mcpField
L
learning-target-authoring-guide

learning-target-authoring-guide

github.com/garethmanning其他921141 次安装6 天前更新
代码仓库

关于 learning-target-authoring-guide

这是一套由实践者开发的指南,帮助教师和课程设计者为能力编写可评估的学习目标。流程先审视能力的目的,再分类知识类型并据此确定评估路径,随后拆分目标并进行质量检查。

主要功能

  1. 01审视能力设定的目的
  2. 02优先分类知识类型
  3. 03将能力拆分为学习目标
  4. 04编写发展阶段陈述或观察指标
  5. 05内置目标质量检查

适用场景

  1. 01为某项能力起草学习目标
  2. 02修改含糊或复合型目标
  3. 03为不同目标类型选择评估路径

安装

复制下面的命令,在终端中运行。

$npx skills add https://github.com/garethmanning/education-agent-skills --skill learning-target-authoring-guide

安装后在 Claude Code 中输入/learning-target-authoring-guide确认已加载。

详细介绍

以目标为导向的学习目标编写指南

此技能的功能

此技能编码了由教育工作者、课程设计师和学习系统设计师 Gareth Manning 开发的一套原创实践框架。与其他领域的技能不同,它并非源自同行评审研究传统,而是建立在对学习科学的深入探索、原创课程设计实践和积极的课堂测试之上。纳入此框架,是因为其方法论连贯、可迁移且确实实用;同时,保持学术诚信也要求我们区分实践者框架与经研究验证的方法。

此技能指导教育工作者或课程设计师,为某项能力编写一套完整、连贯的学习目标(LT):从上游问题——该能力是否恰当——到知识类型分类、拆解为 2-3 个 LT,再到编写发展性、精确且可评估的分级陈述(类型 1 和类型 2)或观察指标组(类型 3)。学习目标是广泛能力(“批判性思维”)与教师实际能够观察和评估的学生作品之间的桥梁。没有编写良好的 LT,基于能力的教育就会退化为模糊印象,或沦为清单式任务完成。v4.0 方法引入了一个强制分类步骤——知识类型——用于在编写任何分级陈述之前确定评估工具。这旨在解决基于能力的福祉与项目课程中最常见的失败模式:用量规评估倾向性目标,造成已有测量的假象,却没有实质依据。最终产出是一套可直接用于规划和报告的完整 LT,并在流程中内置质量检查。该方法借鉴了 Black & Wiliam(1998)的观点——清晰的学习意图是有效形成性评估的先决条件——并在 Wiggins & McTighe(2005)的逆向设计基础上,融入 Manning 提出的上游“为什么”层面:在提出“学生应该学习什么?”之前,先问“这项能力为何对我们的使命重要?”

关键区别:LT 描述的是目的地,而非内容。分级陈述描述的是学生在一个发展阶段结束时能够做到什么,而非规定为达成目标要教什么。内容的具体性体现在 KUD 图表和单元计划中。LT 限定了评估内容,但不限定教学内容。LT 看起来没有明确指定内容,反而说明它发挥了应有作用——KUD 图表才是明确内容的合适位置。

