Abstract

The basic principle of biological theories holds that crime is caused by a genetic or physical defect. Indeed, biological theories assume that people who commit crimes are ‘born criminals’ and their biological features may differ from non-criminals. The aim of the present study was to compare the Facial Ratios, Facial Masculinity, and Digit Ratio between ex-convicts and non-convicts in Kano state and to generate equations for criminal discrimination using linear measurements, ratios, and indices of the face. Five hundred and seventy-one (571) subjects were selected comprising of 252 ex-convicts and 319 non-convicts, using random sampling methods with age range of the participants between 18-50 years. Digital Vernier and spreading callipers were used to measure the facial linear dimensions of face directly. Facial ratios and indices were derived from facial distances. The lengths of the 2^nd and 4^th digits were determined using a direct method of measurement. The independent-samples t-test was used to determine the mean differences in facial ratio, facial masculinity, and 2D:4D between ex-convicts and non-convict participants. Binary logistic regression analyses were used to discriminate the face. It was observed that there were significant (P<0.001) differences between ex-convicts and non-convicts in the following facial ratios and indices, upper facial width upper face height, face height, nose width with, and nasal index (P<0.005). The upper facial width index shows higher significant difference among the facial parameters. The present study found that 4D both right and left sides show significant differences only between the study populations, none of the digit ratios show significant differences between ex-convicts and non-convicts. The best predictor that discriminates between criminals and non-criminals was facial width followed by eye length with explained variance of 2.8 and 4.4% in the 1^st and 2^nd steps in the models respectively. The overall accuracy of prediction of the best predictor was 78.1%. The prediction accuracy was in favour of criminal group compared to their non-criminal counterparts. The models are significantly fit for the discrimination of criminals from non-criminals using facial dimensions.