The National Assessment of Educational Progress (NAEP) focuses on providing information on what demographic subgroups of America's students know and can do. Because the NAEP does not report for individual students, it could be classified by some students as a low-risk test. Consequently, some students may lack proper motivation for giving their best efforts. This paper describes a method to identify students who may have randomly marked or systematically marked (pattern marked) their responses. The magnitude of the difference between responses (in absolute value) is used to compute a variance for each student's response string. The variances for students who take a particular block of items are compared; students who have a very small variance may have lacked motivation and are classified as potential pattern markers. Approximate sample sizes for each block (for both mathematics and science) were 3,700 students in the age 13 years/grade 8 cohort and 3,600 students in the age 17 years/grade 12 cohort of the 1990 NAEP. Subgroup analysis indicates that males are identified as potential pattern-markers more frequently than are females. Black and Hispanic American students are classified as potential pattern markers more often than are White students. Pattern marking tends to occur more frequently when the item block is near the end of the test. Seven tables, five figures, and three references are included. (Author/SLD)