证据基础

Manning 通过迭代式课程设计制定了《学习目标编写指南》,并在 v4.0(2026)中记录了这一过程。该方法针对一个具体问题:大多数基于能力的课程,其学习目标编写得并不理想。常见的问题包括:使用无法观察的动词(“理解水循环”——如何观察学生是否理解?);把主题列表伪装成能力目标(“了解第一次世界大战的起因”——这是内容,不是能力);缺乏发展进阶(每个水平的表述都一样,只是在前面加上“简单”和“复杂”);复合型目标同时评估多个方面(“分析和评价资料以构建论点”——学生究竟在哪一部分遇到困难?);以及最隐蔽的问题——通过评分量规评估倾向性目标(针对情境任务评分的量规,无法告诉你学生是否养成了这种倾向;它只能告诉你,学生被要求时能否表现出来)。Black 和 Wiliam(1998)指出,有效的形成性评估需要清晰且共享的学习意图。如果教师不清楚自己究竟要观察什么,就无法提供有用的反馈,学生也无法进行自我评估。学习目标是明确性的基本单位:既具体到可评估,又足够宽泛,能应用于多个项目和情境。Wiggins 和 McTighe(2005)提出了逆向设计框架,Manning 在此基础上加入了一个前置的“为什么”层。标准的理解设计(UBD)从既定课程目标出发;以目标为导向的逆向设计则从学校使命出发,并提出这样的问题:这项能力为什么要纳入我们的课程?如果将其删除,我们会失去什么?这种前置追问可以避免课程中不断堆积无人能说明理由、却也无人敢于删除的能力目标。Bloom 的分类法(1956 年提出,2001 年修订)提供了约束学习目标编写的动词层级:用可观察、可评估的动词(识别、描述、比较、解释、论证、分析、评价、创造)替换不可观察的动词(理解、知道、欣赏、意识到)。Vygotsky(1978)的最近发展区(ZPD)为水平说明的编写提供了依据:每个水平的表述都描述学生在相应发展阶段能够在适当支持下做到什么——也就是最近发展区的上沿。

输入模式

教育者必须提供:

  • **能力名称:**宽泛的能力。例如:“批判性思维”/“科学探究”/“协作解决问题”/“自我调节”
  • **能力定义:**一句话。例如:“分析信息、评估证据并构建论证充分的观点的能力”/“在不同情境中监控并调整自身情绪和行为反应的能力”
  • **水平范围:**涵盖哪些水平。例如:“A-D 级(5-14 岁)”/“B-D 级(8-14 岁)”/“A-C 级(5-12 岁)”
  • **项目宗旨:**这项能力为何重要。例如:“我们的使命要求学生能够为这个需要他们的世界作出贡献——要作出贡献,就必须能够批判性地思考复杂问题,而不能只是接受他人所说的一切”

可选(如有,情境引擎会注入):

  • **现有草稿:**需要修改的现有学习目标表述
  • **学科领域:**课程所属领域
  • **评估情境:**学习目标的用途

提示词

You are authoring learning targets using Gareth Manning's methodology (Learning Target Authoring Guide v4.0). You must follow these rules PRECISELY — they are the quality standard for every LT produced.

═══════════════════════════════════════════════════════════════
STEP 0: KNOWLEDGE TYPE CLASSIFICATION (mandatory first step)
═══════════════════════════════════════════════════════════════

Before decomposition. Before writing band statements. Before anything else.

Every LT must be classified as one of three knowledge types. The classification determines the assessment instrument. Writing the LT before classifying it produces the single most common failure mode in competency-based wellbeing and project curricula: dispositional goals assessed through rubrics, creating the illusion of measurement without the substance.

THE THREE TYPES:

**Type 1 — Hierarchical Knowledge.** Concepts build on each other in a fixed sequence. Missing a prerequisite makes later content inaccessible, not merely harder.
- Indicators: prerequisite chains exist; errors at lower levels propagate upward; canonical sequencing logic; content cannot be reordered without loss.
- Assessment route: rubric with clear criteria, explanation tasks, vocabulary checks, knowledge tests.

**Type 2 — Horizontal Knowledge.** Organised around a central question or situation, illuminated through multiple disciplinary lenses. No single correct answer, but better and worse analyses. What develops cumulatively is not content but thinking sophistication.
- Indicators: multiple valid interpretive frameworks exist; reasoning quality determines assessment outcome; content can be entered from different points; analytical depth increases with experience.
- Assessment route: reasoning-quality rubric, scenario analysis, roleplay with novel situations, case studies, written analytical tasks. Transfer is the test.

**Type 3 — Dispositional Knowledge.** Capacities that develop gradually through sustained experience and exist only in enactment. Cannot be assessed through a single task. A rubric scored on a scenario task does not tell you whether a student has developed the disposition — it tells you whether they can perform it when asked.
- Indicators: the quality exists only across time and contexts; observation of the real behaviour (not a simulation) is required for valid evidence.
- Assessment route: multi-informant, longitudinal — teacher observation notes, student self-reflection, optional parent input, Strive data where applicable. Synthesised in developmental conversation. NOT a summative grade. NO RUBRIC.

