The Purdue nonverbal test (PNT) was used to analyze intellectual abilities of 500 candidates for a driving license. The candidates were divided in five age groups of 100 subjects in each: 16-19, 20-24, 25-39, 40-54, and over 54. The distribution of result differs greatly from the normal distribution (P < 0.01) for all age groups of candidates. The distribution of the results of the candidates over 54 is positively asymmetric, while the distribution of the results of all other age groups is negatively asymmetric. These results suggest that the PNT is poor in discrimination. The percentage of candidates categorized according to the PNT as "capable for handling a motor vehicle" was higher than expected, while the percentage of candidates categorized as of "limited capability" was below expectations. According to these criteria, no candidate was found "incapable", which is in sharp nonconformity with the expected 5 per cent (5%). The above results indicate that the screening of candidates according to the criteria of intellectual ability is inadequate because of inadequate criteria, especially those which are supposed to determine borderline and lower intellectual ability. The results of the research show that admission of candidates whose intellectual ability borders or is below intellectual insufficiency parameters renders inadequate the prevention of risks in traffic. The risk in traffic is expected to increases with every single driving license issued to a person whose intellectual ability is inadequate for safe driving.