Dialogic Facilitation of Statistical Modeling Practice¶

Strategy · All strategies
Evidence · 1 claim (1 for) · 1 study (1 qualitative), q2 · 1 of 1 report an effect size · 1 claim rests on one study

Description¶

Statistical modeling — mapping a real situation into a model, imagining it as one instance of a repeatable long-run process, judging how well a model fits, and using models to make inferences under uncertainty — is rarely taught as an integrated practice; more often, its component skills (running a simulation, computing an IQR) are taught as separate procedures. This strategy decomposes statistical modeling into four foundational aspects and pairs each with specific, recurring forms of dialogic teacher talk that push students from a single, static view of a phenomenon toward a coordinated modeling practice — for example, moving students from running a model once to systematically reasoning about a collection of runs, or from matching a single sample's shape to also weighing its variability.

Design Implications¶

Context¶

Requirements¶

• A modeling tool or activity capable of producing repeated simulated samples (the source study used spinners and TinkerPlots software), not just a single static representation
• Sustained instructional time across multiple lessons (the source study used 4 weeks, 15 lessons) — the four aspects and their coordination develop gradually, not in a single lesson

Constraints¶

• The teacher moves are content-general (applicable to any repeated-sampling context) but were only documented within one statistics unit; transfer to non-statistical modeling domains is untested

Implementation Variability¶

• Requires a teacher willing to press for elaboration and rerun models repeatedly rather than accept a first answer, which can slow initial pacing; the source study's teacher had no specialized statistical training, suggesting the moves are learnable without disciplinary expertise

Target Learners¶

• Middle-school students (studied at grade 6) learning to reason about variability, sampling, and inference

Target Learning Goals¶

• Dialogic teacher support cultivates all four aspects of statistical modeling practice over a multi-week unit [+W]

Instructions¶

1. Mapping situations to models: ask students to describe or justify why a model component corresponds to a real situational element; "animate" an actor's history to counter equiprobability assumptions (e.g., "what has this player been doing all year?"); press for concrete examples of any component a student names.
2. Simulating long-run processes: prompt rerunning a model ("what happens if you run it again?"); use repetition language ("over and over," "again and again"); ask students to anticipate a distribution's shape before running; ask students to interpret individual dots within a sampling distribution to keep sample and sampling-distribution distinct.
3. Evaluating model fit: extend fit criteria beyond shape and center to include variability (IQR, range); introduce "possible values" reasoning (a value absent from an empirical sample may still be a plausible outcome of the process); juxtapose alternative models that fit comparably but represent the process differently.
4. Making model-based inferences: pose a hypothetical sample statistic and ask students to judge whether it reflects chance or a real difference; make the counterfactual logic explicit ("if X were like all the others, it would be..."); use a sampling distribution to estimate how likely an observed (or more extreme) result is; explicitly distinguish "possible" from "likely."
5. Periodically coordinate aspects explicitly in a single discourse move — connecting a model's construction, its established fit, and a new inference question in one continuous exchange.

Related Strategies¶

• (none yet linked)
• Adopt a dialogic format to align student and teacher perspectives on instructional quality within schools

Examples¶

• A teacher, prompted to interpret an unusual sample statistic reported by another class, said: "Mr. B's class measured a median of 517. Do you think they measured the same table, or by chance?" — explicitly linking model fit and inference in one move.
• A teacher extended students' fit criteria beyond shape by asking, "Do we want a small range? Why?" after students had already agreed two simulated samples "looked the same."

Key Sources¶

• Wisittanawat, P., & Lehrer, R. (2024). Teacher Cultivation of Classroom Statistical Modeling Practice: A Case Study. Cognition and Instruction, 42(4), 505-547. https://doi.org/10.1080/07370008.2024.2397969