Past DA studies have investigated how people maintain paranormal belief (e.g. Lamont, 2007b, Wooffitt, 1992) as well as a position of disbelief (Lamont et al, 2009); however no current research has explored how people maintain paranormal belief specifically in relation to fraud. The current study’s objective was to use DA in exploring how individuals express and maintain a paranormal belief, whilst discursively overcoming the issue of fraud. Four common themes emerged from the data, which were all naturally occurring texts; 1. Framing ‘failure’ as supportive of belief, not fraud, 2. Acceptance that some psychics are fake, and even genuine ones cheat; but not all the time, 3. Drawing boundaries between real and fake, and 4. Reference to more reliable evidence which supported the existence of paranormal phenomena. The discursive practices used in these themes are in line with other relevant DA literature; the use of extreme case formulations (Pomerantz, 1986), avowals of prior scepticism (Lamont, 2007b), consensus (Edwards & Potter, 1992), the use of framing (Goffman, 1974), three-part lists (Jefferson, 1990) and other forms of fact construction including stake and interest, minimisation, and maximisation were used frequently throughout the analysed data (Edwards & Potter, 1992; Potter, 1996).