Experimentation

Element · All elements
Evidence · 13 claims (7 for, 4 mixed, 2 against) · 22 studies (9 quant-synthesis, 7 causal, 3 theoretical, 2 review, 1 associational), q1–q4 · 8 of 22 report an effect size · 7 claims rest on one study

Description

Experimentation asks learners to pose a question or prediction, systematically vary conditions, observe outcomes, and revise their understanding based on evidence. It functions as both a learning activity and a model of scientific reasoning: the learner generates a hypothesis, gathers data, and confronts discrepancies between expectation and result.

Design Implications

Experimentation supports learning by making learners active generators and testers of ideas rather than passive recipients Active learning improves exam performance relative to lecture alone. [+S]. Its effectiveness depends heavily on structure: unguided "discovery" tends to overload working memory and leave misconceptions intact, whereas experimentation with prompts, predictions, and feedback produces robust gains Unguided discovery is less effective than guided instruction. [-S]. Requiring explicit predictions before observing outcomes is the highest-leverage design move — prediction creates a commitment that evidence can confirm or violate, and violated expectations drive conceptual change Cognitive disequilibrium motivates conceptual change. [+M].

Context

Requirements

• A manipulable system or dataset where variables can be isolated and outcomes observed (simulation, lab, dataset, or worked domain)
• A prompt for explicit prediction or hypothesis before the outcome is revealed
• Timely, interpretable feedback linking the outcome back to the hypothesis
• Scaffolds for the reasoning cycle itself: question → predict → test → interpret → revise (Articulation prompts or structured lab reports)

Constraints

• Unguided experimentation with minimal scaffolding produces poor outcomes for novices, who tend to run unsystematic trials and confirm prior beliefs Unguided discovery is less effective than guided instruction. [-S]
• Learners often exhibit confirmation bias — testing cases they expect to succeed rather than disconfirming ones — unless prompted to consider alternative hypotheses [~M]
• In domains with high element interactivity, the overhead of managing the experiment itself can crowd out the target learning Cognitive overload degrades learning. [-M]
• Slow feedback loops (multi-day labs) weaken the prediction–outcome link; simulations with immediate feedback preserve it

Target Learners

• Learners with some prior knowledge who can generate a plausible first hypothesis; pure novices need a model or worked case first
• Learners holding misconceptions — the prediction–violation cycle is one of the most reliable ways to surface and confront them Cognitive disequilibrium motivates conceptual change. [+M]
• Less effective as a first exposure for complete novices, who lack the knowledge to interpret outcomes meaningfully

Target Learning Goals

• Conceptual change: replacing intuitive but incorrect models with evidence-consistent ones
• Scientific inquiry skills: variable control, hypothesis generation, evidence evaluation
• Transfer: reasoning from principles rather than recalling procedures Analogical reasoning improves transfer when learners compare cases. [+M]

Affordances

• Active Learning — experimentation enacts this principle by requiring learners to generate and commit to ideas rather than receive them, producing the retrieval and decision-making that strengthen encoding
• Cognitive Disequilibrium — a failed prediction creates exactly the expectation–evidence conflict that motivates schema revision
• Autonomy — letting learners choose what to vary and investigate supports intrinsic motivation, provided the choice space is bounded Autonomy supports intrinsic motivation. [+M]
• Cognitive Load Management — well-designed simulations and structured lab templates offload procedural overhead so working memory is spent on the hypothesis–evidence relationship

Claims

• CVS skill is the strongest predictor of performing valid (unconfounded) experiments in simulation-based inquiry [~W]
• Children's confidence in experimental conclusions increased after training and was higher for unconfounded than confounded experiments, with no training-condition differences [~W]
• Students' experiment-evaluation strategies shifted toward correct CVS after instruction, though 23% still used incorrect strategies [+W]
• Individual children in both training conditions were equally likely to become good experimenters or nonlearners after CVS training [~W]
• Instruction increased robust CVS use — designing unconfounded experiments plus providing a CVS rationale — in individual interviews [+W]
• Causal inference always goes beyond the data, even in randomized experiments, because experimental manipulations may affect other variables that caused the effect [~W]

The evidence by question

• Does adding guidance to inquiry improve what learners learn, and does the kind of guidance matter? Guidance during inquiry learning improves post-inquiry learning outcomes (d = 0.50) [+M]
• Should novices be left to discover the target content without support? Minimal guidance is less effective for novices than explicit instruction [-S]
• Does conflict-based instruction improve conceptual learning more than standard exposition? Conflict-based instruction improves science conceptual learning, though staged contradictions helped only learners who reported being confused, and no study isolates disequilibrium as the mechanism [+S]

Related Elements

• Application — experimentation is application with a hypothesis-testing structure layered on top
• Articulation — learners must state predictions and interpretations for the cycle to produce conceptual change
• Assessment — experimental artifacts (lab reports, revised hypotheses) provide authentic evidence of reasoning
• Case Studies — published cases can serve as "experiments others ran," useful when live manipulation is impractical
• Analogies — comparing multiple experimental outcomes supports abstraction of the underlying principle

Patterns That Use This Element

• 5E Learning Cycle — the "Explore" and "Explain" phases center on structured experimentation before formal explanation
• Cognitive Apprenticeship — the exploration phase, where learners test strategies in the target domain

Examples

PhET Interactive Simulations — University of Colorado physics/chemistry simulations designed around variable manipulation and immediate feedback; research-validated for conceptual gains when paired with prediction prompts.

Concord Consortium — Technology-enhanced science curricula (e.g., SmartGraphs, Molecular Workbench) embedding prediction-then-test cycles with embedded assessment.

WISE (Web-Based Inquiry Science Environment) — Scaffolds the full inquiry cycle — prediction, data collection, reflection — with embedded notes and stepwise guidance.

BSCS 5E instructional model — Published curriculum framework sequencing engagement, exploration, explanation, elaboration, and evaluation; the exploration phase is structured experimentation. - Use simulation as a non-threatening alternative to direct experimentation when testing office reorganizations

Key Sources

• Klahr, D., & Dunbar, K. (1988). Dual space search during scientific reasoning. Cognitive Science, 12(1), 1–48. doi:10.1207/s15516709cog1201_1
• Alfieri, L., Brooks, P. J., Aldrich, N. J., & Tenenbaum, H. R. (2011). Does discovery-based instruction enhance learning? Journal of Educational Psychology, 103(1), 1–18. doi:10.1037/a0021017
• Kirschner, P. A., Sweller, J., & Clark, R. E. (2006). Why minimal guidance during instruction does not work: An analysis of the failure of constructivist, discovery, problem-based, experiential, and inquiry-based teaching. Educational Psychologist, 41(2), 75–86. doi:10.1207/s15326985ep4102_1
• Hmelo-Silver, C. E., Duncan, R. G., & Chinn, C. A. (2007). Scaffolding and achievement in problem-based and inquiry learning: A response to Kirschner, Sweller, and Clark (2006). Educational Psychologist, 42(2), 99–107. doi:10.1080/00461520701263368
• Kuhn, D. (2007). Is direct instruction an answer to the right question? Educational Psychologist, 42(2), 109–113. doi:10.1080/00461520701263376

Page metadata

Field	Value
type	element
id	experimentation
title	Experimentation
description	Learners form predictions, manipulate variables, and test hypotheses against observed outcomes rather than receiving conclusions directly.
status	review
canonical	True
generated	by claude/unspecified on 2026-08-30
trust	unverified — no one has confirmed this page's content

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