This chapter discusses the methodological challenges associated with studying personal background variables in first language (L1) attrition from the perspective of Complex Dynamic Systems Theory (CDST). It starts with a review of extant research which, despite concerted efforts to design rigorous, comparable studies, to date has not turned up strong, unambiguous predictors for L1 attrition. I argue that this failure lies in the nature of language as a complex dynamic system, and consequently in the properties of variables, their interaction, and varying contribution to the process and outcome of L1 attrition, and indeed to L1 and L2 (second language) acquisition in the larger context of multilingual development. CDST provides a challenge not just for common empirical and analytical approaches to attrition, but for the very notion of ‘predictor’. The chapter concludes by discussing some ways in which the current stalemate may be overcome.