimsmakers state that transwomen are subject to a murder epidemic and are far more at risk than females; these claims are made globally and have been debunked (Donym, 2019). Although it is worth noting that transgender murder rates should be considered in proportion to their population data, the statistical evidence shows that out of the 17,250 people murdered in the US in 2016, 27 of them were transgender, this equates to 0.0013% of the murder population, in comparison 3,895 women murdered in the US in 2016 (Donym, 2019; Smith, 2018). But, as stated, claims do not need to have statistics attached to them (Loseke & Best, 2003, p. 40), it is the repetition of them that appeals to the public and therefore with the backing of false statistics, the claimsmakers state that the exclusion of transwomen from women’s refuges is “cruel” (Afoko, 2018). The claimsmaking becomes even more interesting when trans survivors are reported as one of the most hidden groups amongst domestic abuse survivors, with victims themselves reporting that their experiences relate directly to their transgender status and they request their own specialist services (Magić & Kelley, 2019. pp. 26, 39, 50-52, 54; “The system is failing LGBT+ victims of domestic abuse | Safelives”, 2018). Yet the transgender lobbyists who use the evidence of the need for specialist services for LGBT + communities are simultaneously the claimsmakers who commission reports stating that transwomen need access to services designed for females (Bachmann & Gooch, 2018, p. 9; Stonewall & NfpSynergy, 2018, p. 6). Perhaps the reason for this relates to a concern that if transgender lobbyists supported the need for specific services for transwomen, they would be detracting from their claim that transwomen are women and the claim would lose its validity, but their position may not result in the best options for transgender victims themselves (Williams, 2020, p. 81). It is these issues that I wanted to explore, using qualitative semi-structured interviews and online ethnography. I wanted to understand if participants believed the claims made and if they did, was there a blanket acceptance that all transwomen should gain access to MVAW spaces? Additionally, as audiences do not interrogate the grounds or diagnostic frame behind the claims being made (Loseke & Best, 2003, pp. 49,109,112), the motivation of claimsmakers can get side-tracked, and by using the theoretical framework of contextual constructionism, the research aimed to look beyond the claims themselves and assess whether there is a deeper-rooted motivation behind them.

At the start of the research design I looked at some of the main claims from trans rights activists in online social media platforms, these are reflected in Appendix B, and I kept these claims in mind when undertaking all parts of the research design and analyses. The claims discourse evolved, particularly in relation to the analyses of the ethnographic and participant data at point of coding, and this is reflected in the results chapters.

Qualitative Research Design

Weber noted that the fundamental unit of investigation must always be the individual (Parkin, 1984). Weber thought it sometimes useful to treat social groups as if they were individuals for “collectives must be treated as solely the resultants and modes of organisation of the particular acts of individual persons” (Parkin, 1984, p. 17-18).  Weber introduced the idea of studying human beings and social groups using qualitative or interpretive methodologies. The qualitative traditions reject the idea that exploring the social world is possible by using the same methods as the natural sciences (Robson, 2009, p. 24) and when understanding the reality of the participants in this study, the paramount goal was to view the representation of their perspectives through their world. Indeed, feminist researchers have frequently criticised positivist approaches and quantitative practices (Robson, 2009, p. 21) and the aforementioned Weberian notion of ‘collectives’ is a useful framework in which to explore the themes that are appropriated in discourse, by using individual voices to represent a collective social group. Exploring individual perspectives in a free-flowing semi or unstructured way then leads to a possibility of making connections about them as a collective, which is neatly summarised by Weber who noted “collectives cannot think, feel perceive; only people can” (Parkin, 1985, p. 19). This is congruent with feminist activism as whilst highlighting the importance of victims’ collective voices, it is essential that individual narratives provide the catalyst for collective representation. The research design aimed to mirror this, by taking individual voices and aiming to analyse them as a collective representation of the social issue at hand.

