A total of 54 videotaped mathematics activity observations drawn from 27 different pre-kindergarten classrooms in an urban public schools system in the U.S. were coded using 2 observational measures: (a) a general measure, the Classroom Assessment Scoring System (CLASS); and (b) a math-specific measure, the High Impact Strategies in Early Mathematics (HIS-EM). Criterion validity analysis revealed varying levels of convergent and discriminant patterns between the two tools. Indications of these results are twofold: (1) to some extend both tools document the similar aspects of quality of instruction (e.g., instructional support), and (2) a math-specific measure can be more sensitive to document the various levels of mathematics specific instructional support provided during teaching episodes (e.g., math representations, establishment of math learning communities). Results also indicated that the quality of mathematics instruction observed was revealed to be low to medium levels. (author abstract)