Commentary on Hobday et al. (2015): Inconsistent results beneath consistent conclusions-the need for a new approach to analysing alcohol availability Alcohol availability studies keep producing unexpected findings alongside results that are in line with previous research. Such contradictory results should be seen as an opportunity for a renewed focus on theory development.More off-premise outlets are associated with fewer alcoholrelated injuries. Bar sales do not affect alcohol-related injury admissions. Areas with higher young male populations have lower levels of alcohol-related injury. These are some of the findings which do not feature in the conclusions of Hobday et al.'s well-executed analysis of the relationship between alcohol availability and alcohol-related injury [1]. What follows is not a criticism of the authors' work, which is good quality, offers useful findings on extended trading permits and merits its publication. Instead, we offer a reflection on a case study which is symptomatic of persistent problems in this area of alcohol policy analysis.Hobday et al. find evidence that the number of alcohol outlets in an area (in this case on-trade outlets and particularly those with extended trading permits) is associated positively with an outcome of interest to public health (in this case, night-time injury admissions to emergency departments). What the conclusions do not tell us is that, like many researchers before them, the authors' results tell an inconsistent story and, at times, are plain contrary [2]. From a policy perspective, the persistent identification of inconsistent findings is challenging and difficult to accommodate within compelling advocacy messages. It is easy to see why they are neglected. However, scientists should see such inconsistencies as a positive: they raise questions to be answered and a spur for new thinking.To their credit, Hobday et al. seek to examine their unexpected results, but their arguments are not always convincing. The economies of scale afforded by large off-trade outlets will probably prompt higher sales, but this does not explain why, after controlling for sales per outlet, a reduction in the number of off-trade outlets is associated with an increase in alcohol-related injuries. Perhaps drinkers are switching to the on-trade, a sector associated more closely with acute alcohol-related harm, but this explanation leads us to the puzzling finding that on-trade sales are unrelated to alcohol-related injury admissions. It is reasonable to conclude that increased on-trade outlet numbers are associated with increased injuries because the drawing together of crowds presents more opportunities for injuries to occur; however, public health researchers and advocates have long argued that intoxication levels are a partial cause of alcohol-related violence. Perhaps the lack of a relationship between on-trade sales per outlet and alcohol-related injuries is due to high sales outlets having more customers, rather than more intoxication. It may be that sales per outlet is a ...