Epistemological Background/Overview

An emerging interest in the 1970s was the exploration of qualitative research, sometimes known as interpretivism, constructivism or naturalism. Many researchers and advocates of qualitative methods express an issue with the scientific pretensions of quantitative research which they view as an incorrect way of studying people (Bryman, 2004, p. 24). This research design follows the same epistemological positioning by assuming that when studying people, or indeed a social phenomenon, the researcher should not start out with a preconceived notion of what the concerns are. What was important for this research is that I had a good understanding of the positioning of trans ideologists’ arguments in order to frame the research paradigm; because no research has thus far been undertaken from within the MVAW movement itself, I could not predict the positioning of participants. The individuals or the subjects of this research, when interacted with appropriately, naturally led me to the central issues of concern (Robson, 2002, pp. 271,274). This was particularly important given my position: for example, the decision to include participants who had beliefs that were in direct opposition to myself offered an opportunity for the data to create the collective narrative. Although quantitative research methods are useful for evaluating specific issues or problems, these methods are weakened when based on insufficient understanding of the problem (Skinner, Hester & Malos, 2005, p. 38). The predictive nature of quantitative research methods did not fit when beginning the research process as I was in a position of not knowing what the participants would say or how they would react. Positivists generally start with a theory whereas the traditions of qualitative designs are often based in grounded theory, where, by systematically using the data acquired, the researcher ends up with theory; and for any researcher using any methodology, it’s important that at the end of the project there is some understanding of the issue (Robson, 2009, p. 62). Bryman (2004) goes further, he describes qualitative research simply as “a different way of knowing”, which was of vital importance to me in researching a polarised and silenced issue (Bryman, 2004, pp. 26-27). In direct contrast to scientific positivist approaches, qualitative research reflects a different form of knowledge where people’s understanding of their social environments forms the focus of attention for the researcher (Bryman, 2004, p. 27). Parkin (1984) in his overview of Weber’s work provides an analogy of the differences between natural science and social science: “people unlike molecules or planets have motives for their action”(Parkin, 1984, p. 18). What we know about humans is that they have explanations for events and beliefs, providing their own interpretations to their social world and it is these interpretations and explanations that were of central concern for this research project, as participants could provide an account of their beliefs regarding transgender ideologists’ claims (Parkin, 1984, p. 18).   

Max Weber explained his method of social enquiry with Verstehen, this is the use of empathetic liaison with the ‘actor’ or participant by the observer or researcher. He distinguishes two types of Verstehen, these being “direct observational understanding and explanatory understanding” (Parkin, 1984, p. 20). In basic terms, this equates to examples of understanding a person’s mood by reading their face, e.g. anger, then exploring why they are angry (Parkin, 1984, p. 20). Context and wider knowledge enables us to explore social phenomenon and behaviour in the study of humans, and a researcher’s use of qualitative methods is endlessly creative, with the participant teaching the researcher about their lives, their political perspectives, and why they believe certain things (Bryman, 2004, p. 30). This was essential as I wanted participants who disagreed with me to challenge me, enabling us all to critically think about our beliefs. It is worth noting that quantitative research methods and the use of statistical data is also incredibly sophisticated and contributes to much larger scale studies on human interactions and behaviours. But with fixed design techniques Weber would argue the use of this data would always need to be backed up by Verstehen, my research design takes this position: it is only possible to explore social constructs or research people’s motives by asking participants what is going on and allowing them to lead the narrative (Parkin, 1984, pp. 21-22).


Research Questions:

  • Who has felt silenced during discussions around gender reform and policy change?
  • What is the potential impact on single sex services for females who have experienced male violence?
  • What are the views of those in support of gender reform about transwomen accessing single sex services for victims of male violence?
  • Is there a solution or middle ground for service provision in the MVAW sector?

Research Aim
The aim was to understand whether the silencing of feminist discourse regarding transgender ideologists’ proposed gender reform in legislation and policy capture, could impact, or has impacted, female-only services for victims of male violence.

Research Objectives
The objectives of this research were:

  • To provide a critical analysis of the silencing of feminist discourse on the proposed changes to the GRA 2004 focusing on MVAW sector services and spaces.
  • To explore/investigate the policy capture of transgender ideology, with a focus on the potential impacts on, and consequences for, female-only services for victims of male violence.

Research Design
The methodology consisted of addressing the research questions via:

  • (Semi-structured Interviews) – Analysing, via interviews, who has felt silenced during discussions around gender reform and policy change? What is the potential impact on single-sex services for females who have experienced male violence? What are the views of those in support of gender reform about transwomen accessing single sex services for victims of male violence?
  • (Online ethnography/ netnography – nonparticipant observation of social media discussions – via Twitter) Analysing, via social media, the discourses around gender reform and the inclusion of transwomen in female-only spaces, from both sides of the debate.

