The activities with Minitool 1 afforded more than informal reasoning about majority, outliers, chance, and reliability; they also supported the visual estimation of the mean (Figures 3 and 4). After this strategy had spontaneously emerged in the exploratory interviews, we incorporated instructional activities to evoke this strategy in other classes as well (Bakker, 2003). Minitool 1 supported the strategy with the movable vertical value bar. Students said that they cut off the longer bars, and gave the bits to the shorter bars. Several students in different classes could explain that this approach was legitimate: The total stays the same, and the mean is the total divided by the number. When students said that brand D is better because its mean is higher, they used the mean to say how good the brand is. In that case, the mean is not just a calculation on a collection of data, but refers to a whole subset of one brand. As we intended, they learned to use the mean as a representative value for a data set and to reason about the brand instead of the individual data values.