In this work we examined the content of approximately 950 mathematics questions used in measuring achievement among 4th, 5th and 6th graders in the US in an attempt to determine how the current standardized examinations, which were developed to reflect new national curricular mandates, compared to the old measures used for the same purpose. Using Bloom's Taxonomy and TIMSS characterization of cognitive demands of tasks, we compared the quality of knowledge assessed by the items. Notable differences were found between the new items measuring skills and competencies compared to compatible items on the old tests. We noted that previous assessments primarily tested children's procedural and skill-based knowledge of mathematics, and often allowed children to arrive at a right answer in absence of understanding. In contrast, the new assessments require that children have a conceptual understanding of mathematics, attend to precise mathematical language, communicate their ideas, and utilize abstract reasoning skills.