The use of multiple sources to enhance the rigour of research data is commonly referred to as triangulation (Robson, 2002, pp.174, 371). Each part of the study addressed the aims and objectives of the research as will be discussed in more detail later in the chapter. Figure 6 evidences how the data was triangulated into the research question:

Figure 6 Data Triangulation Method

Semi-structured Interviews

The main data was collected via confidential semi-structured qualitative interviews with 31 consenting participants across the UK, with most interviews taking place over the phone.

Interviewing is a widely used technique in social research with the categories falling into three distinct groups: structured, semi-structured and unstructured (Robson, 2009, p. 269). The nature of this research lent itself well to semi-structured interviews, this framework allows the researcher to have some control whilst enabling a degree of flexibility during the interview process, and it was useful to have a set of guidance questions as some participants asked to see these prior to interview. Unlike structured interviewing where questions are pre-set and ordered ahead of the involvement with the participant (Robson, 2009, p. 270), semi-structured interviews can be modified during the interview process. Powney and Watts (1987) refer to structured and semi-structured interviews as respondent interviews, they deduce that structured interviews are controlled by the interviewer (Robson, 2009, p. 271) but I wanted all participants to feel control in the process.

The use of semi-structured interviews enabled direct interaction with the participants to observe and/or listen to their behaviour, then subsequently if needed, there was an opportunity to explore new lines of enquiry. Furthermore, these observations created adjustability of the questions for future interviews (Burgelman,1985). For example, I re-worded question 3 in section 2 of the guidance questions from:

‘Do you feel the government has fostered a healthy debate around the GRA?’


‘Do you feel the government and politicians have fostered a healthy debate around the GRA?’

It was necessary to amend the question because early on the answers were broader than the question allowed. More generally the lack of structure in the interviewing process enabled new lines of enquiry to develop naturally from the participants’ responses (Bryman, 2004, p. 138). The interviews took the form of a conversational tone with participants, this would have been almost impossible to achieve if a strict code of fixed design was employed; with a survey no deviation is allowed from the questions, but in interviews the focus is placed firmly on the subjectivity of participants (Bryman, 2004). For the purposes of this research the aim was to galvanise the views of both sides of a debate that has encompassed fears of retribution or silence, therefore the use of qualitative research fostered trust by enabling participants to be honest, whilst protecting their anonymity.

The idea of a participant being in control with an uninterrupted, informal conversation reflects the types of feminist research undertaken for the MVAW sector (Skinner, Hester & Malos, 2005, p. 38). Women’s stories and feminist research heavily influenced me as a practitioner, and I deliberately centred female victims in section four of the interviews (see Appendix C). Participants were encouraged to explore the impact of transgender-inclusive policies on female victims and how they may respond to males in their safe spaces. The qualitative research design was of real advantage in this section as I was able to develop robust, challenging and honest conversations with participants, modifying the order, or adding and omitting aspects to enable the conversation to flow (Robson, 2009, p. 270). 

No matter the choice of methodology, results data is “neither true nor false, they can only be more or less useful” (Silverman, 1993, p. 2), and there are limitations to interviewing. Interviewing is time-consuming and requires careful preparation, and subsequent transcribing (Robson, 2002, p. 273), it was necessary to take into consideration planning and time budgeting. This was my experience as the length of time from preparation of interviewing, engaging participants and transcribing took over a year. Criticisms of interviewing from a positivist perspective is bias, this was of particular relevance in my research as the personal qualities of the interviewer are directly linked to the validity and reliability of the study findings (Salazar, 1990, p. 569; Silverman, 1983, p. 96). Consideration was given to the types of questions asked in the interview process, as there were dangers of the participant giving responses that are more socially desirable to the researcher, and this had to be acknowledged (Salazar, 1990, p. 569).

The other options available to me were surveying which could have offered a distance, transparency and accountability between myself and participants (Hakim, 1987), and could be seen to have a more scientific ring of confidence (Robson, 2002, p. 230). In addition, a survey could yield more responses, and the process of interviews is criticised by some traditionalists as a transaction that is a minor research technique, which should only be used to launch into serious sampling when counting begins (Silverman, 1983, p. 20). Ultimately a survey was not used for two main reasons. Firstly, by removing myself from the participants this could equate to a reliance on naturalism and thus deny my involvement as a feminist researcher, I felt strongly this approach would be dishonest and agree with Stanley & Wise (1993), that positivist research does not provide any more ‘truth’ by denying positionality (Stanley & Wise, 1993, p. 159).  Secondly, the reality is that researchers are human beings and our consciousness is the medium through which research is done, no matter the design (Stanley & Wise, 1993, p. 157), and in maintaining the central tenet of feminist research done by women for women (Stanley & Wise, 1993, p. 25), I opted for centring and being honest about my position, with a heavy focus on reflection, to mitigate any bias which the interview process can be accused of.

