How to Teach with Model-Eliciting Activities

Model-eliciting activities (MEAs) are usually given to students to work on in groups in a classroom setting. After students have produced a solution and written their reports, they share their solutions with the class. These presentations may lead to further class discussion and groups going back to re-examine and revise their models. A course might use one MEA or multiple MEAs.

A model-eliciting activity is typically used to "plow the field" so that seeds of learning more formal content may be planted. For example, an MEA on developing a SPAM filter to use for email messages can be used to teach the students ideas of Type 1 and Type 2 error (e.g., classifying an email as SPAM when it is not is like a Type 1 error) or to prepare students to learn about Bayesian methods or Logistic Regression.

Another way to use MEAs is to engage the students in statistical thinking. By considering a real-world problem, a set of data, and the need for a solution, students are experiencing the statistical enquiry cycle (see Wild and Pfannkuch, 1998).

Model-eliciting activities can be followed by an assessment, but we think that the best time to assess student learning is after the formal instruction that follows an MEA or at a later point in time. Assessments may consist of a parallel type of problem to solve, using the correct statistical procedure, or a different type of problem that allows students to transfer their learning to a novel problem or context.

A model-eliciting activity typically lasts 50 - 75 minutes. The reading and individual students' responses in the first part of an MEA can take place prior to class and comparison of student reports generated in the last part can take place at a subsequent class or via an online class management system.

For more detailed instructions and a list of model-eliciting activities, see the examples page.