Skip to content

Evidence-Based Reasoning

Element · All elements
Evidence · 3 claims (1 for, 2 mixed) · 6 studies (2 quant-synthesis, 2 associational, 1 causal, 1 design), q1–q4 · 1 of 6 report an effect size · 2 claims rest on one study

Description

Evidence-based reasoning asks learners to evaluate two or more competing models or explanations against real-world data, deciding which model best accounts for the evidence and articulating why. Rather than receiving a canonical explanation, learners weigh alternative claims, coordinate evidence with theory, and revise their understanding — a process that targets epistemic cognition as much as domain content.

Design Implications

Evidence-based reasoning supports conceptual change by requiring learners to confront the inadequacy of intuitive or naive models against data, which is more effective than simply presenting the correct model [~M]. The activity gains its power from genuine competition between plausible alternatives; when one model is obviously correct, learners engage in pattern matching rather than evidential reasoning. Structured argumentation supports — such as evidence–claim linking frames — improve the quality of learners' reasoning, especially for novices Argumentation quality improves when learners are given explicit scaffolds for connecting claims to evidence. [+M].

Context

Requirements

  • Two or more genuinely plausible competing models, each with some evidential support
  • Real or realistic data that learners can interpret and that discriminates between the models
  • A structure for articulating justification — e.g., claim–evidence–reasoning frames (Argumentation)
  • Opportunities to revise initial commitments after evaluating evidence (Cognitive Conflict)

Constraints

  • Unguided "discovery" of which model fits the evidence often fails; learners with low prior knowledge may not notice evidential inconsistencies without scaffolds [~S]
  • Learners frequently reason backward from prior beliefs, discounting disconfirming evidence (belief bias); simply presenting contradictory data rarely produces conceptual change [-M]
  • Heavy data-interpretation demands can overwhelm working memory, leaving no capacity for the reasoning itself; pre-chunking data or providing interpretation supports mitigates this [~M]
  • Less effective when the domain has no genuine evidential ambiguity or when models differ only in terminology rather than substance

Target Learners

  • High school and higher-education students who hold intuitive models that conflict with scientific accounts
  • Learners with moderate prior knowledge — enough to interpret evidence, but with misconceptions worth confronting
  • Advanced learners and research trainees developing disciplinary norms of evidential justification; novices need substantially more scaffolded data interpretation

Target Learning Goals

  • Conceptual change: replacing intuitive models with evidence-consistent ones
  • Epistemic cognition: understanding what makes a claim warranted and how evidence bears on theory
  • Critical evaluation: assessing competing explanations and articulating justified judgments
  • Argumentation skill: constructing and critiquing evidence-linked claims

Affordances

  • Cognitive Load Theory — competing-model tasks must be designed so data interpretation is manageable; chunking evidence and providing interpretive schemas preserves working memory for the reasoning that matters
  • Active Learning — learners generate judgments and justifications rather than receiving conclusions, producing deeper processing than passive presentation of the correct model
  • Cognitive Apprenticeship — making expert evidential reasoning visible (how a scientist weighs conflicting data) lets learners appropriate disciplinary norms of justification

Claims

  • Argumentation — the discourse structure through which evidence-based reasoning is expressed and refined
  • Cognitive Conflict — the mechanism by which evidence contradicting prior beliefs opens the door to conceptual change
  • Case Studies — a common vehicle for supplying realistic evidence against which models are tested
  • Justification — the explicit articulation of why evidence supports one model over another; often scaffolded with claim–evidence–reasoning frames

Patterns That Use This Element

Examples

Model-Evidence Link (MEL) activities — Developed for climate science education (e.g., through the University of Colorado Boulder and Penn State's MEL projects), learners link evidence such as ice-core data or spectral measurements to competing explanations of climate change and rate the plausibility of each connection. See https://mel2.meteorology.psu.edu

Case-Based Learning — Published case curricula such as National Center for Case Study Teaching in Science present data-rich scenarios where learners must adjudicate between competing diagnoses or mechanisms.

Structured scientific argumentation curricula — Osborne and colleagues' argumentation toolkits for secondary science provide claim–evidence–reasoning frames that scaffold learners in evaluating alternative explanations of phenomena.

Key Sources

  • Chinn, C. A., & Brewer, W. F. (1993). The role of anomalous data in knowledge acquisition: A theoretical framework and implications for science instruction. Review of Educational Research, 63(1), 1–49. doi:10.3102/00346543063001001
  • Kuhn, D. (2010). Teaching and learning science as argument. Science Education, 94(5), 810–824. doi:10.1002/sce.20395
  • Nussbaum, E. M., & Sinatra, G. M. (2003). Argument and conceptual engagement. Contemporary Educational Psychology, 28(3), 384–395. doi:10.1016/S0361-476X(02)00038-3
  • Berland, L. K., & Reiser, B. J. (2009). Making sense of argumentation and explanation. Science Education, 93(1), 26–55. doi:10.1002/sce.20286
Page metadata
Field Value
type element
id evidence-based-reasoning
title Evidence-Based Reasoning
description Learners evaluate competing models based on real-world evidence and refine their understanding.
status review
canonical True
generated by claude/unspecified on 2026-08-29
trust unverified — no one has confirmed this page's content

Frontmatter is the page's machine-readable record — see the Schema & Guide.