Hair analysis is a useful tool in both clinical and forensic fields: it allows information about drugs of abuse (DOA) consumption to be obtained. However, in spite of analytical results, sometimes patients continue to deny using drugs or, on the contrary, insist on describing themselves as severe drug addicts; indeed there are often considerable difficulties in getting truthful statements about the real amount of drugs used. In this study we have tried to compare cocaine concentration in hair samples with self-reported drug intake. We enrolled 113 subjects (61 Africans, 52 Caucasians) who had been recently sent to jail. They were asked to tell about their use of illicit drugs during the last three months and then submitted to hair analysis. Hair segments (3 cm) were analyzed by GC-MS for amphetamines, cocaine and opiates. Useful data was obtained from 82 subjects, separated into two main groups on account of ethnic origin (African or Caucasian) and divided further into daily, weekly and monthly users. The results showed qualitative results and self-reported consumption to be in good agreement, although the correlation between frequency of consumption and concentration in hair revealed sometimes higher concentrations in contrast with the admission of low consumption. There was a definite separation between occasional and daily use (especially in Caucasian people), while concentrations found where weekly use was reported were more variable. Concentrations of cocaine measured in Africans' hair were much higher than in Caucasians'. Even if this study is exclusively based on self-report, it provides some interesting information in order to differentiate the frequency of consumption, and especially underlines the great importance of ethnic bias on hair analysis.