Skip to content

Evidence-Centered Design

Research Method · All research methods
Evidence · 4 claims (3 for, 1 mixed) · 2 studies (2 design), q2 · 0 of 2 report an effect size · 4 claims rest on one study

Description

ECD is a principled assessment design framework that structures the rationales, choices, and consequences of assessment development. The article describes how the framework identifies layers and models: "These models are then glued together by (4) the assembly model and (5) the presentation model that make the assessment deliverable." Domain modeling drives the whole process, laying out the assessment argument in narrative terms before formal models are built. The authors adopt ECD as a natural bridge between the demand for high fidelity of epistemic games and the demand for rich assessment data.

Accounts

  • Evidence-Centered Design framework structures assessment design through student, task, evidence, assembly, and presentation models: ECD is a principled assessment design framework that structures the rationales, choices, and consequences of assessment development. The article describes how the framework identifies layers and models: "These models are then glued together by (4) the assembly model and (5) the presentation model that make the assessment deliverable." Domain modeling drives the whole process, laying out the assessment argument in narrative terms before formal models are built. The authors adopt ECD as a natural bridge between the demand for high fidelity of epistemic games and the demand for rich assessment data. (Rupp et al. (2010))
  • Evidence Centered Design as a principled approach to rubric development: The article applies Evidence Centered Design (Mislevy, Almond, & Lukas, 2003) as a principled assessment development approach. As the article states, "Ze XVed an eYidence-cenWeUed deVign aSSURach (MiVleY\, AlmRnd, & LXkaV, 2003) WR deYelRS Whe UXbUicV", because the approach helps ensure clear connections between what is observed in assignments and student work and what is inferred about them. The process involved articulating claims, defining assessment targets, considering what evidence would imply targets are met, and considering tasks that would bring about the evidence; each decision in the development pathway needed adequate justification. (Iwatani et al. (2021))
  • ECD-based framework treating GenAI as a design variable with three sequential governance stances: The framework applies Evidence-Centered Design's three models—student, evidence, and task—as sequential decision criteria for GenAI governance in STEM assessment. It analyzes how GenAI reshapes each model and, as the abstract states, "specifies when to restrict, scaffold, or require GenAI use in STEM assessment." Decisions are organized across two branches (AI interaction competency is or is not the target construct) with a construct revision loop, and the authors note that "Governance decisions are not static" and require ongoing reassessment as capabilities evolve. (Gao et al. (2026))
  • Evidence-centered design as the framework linking MCD learning objectives to observable response evidence: The article frames MCD development within evidence-centered design (ECD), in which "the student models specify the knowledge an assessment makes claims about, and the evidence models specify the link between observed responses and those claims." ECD comprises five models (student, evidence, task, assembly, and delivery); this paper focuses on the student models (the 14 LOs) and evidence models (the Q-matrix and DINA model). The authors treat DINA fit and classification accuracy as validity evidence for the student models, testing whether the LOs defined a priori are measurable and separable in student responses. (Vy Le et al. (2026))
  • Evidence-Centered Design (ECD) as the methodological foundation for stealth assessment in game-based learning environments: Evidence-centered design is a systematic approach for designing knowledge assessments in terms of evidentiary arguments, featuring three core models: a task model defining the activities students undertake, an evidence model that processes low-level action sequences into machine-interpretable evidence, and a competency model that dynamically estimates student mastery. In this article, ECD grounds the stealth assessment framework in Geniventure, where the task model consists of "60 challenges across six game levels," the evidence model links gameplay behaviors to targeted genetics concepts, and the competency model estimates mastery of 16 concepts aligned to post-test scores. Stealth assessment conducts real-time processing of data from these three models to inform adaptive scaffolding without disrupting gameplay. (Henderson et al. (2020))
  • Evidence-centered assessment design (ECD): competency, evidence, task/action, and assembly models: ECD is "a conceptual design framework to help in the creation of coherent assessments" supporting "a broad range of assessment types, from classroom quizzes to simulation-based assessments." The competency model defines knowledge, skills, abilities and other attributes as probability distributions per variable; the evidence model specifies what student actions count as evidence and how it statistically links to competency variables; task/action models express situations that evoke required evidence; and the assembly model specifies how the models work together to form a valid assessment. Statistical machinery such as IRT or Bayesian networks links observable data to unobservable competencies. (Shute et al. (2013))
  • Evidence-centered design (ECD) framework for stealth assessment: ECD is a framework for designing assessments as evidentiary arguments, consisting of three main models: a Competency Model defining claims about student competencies, an Evidence Model specifying what behaviors constitute evidence, and a Task Model defining tasks that elicit that evidence. In stealth assessment, ECD formalizes claims about the learner and the evidence supporting them, allowing continuous performance data from complex interactive environments like video games. The article uses ECD to build the creativity assessment in Physics Playground, with game levels serving as tasks. (Shute et al. (2020))
  • Evidence-centered design (ECD) as a conceptual framework for integrating assessments into educational games: ECD is a conceptual design framework for crafting educational assessment tests, "emphasizing consistency between the evidence gathered and interpreted and the underlying knowledge addressed by assessments". In game-based environments it guides researchers to collect evidence making valid claims about students' competency levels during gameplay. Researchers employing ECD specify the competences to be assessed, outline corresponding assessment tasks, and collect evidence such as interactions and behaviors during gameplay to infer students' mastery levels. The review reports ECD was the most-cited framework (N = 6) and recommends integrating it when crafting GBAs. (Su et al. (2024))
  • ECD-based LLM conversation design process with a five-module prompt structure: A design framework for building conversation-based assessments with LLMs via prompt engineering, applying Evidence-Centered Design so conversations stay on the domain topic and elicit construct evidence. The prompt comprises five modules: (1) introduction and domain using the persona and audience patterns, e.g. "Your name is John. You are humorous and reliable study partner"; (2) a conversational schema with Socratic questioning, misconception handling, and turn-length limits; (3) persona behavior modeling via negations and response-type handling instructions; (4) restating of the four-step process; and (5) evidence instructions triggering a terminate command and an XML Toulmin evidence diagram aligned to constructs and standards. (Diego Zapata-Rivera et al. (2024))

Claims

Key Sources

  • Rupp, A.A., Gushta, M., Mislevy, R.J., & Shaffer, D.W. (2010). Evidence-centered Design of Epistemic Games: Measurement Principles for Complex Learning Environments. Journal of Technology, Learning, and Assessment, 8(4). http://www.jtla.org
  • Iwatani, E., Means, B., Seylar, J., and Hardy, A. (2021). Rubrics on historical thinking skills for assignments and student work: Initial validity evidence. Digital Promise. https://doi.org/10.51388/20.500.12265/112
Page metadata
Field Value
type research-method
id evidence-centered-design
title Evidence-Centered Design
description ECD is a principled assessment design framework that structures the rationales, choices, and consequences of assessment development.
status review
moved_from theory/ecd-framework-models-epistemic-games, theory/evidence-centered-design-rubric-development
generated by process:sweep-kinds on 2026-10-10
trust unverified — no one has confirmed this page's content
sources 2 — rupp-2010, iwatani-2021

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