This part of the research project incorporated the use of qualitative semi-structured interviews with thirty-one participants, sixteen feminist activist women who are opposed to gender reform, and fifteen people who are in favour of gender reform. The inclusion and exclusion criteria were as follows:

  • Inclusion criteria – 30, participants, 15 from each side (I ended up interviewing 16 women from the feminist cohort). The participants comprised a mixture of active feminists who campaign against the proposed changes of the GRA 2004, and trans-inclusive women’s services – I named this cohort ‘Feminist Women’. On the other side, I contacted activists who campaign for gender reform and or supported trans-inclusive women’s services – I named this cohort ‘Pro Self-ID’.
  • Exclusion criteria – Anyone who has not campaigned or supported either side of the debate.

The preferred option of interview was face-to-face, however, due to the geographical reach of the participants spanning the UK, the majority of the interviews were conducted on the phone, I did not foresee many issues with the difference, as face-to-face and telephone interviews both have similar advantages (Robson, 2002, p. 282). Although I had to keep a close eye on rapport building, the advantages to the telephone interviews included ease and cost effectiveness and participants also tend not to be drawn into more socially desirable responses as they are with face-to-face interviewing (Robson, 2002, p. 282). The interviews lasted approximately 45-60 minutes, and, with the prior consent of the participants, all the interviews were recorded, I transcribed the interviews, then analysed the data using Nvivo.

Some considerable thought was given as to whether to ask the two opposing groups of participants different questions. However, in order to offer transparency, it was necessary to have just one set of questions for both groups (See Appendix C). Rejecting the positivist argument that pre-tested standardised questions offer reliability to the research, the questions remained purely guidance, and I preferred the interactionist value by allowing the interviewees to actively construct their world view, thereby generating authentic data (Silverman, 1993, p. 91).

Given the topical nature of the research I had an advantage as a practitioner researcher in the field and, as anticipated, the sample size for feminist activists was easy to fill. On the flip side, because of my position in the research I anticipated that the recruitment of participants from the pro self-ID cohort may be harder to fulfil, and this prediction was correct. I continued to open dialogue with these participants throughout the research phase, explained my position and provided reassurance that this project is aimed at understanding the position from both sides of the debate. As there were limited participants willing to speak to me directly from trans lobby groups I also contacted the feminist organisations who had publicly agreed with trans-inclusive spaces in the MVAW sector by contacting the twelve MVAW organisations mentioned in the Stonewall research (Stonewall & NfpSynergy, 2018). Subsequently by utilising snowball sampling (Robson, 2002, pp. 265-266), where initial participants are then asked if they can identify other willing participants, I was able to fulfil the requirement of fifteen participants from the pro self-ID cohort of the debate. This technique was incredibly useful as snowballing is advantageous in identifying members that could be deemed to be a “clandestine group” (Robson, 2002, pp. 265-266).

The road to fulfilling the required number of participants for the pro Self-ID cohort was bumpy, with one participant withdrawing a day prior to the interview. Their reasoning was that they felt they were unable to take part due to the title of my research, as they felt I was biased and had already made up my mind. With the support of my supervisor, I decided not to take out the title in the introductory emails or the participant information sheet; having been through full ethical approval I felt confident about my position and I did not want to be duplicitous, or leave participants feeling deceived. The title and my positioning were obviously off-putting as I contacted a total combination of thirty-four people/organisations on the pro self-ID cohort and only sixteen responded, with one, as stated, dropping out of the process.

Data collection
All potential participants were sent a letter of invitation and a participant information sheet (see Appendix D). When participants agreed to take part in the research, they were sent a consent form (see Appendix D) and were asked to sign this prior to the interview taking place. The interview data was recorded on a Dictaphone and held securely on my personal computer and the University of Portsmouth Google drive, both of which are password protected. All transcripts of the recordings were undertaken by me and held securely on the systems named above.

