Abstract Computer-assisted content analysis has many advantages compared to a manual scoring system, provided that computerised dictionaries represent valid and reliable measures. This study aimed to assess the inter-coder reliability, alternate- form reliability and scoring consistency of the Body Type Dictionary (BTD) (Wilson 2006) based on Fisher and Cleveland’s (1956, 1958) manual body boundary scoring scheme. The results indicated an acceptable inter-coder agreement with barrier and penetration imagery in the sub-sample (N = 53) of manually coded Rorschach responses. Additionally manually coded scores showed an acceptable correlation with the computerised frequency counts, and thus indicating an alternate-form reliability. In the full data set (N = 526), barrier imagery in the Rorschach responses only correlated with the picture response test, showing low scoring consistency, which might disconfirm the notion of body boundary awareness representing a stable personality trait but instead it might be dependent on the level of cognitive dedifferentiation.