For the teaching of modeling correlation and regression an 11th-grade unit was designed on the basis of authentic practices in which practitioners use linear regression models. These practices include identifying suitable sports training programmes, dike monitoring and calibration of measurement instruments. The question we intend to answer is how educational versions of such practices can support students in learning to make data-based predictions in authentic situations. For example, the authentic practice of dike monitoring is used to motivate students to study scatter plots of dike height over time and learn how to model a regression line for finding a possible trend to predict when the dike needs to be heightened. The analysis of three teaching experiments shows that most students worked eagerly and with good results, but many continue to struggle with coordinating statistical and contextual knowledge.
展开▼