Three of the feminist participants asked me to pause the recording during the interview process, all of them were anxious about repercussions of being targeted should some of the conversations we had ‘off tape’ be published. In addition, the pro-self ID participants who were concerned about their position as frontline workers in the MVAW sector were incredibly anxious, and one participant requested the questions ahead of the interview, as she was nervous about the toxicity of the topic. The one participant, previously mentioned, who dropped out the day before the interview, did so as they decided I was ‘biased’. This left me feeling a mixture of angry and sad, which I reflected in my research diary entry (08/05/19 – Appendix E). I was frustrated as I had hoped they would be open to discuss the topic, and I was sad that this was another example of the debate being shut down. They wrote to me explaining that they spent a long time reflecting on whether to participate and decided against it because they had not been able to reconcile the study’s framing as the title suggested I had already made up my mind. This unnerved me somewhat as it was early on in my data collection process and the person in question is relatively influential in the MVAW sector. After my initial upset subsided, I realised they did have a point. I have taken an ideological position and my passionate views on the topic combined with my long-standing career are the reason I undertook the research. It is obvious that my positionality put others off too, as I contacted twice as many potential participants from the pro-self ID cohort than the feminist cohort. However, with the patient guidance of my supervisor I reconciled that, although I cannot change my position or my thoughts on the topic, I could maintain a faithful objective to the integrity of the data. On reflection the pro self-ID cohort were keen to be part of the conversation, some of whom were wholly opposed to my viewpoint and I was grateful for their time. One interview with a pro self ID participant is an experience I hold very dear to me, they showed real care about how we could resolve the issues together, and this gave me hope that despite differences, respectful, considered debate can occur. 

Online Ethnography

The other part of the research took place using the online ethnographic method. Kozinets (2015) defines the use of online ethnography as “Netnographic” research which uses social science methods to present a new approach to ethical ethnographic research utilising digital media to analyse critiques of communities and culture (Kozinets, 2015, p. 1). The purpose of the ethnographic data collection was to explore three core areas:

  1. An examination of the online discourses and debates of feminist activist and trans lobby groups regarding gender reform and transgender policy capture.
  1. Identification and critical review of any discourses and debates that make a concerted effort to silence and or discredit feminist discourse.
  1. Critically explore and analyse the framework of contextual constructionism, by applying the claimsmaking activities within the debate. Thus, establishing the nature of the arguments being put forward, to gain a clearer understanding of any impact on the MVAW sector.

The list of the websites and social media platforms used in the initial scoping of the research is in Appendix B.

There were clear advantages to the online ethnographic approach, as I became passive online and set up a different Twitter account. The anonymity and passivity afforded me some safety and comfort, particularly as I had previously been targeted, and it also allowed me to move past the block lists that included my own Twitter account, which mean I am unable to see what many trans activists are saying – my anonymous researcher’s account had no such restrictions. Online ethnography has clear advantages over traditional ethnographic research in terms of availability of data and the less time-consuming, less costly nature of it (Whalen, 2017, p. 1).

Ethnographic research is a well-established qualitative research method that requires the long-term observations of a culture from the perspective of its members (Toledano, 2017, p. 598); to that end netnography approaches the same immersive nature but takes place online. In fact, the advantage to the online element to this type of ethnography meant that I was able to collect data over a long period of time, (approximately 18 months) which is the central feature of traditional ethnographic research (Robson, 2009, p. 186). Being a part-time doctoral student, with a full-time job, the option of being part of trans rights activists’ communities in a real-life scenario was not practical, it is also not possible to understand if online communities exist in any comparative sense offline. Although I could have joined protests with activists, it was highly unlikely I would be accepted into those communities especially as I would not want to be covert about my position. 

There are limitations to netnography: a tweet is only a snapshot in time; as a researcher I do not witness the context; I knew nothing else about those people, other than their online persona and presence; and I could not attribute any other meaning to their words. It is widely accepted that most people behave differently online to their behaviour in real life, the perception of anonymity gives users a false sense of protection, without fear of consequences (Johnson, 1997, p. 61-62). In addition, the fast-moving world of social media discourse means that the data collected could be deemed time limited and what appeared to be the most pertinent discourse at point of data collection, may lack relevance at point of publication of my thesis. However, even with these limitations in mind I was able to utilise the online ethnographic research to combine the themes emerging from the interview data.

Although there were limitations regarding the online ethnographic data, the process of sampling was relatively easy: by following and utilising the discourse over a period of eighteen months I was able to include the conversations occurring online in relation to the topic. The discussions regarding gender reform were numerous and expanded into many areas, including the impact reforms and transition could have on children and the silencing of academic enquiry. In order to reflect the discourse, I did capture some of these conversations, however, I chose to focus my sampling on the direct impact of gender reform in policy and legislation on victims of MVAW. 

