The behaviour of locusts has been studied extensively using two approaches: (1) analysing a single individual's response to a group stimulus or (2) using group conditions to look at aggregation patterns. The second approach has, in contrast with the first one, not been improved in terms of statistical analyses since the 1960s. In the present study, we propose a spatial statistics approach of point-pattern analysis to improve the group-based assessment of behavioural phase characterization. This diagnostic tool was developed and tested in the laboratory with comparative analysis of solitarious (isolation-reared) and gregarious (crowd-reared) desert locusts, Schistocerca gregaria (Forskål) (Orthoptera: Acrididae). The spatial distribution patterns of 10 either solitarious or gregarious third-instar hoppers were characterized with nearest neighbour distance measurements in a circular arena. The temporal sequence of spatial disposition of locusts was recorded with a digital camera taking photographs at regular intervals. The approach of point-pattern analysis focused on the spatial distribution of observed events and allowed us to make inferences about the underlying process that generated them. The results confirmed that our diagnostic tool could identify that crowd-reared hoppers tended to aggregate more to conspecifics than isolation-reared ones. We could also verify that isolation-reared hoppers were less active than crowd-reared ones, but this was only true at the beginning of the experiments. The spatial statistics approach proposed in the present study could help with observations of phase-related differences in the behaviour of locusts.