Model-Eliciting Activities (MEAs)¶

Strategy · All strategies
Evidence · 4 claims (4 for) · 11 studies (4 quant-synthesis, 3 causal, 2 review, 2 theoretical), q2–q4 · 5 of 11 report an effect size

Description¶

Model-Eliciting Activities are open-ended, realistic problems in which teams of learners must invent, document, test, and revise a generalizable procedure or model — a ranking system, a selection rule, a predictive tool — that a hypothetical client could actually use. Developed by Lesh and colleagues in mathematics education, MEAs are structured around a client letter and a "readiness questions" sequence, and they require learners to externalize their thinking in a written product that can be interpreted, critiqued, and improved by others. The activity is "model-eliciting" because the problem cannot be solved by applying a known procedure; learners must construct the procedure itself.

Design Implications¶

MEAs operationalize model-building as the unit of learning: learners reveal and refine their mental models through cycles of expression, testing, and revision, which makes them a strong vehicle for productive struggle before formal instruction Productive failure improves conceptual learning. [+M]. Because teams must reconcile multiple interpretations of a messy situation, MEAs create the social conditions for articulating and negotiating reasoning rather than merely computing answers. The written model product also functions as a rich formative assessment artifact, enabling feedback at the process level rather than the answer level Feedback is most effective at task and process levels. [+S].

Context¶

Requirements¶

• A realistic, client-driven scenario with genuine ambiguity and multiple defensible solutions
• A written deliverable (memo, procedure, rubric) that forces externalization of the model
• Team structures with interdependent roles so that ideas must be shared and defended (Collaborative Learning)
• Time for at least one full test-and-revise cycle; single-pass completion undermines the design
• Instructor facilitation through Coaching and process questions, not solution delivery

Constraints¶

• Requires substantial class time (typically 1–2 full sessions plus revision); inefficient for teaching well-defined procedures [~M]
• Teams can converge prematurely on a workable-but-shallow model if the instructor does not press for testing against edge cases [-M]
• Assessment is labor-intensive: grading open-ended written models reliably requires rubrics and calibration [-M]
• Learners with very low prior knowledge may flounder without readiness questions or scaffolds, producing frustration rather than productive failure [~M]

Implementation Variability¶

• Model-Eliciting vs. Model-Exploring: once a model exists, follow-up "model-exploring" activities apply and extend it with less invention required
• Discipline adaptations: engineering (first-year design courses), statistics (designing a ranking metric), biology (classification rules), business (vendor selection procedures)
• Scale: full MEAs, "quick MEAs" (single session), or warm-up readiness questions used alone as activation

Target Learners¶

• Undergraduate and secondary students with enough domain familiarity to interpret the scenario but not enough to apply a canned procedure [~M]
• Teams benefit from heterogeneous prior knowledge; peer negotiation of models supports persistence through difficulty Self-efficacy predicts academic persistence. [+M]
• Less suitable for complete novices without readiness scaffolds, or for learners who need fluency practice on established procedures first

Target Learning Goals¶

• Conceptual understanding: distinguishing surface features from underlying structure of a problem
• Mathematical and scientific modeling: translating a messy situation into formal relationships
• Communication and argumentation: justifying and defending a procedure in writing (Argumentation)
• Teamwork and metacognition: monitoring whether a model actually works under varied conditions

Instructions¶

1. Present the client letter — a realistic stakeholder describes a need in non-textbook language; use Advance Organizers sparingly so the ambiguity is preserved
2. Readiness questions — individual short questions that activate relevant knowledge and ensure every team member engages before group work (Activation)
3. Team model development — teams invent and document a generalizable procedure; instructor circulates as Coaching, asking teams to test their model against cases it might fail
4. Test and revise — teams apply their model to new data or a peer team's scenario and revise; contrasting multiple team models supports abstraction Multiple contrasting cases support abstraction. [+M]
5. Deliver and debrief — teams submit the written model; a whole-class Class Discussion compares approaches and connects to formal disciplinary content (Articulation)

Related Strategies¶

• Case-Based Learning — shares realistic anchoring scenarios, but MEAs require inventing a procedure rather than analyzing a decision already made
• Problem-Based Learning — similar ill-structured team format; MEAs differ in requiring a documented, generalizable product
• Productive Failure — MEAs often function as a productive-failure phase preceding direct instruction on canonical methods

Examples¶

• Purdue University first-year engineering — the "Little Betsy" and "Nanotechnology" MEAs (Diefes-Dux et al.) have teams design procedures for ranking or quality-control decisions; widely replicated across engineering programs (https://engineering.purdue.edu/ENE/Research/immex)
• Statistics courses — teams design a numerical index (e.g., "best college for the money") from a messy dataset, then defend the weighting scheme
• Modeling in middle-grades mathematics — Lesh's original "paper airplane" and "campground" problems, in which students build ranking or packing procedures for a client

Key Sources¶

• Lesh, R., Hoover, M., Hole, B., Kelly, A., & Post, T. (2000). Principles for developing thought-revealing activities for students and teachers. In A. E. Kelly & R. A. Lesh (Eds.), Handbook of research design in mathematics and science education (pp. 591–645). Lawrence Erlbaum.
• Lesh, R., & Doerr, H. M. (2003). Beyond constructivism: Models and modeling perspectives on mathematics problem solving, learning, and teaching. Lawrence Erlbaum.
• Zawojewski, J. S., Diefes-Dux, H. A., & Bowman, K. J. (Eds.). (2008). Models and modeling in engineering education: Designing experiences for all students. Sense Publishers.
• Hamilton, E., Lesh, R., Lester, F., & Yoon, C. (2007). The use of reflection tools to build explicit models of reflective thinking. Norwood, NJ: Lawrence Erlbaum.
• Diefes-Dux, H. A., Moore, T., Zawojewski, J., Imbrie, P. K., & Follman, D. (2004). A framework for posing open-ended engineering problems: Model-eliciting activities. Proceedings of the 34th ASEE/IEEE Frontiers in Education Conference. doi:10.1109/fie.2004.1408556