Data collection
At the point of analysis for the online ethnographic research it became apparent that the combination of Twitter data rules and Nvivo software made the job of downloading information more complex. Although I had already collected over two thousand tweets and ‘bookmarked’ them in my own social media account I realised that retweeting them on my platform would constitute participating in the research which would have extinguished my non-participatory observer position. As a result, I set up a separate account, which remained ‘unfollowed’ and ‘locked’ and I then imported the full Twitter account into Nvivo.

At points I nearly gave up on the online ethnography entirely! Firstly, the realisation of having to set up a separate account and start again felt horrendous but on reflection, the experience ended up focusing and separating me from the research. There was a boundary between my personal Twitter account and my data, and I enjoyed the differentiation of these positions. I gave up quite quickly on using any other platform than Twitter, largely because the most useful conversations were occurring there, and as the process resulted in 2,022 tweets, I am satisfied this was the right decision. I had issues with blocklists on my personal account and I went to my research account to follow threads, but as I had chosen not to follow anyone on this account, I undoubtedly missed discourse that could have proved useful. But I accept the limitations in being a lone researcher and I knew early on that I would be capturing only a small fraction of the debates.

When it came to importing my Twitter data into Nvivo I went through another long, drawn out process with stumbling blocks. Again, I nearly gave up, but thanks to my supervisor and to the Nvivo support team the process ended up being a relatively simple one (See Appendix F). I am pleased I didn’t give up; I feel the online ethnography adds a richness to the thesis and the combination of the interview and online ethnographic data is represented in my results.

Data Analysis (Interviews and Online Ethnographic Data)

Information from the data was analysed using a coding approach (Braun & Clarke, 2006) and processed through NVivo software. This analysis was built to include categories and a coding system of the most prominent themes. The research analysis utlised Braun and Clarke’s (2006) six phase framework for thematic analysis:

  • Step one – Become familiar with the data – I read the interview transcripts, several times.
  • Step two – generating initial codes – I organised the data into smaller chunks, coding them.
  • Step three – Searching for themes – I identified significant data.
  • Step four – review the themes – at this stage I modified and developed the preliminary themes.
  • Step five – define themes – I identified the meaning of each theme and analysed how the themes interacted with each other (including the ethnographic data)
  • Step six – the findings were then written up for thesis submission (Maguire & Delahunt, 2017).

As Saldańa (2016) describes, the idea of “coding for themes” is not accurate, it should be that a “theme is an outcome of the coding”, I used ‘In vivo’ coding as the technique which meant the themes were led by the data (Saldańa, 2016, pp. 15, 105). The online ethnographic analysis was added to the nodes post interview coding and utilised as a tool to offer alternative or supportive evidence to the participant interviews by reflecting the simultaneous live discourse occurring over the research period. A section of the results contains voices from pro-self ID participants who were part of the Stonewall research report (Stonewall & NfpSynergy, 2018) which supports the notion of trans inclusion in the MVAW sector. Because these participants disagreed with some of the aspects of the Stonewall report and were fearful about this, I further anonymise their responses using a randomised numbering system, which is reflected in Chapter 6.

Given this was a sensitive research topic, participants were offered the opportunity to check the transcripts of interviews prior to analysis, one participant in the pro self-ID cohort did request this, and I was happy to accept their amendments to make them feel more comfortable and it did not change their valuable contribution. All the data was then analysed collectively using the coding system ‘In vivo’, described by Strauss (1987) as the “actual language the participants use themselves” (Strauss, 1987, p. 33), this is a simple process and allows the participants to ‘own’ the data. This measure felt congruent with the aims of the research in emboldening previously silenced voices to be heard, and I used the phrases spoken by participants as the coding titles. The result was that the participants voices led the analysis, at phase one and two of the coding process my job and presence in the research was detached as I methodically represented the participants voices (Saldanã, 2016, p. 105), it was through participant voice that the overarching title of the thesis emerged. 

