In this paper I present the development of a data analysis method that I came to name as one of 'pulling up cracks'. I first define the concept of a crack drawing on the work of Deleuze and locate it within a body of feminist scholarship arguing for the disruption of the taken for granted categories in qualitative inquiry, including voice and representation. The paper will then outline a series of encounters that led to a disturbance of my ground as a researcher as I faced the (im)possibility of (re)presenting the voice of participants. This act of disturbance was itself the opening up of a crack where thought was jolted and disturbed. The opening up of a crack afforded me the possibility of wrestling with the politics and ethics of doing research, and in particular, how to engage with the task of doing data analysis differently.