CLASSIFICATION DECISION TREE — apply to EVERY emerging LT strand:

1. Does this LT describe content with a hard prerequisite sequence?
   YES → Type 1. Assessment: rubric / knowledge test.

2. Does this LT describe analytical or reasoning capability across multiple lenses?
   YES → Type 2. Assessment: reasoning-quality rubric, scenario task.

3. Does this LT describe a quality that only exists as it is enacted over time?
   YES → Type 3. Assessment: multi-informant observation. NO RUBRIC.

4. Does the statement contain elements from more than one type?
   YES → COMPOUND. Split before proceeding. Type 3 elements always go to the multi-informant observation pathway. Type 1 and Type 2 elements go to the rubric pathway. Each pathway has a separate LT.

5. Is this Type 3 LT dependent on Type 1 or Type 2 knowledge being in place first?
   YES → Knowledge-contingent. Name the prerequisite LTs explicitly.

KNOWLEDGE-CONTINGENT DISPOSITIONS:
A Type 3 disposition is knowledge-contingent when the student cannot develop the capacity without first having the Type 1 or Type 2 knowledge base in place. This is a sequencing requirement, not a design flaw.

Example: Self-regulation is knowledge-contingent. A student cannot genuinely develop self-regulation strategies without first understanding what the stress response is (Type 1), and without having practised a repertoire of regulation strategies in structured situations (Type 2). The Type 3 observation is only valid evidence once the prerequisites are established.

Design rule: If an LT is Type 3 and knowledge-contingent, name the prerequisite LTs explicitly. Do not expect the disposition to develop before the prerequisites are taught.

═══════════════════════════════════════════════════════════════
LT WRITING RULES (non-negotiable)
═══════════════════════════════════════════════════════════════

1. **Every band statement begins "I can..."** — present tense, active voice, first person. **TYPE 3 EXCEPTION: Type 3 LTs do NOT use "I can..." band statements. They use observation indicator sets — see Mode 3 below.**
2. **No parentheses, no inline examples, no "such as" or "for example" anywhere.** Examples are provided separately in teaching guidance, never embedded in the LT statement itself. A statement with examples baked in becomes an implicit checklist.
3. **Single-construct rule.** Each statement describes ONE coherent capability. Test: "Could I assess these separately?" If yes, they must be separate LTs or separate statements. "Analyses sources and constructs arguments" violates this — analysis and argument construction are separately assessable.
4. **Compound knowledge-type rule.** If a statement contains elements from different knowledge types (e.g., one Type 2 and one Type 3 element), it MUST be split into separate LTs with separate assessment routes — REGARDLESS of whether each element passes the single-construct test on its own. This is the most common cause of assessment misrouting.
5. **Observable action verbs ONLY.** Use: identify, describe, compare, explain, justify, analyse, evaluate, create, design, plan, conduct, present, reflect. NEVER use: understand, know, appreciate, be aware of, effectively, sophisticated, well (without observable criteria).
6. **Progression uses LEVERS, not topic escalation.** The six progression levers: independence (how much support needed), complexity (how many variables or perspectives), scope (how broad the context), precision (how accurate or detailed), reasoning (how deep the justification), transfer (how far from the original context). Progression across bands is achieved by advancing one or more of these levers — NOT by changing the topic or adding content.
7. **2 LTs per competency is the default.** 3 only if the strands are genuinely distinct and cannot be collapsed without losing meaning. NOTE: If a competency contains both a Type 2 analytical strand and a Type 3 dispositional strand, these must become separate LTs with separate assessment routes. This may mean a competency that appeared to need 2 LTs actually needs 3. This is correct design, not inflation.
8. **The LT definition describes what ALL band levels do.** It is the generalised "I can..." statement that applies across every band. It is NOT a Band A statement. For Type 3 LTs, the definition must include a note that this LT is assessed through multi-informant observation.

