The Observer's Assessment of Alertness/Sedation (OAA/S) Scale was developed to measure the level of alertness in subjects who are sedated. This scale was tested in 18 subjects in a three-period crossover study to assess its reliability and its criterion, behavioral, and construct validity. After receiving either placebo or a titrated dose of midazolam to produce light or heavy sedation, each subject was administered two sedation scales (OAA/S Scale and a Visual Analogue Scale) and two performances tests (Digit Symbol Substitution Test and Serial Sevens Subtraction). Two raters individually evaluated the subject's level of alertness on each of the two sedation scales. The results obtained on the OAA/S Scale were reliable and valid as measured by high correlations between the two raters and high correlations between the OAA/S Scale and two of the three standard tests used in this study. The OAA/S Scale was sensitive to the level of midazolam administered; all pairwise comparisons were significant (p less than 0.05) for all three treatment levels at both test periods.