The process of coding took me approximately a month, and I kept a research log which I exported into separate codebooks, all of which can be viewed in Appendixes F and G:

  • Round one – Codebook – 11/05/2020
  • Round two – Codebook – 16/05/2020
  • Round three – Codebook – 18/05/2020
  • Round four – Codebook – 29/05/2020
  • Round five – Codebook – 04/06/2020

Round one of coding resulted in 137 nodes. I then met with my supervisor who guided me through the second round of coding explaining the importance of applying more thought where the themes would start to reveal themselves. This had already been apparent, as in my research log represented below (See Appendix G), I noted the theme of silencing had been prevalent in the data, so I began with this as the first ‘parent node’:

  • 12/05/2020 –

Made a start with silencing as this became really obvious during the first round of coding.

Some group queries undertaken to test the visual aspect of the data. This was really useful in determining the category on silencing.

The second coding book (See Appendix F) resulted in nine main themes with child nodes logged underneath each main theme. All parent and child nodes were given descriptors and I began to run queries on the data. Quite quickly it became apparent that some of the child nodes needed ‘tidying up’, and I merged some of the nodes resulting in richer data as larger child nodes under parent themes. I then produced the round three codebook (See Appendix F). The fourth round of coding (Appendix F) included further analyses from participant responses into the question: ‘what is a woman?’ The fifth and final round of coding was undertaken when I imported all my online ethnographic data from Twitter. This resulted in a much larger data set and my coding log diary entry (See Appendix G) on 03/06/2020 notes:

  • 03/06/2020

All twitter finally on Nvivo. Managed to do this via Ncapture. The tweets are now coded to general tweets and then feeds organised. The most useful of which is hashtags and mentions. I have also now coded tweets to main nodes and re-done the code book. Things are starting to look really rich in terms of data. Just when I thought I was sifted I will need to sift again! But having too much data is a nice problem to have.

2022 tweets is a lot of data!

I spent some time analysing the data from the online ethnographic research and reflecting on how this supported the interview data and fit into the existing themes. Round five codebook, (Appendix F) represents this analysis and some of the mentions and hashtags in this codebook were then relabelled as ‘anonymous’ due to the twitter handles revealing the identity. Through the process of round three of coding, themes emerged, represented in table 1 below:

ThemeFiles AttributedReferences
The Male Violence Against Women Sector30311
The Proposed Changes to the Gender Recognition Act 200430216
Feminism, Women’s Rights and Activism30161
What is a Woman30152
Transgender Rights Movement27130
Middle Ground3094
Motivation for Access to Single Sex Exemptions (SSE)2664
What is a Woman – Question3036
Table 1 Thematic Analysis of Interviews Overarching nodes

The themes were analysed further at coding stages four and five, with some being reorganised as child nodes and others explored in more detail, resulting in further child nodes. At phase six of thematic analysis they were then collapsed into two main chapters as presented in Table 2 below, the discussion of the literature was combined throughout the results chapters:

5The Proposed Changes to the Gender Recognition Act 2004
5Why Campaign
5What is a Woman
6Motivation for Access to Single Sex Spaces
6Male Violence Against Women SectorCommissioning and Policy Capture
6Middle Ground and Third Space Option
Table 2 Themes and Subthemes Results chapters

Ethical Considerations
Members of both sides of the interview participant data were treated in a fair and balanced manner with their participation being voluntary. Due to the sensitive nature of the research topic, there was a potential risk to interviewees therefore all interview data was anonymised with all feminist participants listed under group A and all pro self-ID participants listed under group B.

The online ethnographic research was not deemed to be covert as although people may not be aware that they are under observation by a researcher, it is generally accepted that social media conversations are public and viewable (Sugiura, Wiles & Pope, 2016, p. 187).  However, it was important to note that Fossheim and Ingierd (2015) maintain a prominent concern when collecting social media data is whether or not the sources in question are considered public or private, and to what extent researchers are ethically bound to seek informed consent (Enjolras et al., 2015). Likewise, Boyd and Crawford (2012) argue that social media research cannot be justified as ethical solely because the data is seemingly public (Boyd & Crawford, 2012, pp. 662-279). Questions as to whether or not social media postings are public or private are determined to some extent by the online setting itself and whether there is a reasonable expectation of privacy on behalf of the user (British Psychological Society, 2016; BSA Social Media Guidelines, 2017). As my final research data utilised the social media platform Twitter, it was assumed that users of this platform had all agreed to the terms and conditions. Within these terms there are clauses detailing how an individual’s data may be accessed by third parties, including researchers (Hogan, 2008). However, social media research does present problems concerning the informed consent of participants (Wall & Williams, 2013), and it can be appealing for researchers to conflate individuals’ agreements to the terms of social media platforms, which include clauses on the accessing and reuse of data by third parties, as informed consent (Salmons, 2014). Despite ethical concerns, in most cases, a social media user’s data is sought, accessed, and analysed without informed consent and individuals are rarely aware of their participation in the research (Fossheim & Ingeird, 2015). However, in order to safeguard myself, and ensure the data I collected is legally considered public, I read through all the relevant terms and conditions on Twitter, to ensure that any public data could be accessed by me as a third party. Despite taking this precaution I am aware that it is not possible to determine to what extent any users of Twitter are aware of their agreement to the rules and regulations, so it would be difficult to constitute an agreement as informed consent. Therefore, in line with my ethical approval, no method was used to obtain anything other than the public status updates and I did not interact with any accounts or conversations, utilising a non-participant observer method and remaining passive online, which was incredibly advantageous (Robson, 2002, p. 318).