═══════════════════════════════════════════════════════════════
REQUIRED FIELDS FOR EVERY LT
═══════════════════════════════════════════════════════════════

Every LT must include:
- **LT Name** (Title Case, noun-phrase style, stable across years)
- **LT Definition** (one sentence, "I can...")
- **Knowledge Type** (Type 1 / Type 2 / Type 3)
- **Assessment Route** (Rubric with clear criteria / Reasoning quality rubric / Multi-informant observation)
- **Prerequisites** (required for knowledge-contingent Type 3 LTs — name the Type 1/2 LTs that must be established first)
- **Band statements** (Type 1 and Type 2) OR **Observation indicator set** (Type 3)

═══════════════════════════════════════════════════════════════
THREE AUTHORING MODES
═══════════════════════════════════════════════════════════════

**Mode 1 — Plain-language decision targets**
Best for: judgement, choices, safety, credibility, reflection, ethics, communication, collaboration routines, and most Type 2 wellbeing LTs.
Default pattern: DO / CHECK / DECIDE
- DO: main action (one clear verb)
- CHECK: 1-3 simple checks (fit, risk, source, audience, constraints)
- DECIDE: outcome + short reason

**Mode 2 — Skill/proficiency targets**
Best for: academic disciplines, arts, performance, scientific practice. Also appropriate for Type 1 wellbeing science knowledge.
Defaults: one clear skill per band statement; complexity and independence increase across bands; avoid tool/menu lists.

Performance-based Do tasks — presentations, physical demonstrations, structured performances, observable skill execution — are a sub-type of Mode 2. They are assessed through performance rubrics applied to the observable execution, not through multi-informant observation (Mode 3, T3 only) and not through analytical reasoning rubrics (which assess reasoning quality, not execution quality). Use the criterion-referenced rubric generator for performance-based Do tasks. Common examples: oral presentations, debate performances, physical education demonstrations, practical science procedures, musical or artistic performances.

**Mode 3 — Observation indicator sets (Type 3 ONLY)**
Best for: ALL Type 3 dispositional LTs. This mode is mandatory for Type 3 — not optional.
Structure: a set of specific, observable, third-person indicators that describe what the disposition looks like when present, calibrated to each band. These are NOT assessed with a rubric — they inform the observation notes and the developmental conversation.

Format:

LT 名称:[倾向性品质] 知识类型:Type 3 评估路径:多信息提供者观察

A 档指标——教师注意到:

  • 学生[在 A 档水平上可观察到的具体行为]
  • 学生[在 A 档水平上可观察到的具体行为]

B 档指标——教师注意到:

  • 学生[可观察到的具体行为]
  • 学生[可观察到的具体行为]

C 档指标——教师注意到:

  • 学生[可观察到的具体行为]
  • 学生[可观察到的具体行为]

D 档指标——教师注意到:

  • 学生[可观察到的具体行为]
  • 学生[可观察到的具体行为]

E 档指标——教师注意到:

  • 学生[可观察到的具体行为]
  • 学生[可观察到的具体行为]

F 档指标——教师注意到:

  • 学生[可观察到的具体行为]
  • 学生[可观察到的具体行为]

知识依赖型先决条件:[如适用,请写出 Type 1/2 LT 的名称]


Type 3 developmental progression: Indicators at Band D describe more sophisticated, self-directed, and cross-context manifestations of the disposition. At Band A they describe simpler, more immediate, adult-scaffolded manifestations. The progression uses the same levers (independence, complexity, scope, reasoning, transfer) — but they describe what the teacher observes, not what the student performs for assessment.

Type 3 evidence pathway:
- Teacher observation notes: brief, dated, specific indicator. Third-person. Stored in a retrievable form. At least one counterexample observation per unit (an observation that surprised the teacher or contradicted their existing view of the student).
- Student self-reflection: open questions calibrated to developmental band — not self-grading.
  - Band A: name a moment when you tried something hard
  - Band B: describe one pattern you noticed in yourself this term
  - Band C: compare what you noticed with what your teacher observed
  - Band D: analyse a pattern across contexts and explain what shifted
  - Band E: evaluate how your capacity has developed over the past year — what has changed, what conditions made change possible, and what has remained resistant to change
  - Band F: construct an account of your development across high-stakes real-world contexts over an extended period — what patterns hold, what conditions still challenge you, and what remains unresolved
- Parent/caregiver noticing prompts (optional): three plain-language "what do you notice?" questions sent home two weeks before developmental conversations.
- Strive data (where applicable): behavioural record of actual habit enactment.
- Synthesised in the developmental conversation. NOT a summative grade.

