The main finding of this study was that divergent thinking as indexed by the verbal fluency task was enhanced acutely by cannabis in individuals who were low in trait creativity. Their performance after smoking cannabis improved to the same level as the high creativity group. When not intoxicated, the low creativity group’s performance was significantly worse than the high creativity group. On the category fluency task, the high creativity group significantly outperformed the low creativity group on both testing days so performance was unaffected by cannabis use. The high creativity group had significantly higher SPQ scores than the low creativity group. State schizotypy, as measured by PSI scores, increased significantly for both groups on the intoxicated day, replicating Mason et al. (2009).

The increase in verbal fluency for those low in trait creativity on the intoxicated day supports the notion that cannabis can enhance aspects of creativity (Eisenman et al., 1980). Our findings in a naturalistic study replicate those of a controlled laboratory study in which oral doses (7.5, 15 mg) of THC or matched placebo were administered using double-blind procedures (Curran et al., 2002). It is possible that the high creativity group was performing at ceiling levels on both test days and thus could not increase scores when intoxicated. This possibility is less likely given that the increase in fluency after THC observed by Curran et al. was in the same participants tested over three separate days after low dose, high-dose THC and placebo. Curran et al. suggested that the enhanced verbal fluency following THC reflected some sort of disinhibition of frontal cortical functions which is consistent with cannabis’ known neurobiological effects. This disinhibition may already be present in the high creativity group when not intoxicated and they may already be reaching maximum performance levels that do not change under the influence of cannabis. In the neurobiological model of creativity (Flaherty, 2005), enhanced frontal functioning is thought to play a key role. The observed increase in verbal fluency may be related to the finding that the right hemisphere appears to be involved in recognising remote associates (Rdel, Cook, Regard, & Landis, 1992), and that cannabis acutely enhances right hemispheric activity (Mathew, Wilson, & Turkington, 2002).

In contrast to the verbal fluency, categorical fluency scores on each day did not change significantly within groups. On both test days, the high creativity group scored higher than the low creativity group on category fluency. This difference may be due to the different aspects of divergent thinking the two tasks are tapping. While verbal fluency relies on the phonological cues for word retrieval, categorical fluency relies on semantic associates for retrieval. It has been shown that these two tasks are mediated by different areas of the brain, with verbal fluency linked to the frontal cortex and categorical fluency linked to the temporal cortex (Baldo, Schwartz, Wilkins, & Dronkers, 2006). The high creativity group may have enhanced functioning in the temporal cortex in comparison to the low creativity group and this may not be affected by cannabis use. Acutely, cannabis may also stimulate dopamine release in the mesolimbic pathway which includes the frontal cortex, possibly related to the increase in verbal fluency (Caspi, Moffitt, & Cannon, 2005).

The finding of a significant trait schizotypy difference between groups is in line with Rawlings and Locarnini’s (2008) data strongly linking creativity with schizotypy. State schizotypy (PSI) scores increased significantly on the intoxicated day for both groups, replicating previous findings (Mason et al., 2009). Over days, the increase in PSI in the low creativity group does not appear to be related to their increased verbal fluency as there was no significant correlation between changes in these scores.

The results of the analyses on RAT scores found a trend for a between groups difference, with the high creativity group performing better than the low creativity group. A trend for a decrease in performance following acute cannabis in the high creativity group was also found. While it does not appear that divergent thinking, as measured by verbal fluency is hindered by cannabis use, it may be possible that convergent thinking is affected, leading to the slightly decreased RAT scores when intoxicated. This may have important implications for the possible relationship between cannabis use and creative thinking as whilst divergent thinking encompasses the essence of creativity by allowing for loose associations to be formed between concepts (Reber, 1995), minimal convergent thinking or creative problem solving may hinder the ability to form connections between these ideas. It appears that the level of convergent thinking may largely determine the degree to which one’s thinking is considered either creative (i.e. increased convergent thinking) or psychotic (i.e. decreased convergent thinking), as well as the associated productivity. It is possible the increase in divergent thinking found in the low creativity group may lead to enhanced creative thought, but may not lead to convergence on a specific creative idea. As well as this, there is strong evidence that cannabis impairs episodic memory (e.g. Curran et al., 2002) which may prevent a person from remembering any creative ideas they may have had whilst under the influence of cannabis.

Despite the benefits of a more ecologically valid experiment through cannabis being used in naturalistic settings, issues regarding reliability must be addressed. As creativity is ill-defined and thus has a large subjective component, the verbal fluency task and RAT were used as comparatively easy-to-score measures compared to others used in the literature, in order to ensure high inter-rater reliability. In a number of definitions of creativity, originality appears to be a key inclusive feature. The RAT does not allow for original responses, as there is only one target answer for each of the three-word sets. Whilst the verbal fluency task allows participants more freedom to give a range of responses, the measure is still only based on the quantity of responses. However, Gilhooly, Fioratou, Anthony, and Wynn (2007) found that verbal fluency is related to the production of the number of ‘new’ uses participants came up with on an alternative uses task, with a positive correlation between verbal fluency scores and the number of new uses participants can think of. They hypothesised that this may be due to divergent thinking involving the employment of executive processes, as happens in verbal fluency. In a future related study to support that the tasks did in fact tap into aspects of creativity – not unconnected aspects of cognitive performance – a task involving a measure of originality would be fundamental. Block et al.’s (1992) methodology of analysing original responses by looking at those considered statistically infrequent may be an appropriate alternative.

Another limitation relates to the reductionist approach of measuring creativity using word-based tasks alone. This fails to encompass the multi-faceted nature of creativity as outlined by Guilford (1959). This is highlighted when comparing the currently used state creativity measures to the various aspects of creativity measured in the CAQ. Nevertheless, issues regarding the CAQ stem from its focus on nationally recognised works of creative achievement. Given that the current sample was aged between 18 and 24, a relatively young test group, it may be unrealistic for any potential works of artistic merit to have achieved public recognition; the use of recognised achievements, talents or prizes as a way of scoring creative achievements may not pick up on individuals who may still be creative but may not have received the appropriate recognition to obtain a high CAQ score. Therefore the validity of using the CAQ as a measure of innate creativity for younger individuals may be questionable. However, this has not prevented other studies from using the CAQ on similar populations. Carson, Peterson, and Higgins (2003) used a group with a mean age of 20.7 years in their study on creativity, intelligence and latent inhibition. In the study, they compared eminent creative achievers (those under the age of 21 who reported unusually high scores in specific domains of creative achievement) with those low in creative achievement. While we only looked at the top and bottom quartiles and not at eminent creative achievers, it may be possible higher creativity scores at a young age may still be indicative of trait creativity.