Time series analysis is commonly applied to both chaotic and stochastic systems, which are collectively described as turbulence. However, explicitly time-dependent non-autonomous systems can also generate turbulent dynamics, which makes them useful for describing many physical phenomena. Nevertheless, many of the methods used to analyse turbulence are based around autonomous systems. In this paper, time series from the chaotic, stochastic and non-autonomous Duffing system are analysed using these methods to gauge their suitability to non-autonomous systems. It is found that time-dependent representations are vitally important in the study of this class of systems. Moreover, when time-dependence is neglected in the representation a completely deterministic non-autonomous system is often indistinguishable from a stochastic system.