Commentary on van der Pol et al. (2014): Reconsidering the association between cannabis exposure and dependenceThe existence of a positive, and relatively strong, association between cannabis exposure/tetrahydrocannabinol (THC) dose and cannabis dependence is an assumption on which much cannabis-related research, public policy and health advice has been based. Van der Pol et al.'s [1] findings, however, throw this into some doubt. To understand why there is a discrepancy here, and its import, it is essential to consider how cannabis exposure/dose is typically measured and, thereby, how past studies may have clouded our understanding of the association between cannabis use and dependence.Cannabis exposure has been assessed in a variety of ways, tending to include measurement of some combination of age at first use, duration of use, quantity consumed and frequency of use, with the latter often employed in isolation as a proxy for THC dose [2]. Although known to be only approximations of exposure, these variables are used by researchers for a number of reasons. Foremost among these are the efficiencies associated with the use of self-report questionnaires, which are often completed anonymously by participants and remotely from researchers. The classification of cannabis as an illicit substance within most jurisdictions globally, however, also acts to prevent the vast majority of cannabis researchers from undertaking as thorough an assessment of cannabis exposure/dose as has been demonstrated by van der Pol et al. [1].In the relatively small number of studies where THC dose has been confirmed, this has generally been via 'cannabis cigarettes' purchased from the US National Institute on Drug Abuse (NIDA). These studies, however, do not replicate normal usage of cannabis, such as that assessed in van der Pol et al.'s study. Specifically, NIDA joints range from 3 to 7% in THC content and are typically smoked in a prescribed manner (i.e. uniform puff size/duration and breath hold) (e.g. [3]); in comparison, van der Pol et al.'s participants were using cannabis ranging from 3.6 to 15.7% THC and they smoked it as per usual, such that their smoking topographies varied. Taken together with the finding that total number of puffs and puff volume were significant predictors of dependence in van der Pol et al.'s study, these differences throw at least some doubt on the 'real world' validity of findings from previous studies (whether using real or proxy measures of exposure/dose) and illustrate how important naturalistic studies are to increasing our understanding of the acute and long-term effects of cannabis use, including dependence.Of particular interest here is the 'typical' smoking topography that van der Pol et al. [1] found was predictive of cannabis dependence. Taking more frequent puffs of greater volume when commencing a session is evocative of someone seeking a quick hit, such as to 'take the edge off', then smoking with less urgency as the desired effects kick in. As such, the 'typical' smoking topography may be i...