This paper describes findings from a set of extended observations of 15 public school teachers in an effort to gain insight into classroom assessment procedures. Two elementary school, seven middle school, and six high school teachers were observed over 5 days by dividing time into 10-minute intervals. Attention was given to the amounts of time teacher and students spent on five action categories: (1) formal assessment; (2) informal assessment; (3) integrated assessment and instruction; (4) other on-task actions; and (5) off-task actions. The major role of informal assessment and integrated assessment and instruction was in stark contrast to the emphasis provided in current measurement textbooks, suggesting that many textbooks are not responsive to the actual needs of teachers. Validity issues related to the assessments being judged appeared to be linked to the teacher's understanding of the material being taught, but the degree to which the teacher sought criterion-related evidence of validity appeared to be a function of whether the teacher anticipated discrepancies between observed student performance and actual student ability. (Contains three tables and five references.) (SLD)