In this paper we compare the sketch-map technique with a spatial modelling task based on a set of eight separate items. The aim of the comparison is to verify, thanks to a special spatial reconstruction set, Jeu de reconstruction spatiale (JRS), whether the difficulties often encountered by unskilled persons when asked to draw a sketch map can be avoided. If so, the JRS could help to improve comparisons between different social groups. Thirty university students majoring in geography and twenty-one unskilled staff members were asked to represent the centre of Strasbourg both by drawing a sketch and with the JRS. The results show that, on one hand, the difference between the two groups regarding the number of items mentioned is smaller with the JRS than with the sketch map. On the other hand, in both groups, the same proportion of respondents increased the number of urban elements produced in the second trial with the JRS. This was not the case with the sketch map. Also, the JRS is more stable than the sketch map from one trial to the next. Furthermore, the respondents preferred the JRS to the sketch map. However, the JRS is more suggestive than the sketch map, but only for elements which are more rarely represented (railway tracks, etc). Thus, in terms of scientific research, the JRS makes it possible to improve comparisons between social groups. In terms of applications, it offers interesting possibilities for promoting citizen participation in urban planning.