because they would be afraid to speak to an outsider. She proposed that I started with one of the women, who was on friendly terms with the other women and who also spoke Greek. This way she would go back to the cells and vouch for me after leaving the interview. While I was hesitant, Xenia was correct, and my discussion with this first woman indeed had a snowballing effect; an effect that an ethics protocol would not have accepted. As Alice Gerlach (2018) describes in her thesis about her research inside a women’s detention centre in the UK, after the first day of interviews, women were asking to participate voluntarily. Some of the women also asked to bring their friends too so in some cases I interviewed them together. It quicky became apparent that many women were interested in participating in the project. While the police’s office might have distressed the women (at first at least), for the police officers this was a relaxing environment. In contrast to the detained women, the officers were open about their views from the start, the time they served at the border and their roles as detention officers in the Petrou Ralli detention centre. In fact, at times I wished they had not been so direct. Having to listen to their comments about the dangerousness of immigrants, failing cultures and even explicit stories about violence against those detained, was emotionally very difficult. Interviews were semi-structured but often the shocking stories of women’s journeys on their way to Greece, as the one described in the introduction guided the discussion. I concluded interviews by giving women a debrief form that included contact details for mental health organisations. I sought to build a representative sample of sorts, by speaking to women from different countries. However, in the absence of an interpreter, my discussions with Iraqi women, for example, who did not speak much Greek or English, were very short. Some women brought a friend with them, who could translate. I even put my language skills to use as I had some interviews with women from the Dominican Republic in Spanish. They were digitally recorded and then fully transcribed, along with my fieldwork diaries, and translated into English. In cases where our discussions were not recorded (two detention officers did not want to be recorded), I took written notes during the meeting and afterwards recorded my recollections of the person’s testimony as fully as possible. I attributed pseudonyms to participants and created an excel file with basic demographic characteristics, where applicable. 49