The purpose of this work was to determine the validity of a self-report on recent drug use (cocaine and cannabis) in a sample of university students of both sexes and to explore the role of attitudes toward substance use as related to this report. The subjects (506) were volunteers aged 17-35 years (who received an economic incentive) recruited at the University of Almería (Spain). The results were analyzed on the basis of correspondence between the self-report of recent use and a urine test. Three logistic regression analyses between self-reported use and attitudes toward drugs were also performed. The results show that the convergent validity of the self-report of drug use and the urine test is quite satisfactory, with percentages of agreement varying from .89 to .98 and Kappa of .66 and .56 for cannabis and cocaine, respectively. Sensitivity of the self-report is 57.1% (cocaine) and 91.8% (cannabis), and specificity is 99.4% (cocaine) and 89.6% (cannabis). The differences found in correspondence between the two substances are discussed with regard to self-reported attitudes on drug use.