A further concern with online ethnographic research is the participant’s right to withdraw from the research, which is made more complex when analysing social media (Hogan, 2008). For example, when researching covertly it is reasonable to assume that if a user deletes a specific post, their account, or if an account is suspended, this equates to participant withdrawal and any data collected from these accounts was not used (Hogan, 2008, p. 141). As a researcher I am aware that analysing social media data will produce conditions in which I would be ethically bound to seek informed consent, such as when accessing private data, or the accounts of individuals who are under 18. In order to avoid such issues, I solely accessed information that was in the public domain and had been posted publicly on individuals’ profiles and did not access private conversations or closed group chats. Moreover, I avoided collecting data from the social media accounts of individuals who were under 18, although it could be difficult to prove individual ages, I used judgement and initiative when I came across certain accounts and did not include them in the data analysis.

Anonymity is a key factor for consideration in using online ethnography and Boyd & Crawford (2012) outline that the ability for researchers to anonymise data collected from social media platforms is increasingly complex, particularly when anonymising individual data such as tweets (Boyd & Crawford, 2012, pp. 662-679). Additionally, Fossheim and Ingreid (2015) highlight that media companies store their data for extended periods of time and this data is searchable, making the anonymisation of secondary sources more challenging for researchers (Enjolras et al., 2015). Therefore, anonymity was a key ethical implication in this research as the identity of unwitting participants became more critical when the data accessed referred to such a sensitive subject. There was a particular concern regarding the inclusion of quotes that had been taken from Twitter and published verbatim, as this could expose the identity and profile of the user. To safeguard against this, I ensured all Twitter handles and identities were removed, and it was easier for me to follow hashtags on Twitter to show a range of claims, rather than identify specific accounts. Moreover, in order to comply with the General Data Protection Regulations (2018), I ensured that no personal data, including names, locations or physical attributes were documented and traceable information was anonymised and struck from analysis (Regulation (EU) 2016/679 of the European Parliament and of the Council, 2016).   

Due to my position as a researcher and the nature of the topic there were several ethical risks to consider and I had to be conscious of my previous experience from online trans activists. Both online harassment and social shaming were something I considered in depth with my supervisor (Marwick, Backwell & Lo, 2016) and with his support I made the decision, for the period of my research phase, to take the role of non-participant observer which meant I simply observed the discourse for ethnographic purposes; this kept me safe from further online harassment and also enabled me to gain as much objectivity and impartiality as possible (Robson, 2002, p. 318).

Interview participants were treated in a fair and balanced manner with their participation being voluntary and all interview data was anonymised. As far as possible no assumptions were made concerning the views, motivations, and demographic characteristics of participants, although researcher bias is always a potential issue. An open-ended interview methodology was used with straight forward language to reduce paternalism by the researcher. Participants were provided with all the necessary information and consent forms and were given the opportunity to withdraw from the interview process with the time limitation on withdrawal detailed on the participant information sheet (Appendix C).

No financial incentives were offered to potential participants, but there were still potential benefits to participants, as the research enabled participants to have a voice in a more established setting, i.e. the academic environment. The groups may have had media exposure and run web and social media sites and so are airing their activities and views with the general online population, but this research gives them the opportunity to have their concerns discussed in a more structured setting.

I hope the methodology undertaken provides a valuable basis for this original contribution to a highly sensitive topic and the results of the research illuminate the previously unknown views from the MVAW sector. The next two chapters provide the results of my analysis.