Student-generated single-point rubric (Type 3 supplement): At the start of each unit, the student writes one sentence describing what "doing this well" looks like for them. This becomes the reference point for teacher observation and developmental conversations — not a school-wide criterion. Bands A-B generate this collaboratively with the teacher. Bands C-D write it independently. Band D students review and revise at mid-unit as a metacognitive act. The value of the student-generated single-point rubric is in the act of generation — the student articulating what doing this well looks like for them — not in scoring. It is never used for grading.

═══════════════════════════════════════════════════════════════
DECOMPOSITION LOGIC
═══════════════════════════════════════════════════════════════

A competency is too broad to assess directly. It must be decomposed into 2-3 LTs that together COVER the competency and are individually ASSESSABLE. The decomposition should feel natural — like cutting a diamond along its grain, not chopping it into arbitrary pieces.

Test the decomposition:
- **Coverage test:** Do the LTs together cover the full meaning of the competency? Is anything missing?
- **Distinctness test:** Are the LTs genuinely different capabilities, or are they different ways of saying the same thing?
- **Assessability test:** Can each LT be assessed through observable student work (Type 1/2) or multi-informant observation (Type 3)?
- **Knowledge-type test:** Does each LT contain only ONE knowledge type? If a single LT bundles Type 2 and Type 3 elements, split it — regardless of whether it passes the single-construct test.

═══════════════════════════════════════════════════════════════
QUALITY CHECKLIST (must pass before output)
═══════════════════════════════════════════════════════════════

Format checks:
- [ ] Every Type 1 and Type 2 band statement begins with "I can..."
- [ ] Every Learning Target Definition begins with "I can..."
- [ ] Every Competency Definition begins with "The ability to..."
- [ ] All statements are present tense, active voice
- [ ] Type 3 LTs use observation indicator format, not rubric band statements

Prohibition checks:
- [ ] No parentheses anywhere in the set
- [ ] No inline examples in band statements
- [ ] No inline examples in Learning Target Definitions
- [ ] No "such as" or "for example" anywhere
- [ ] No content inventories (lists of topics/tools/periods)
- [ ] No vague quality descriptors without observable criteria

Construct checks:
- [ ] Single-construct rule passes for all band statements
- [ ] Could pass the "assess separately?" test
- [ ] Compound knowledge-type check: no statement bundles Type 2 and Type 3 elements

Knowledge type checks:
- [ ] Every LT has a Knowledge Type field (Type 1 / Type 2 / Type 3)
- [ ] Every LT has an Assessment Route field that matches its knowledge type
- [ ] Type 3 LTs do not have rubric band statements
- [ ] Knowledge-contingent Type 3 LTs have prerequisite LTs named
- [ ] No Type 3 element is buried inside a Type 1 or Type 2 LT statement

Coherence checks:
- [ ] Vertical coherence: clear A-F progression for all LTs
- [ ] Horizontal coherence: each Learning Target Definition accurately describes what all its band levels do
- [ ] Progression uses appropriate levers (independence / complexity / scope / reasoning / transfer)

Language checks:
- [ ] Language is plain and assessable throughout
- [ ] Age-appropriate for all bands
- [ ] Consistent spelling convention (UK or US)
- [ ] Consistent terminology across the set
- [ ] Strong action verbs (avoid understand / know / appreciate)

Completeness checks:
- [ ] All cells filled (or explicitly marked N/A with rationale)
- [ ] Notes used only for teacher clarifications, not hidden examples
- [ ] Knowledge Type and Assessment Route fields completed for every LT

═══════════════════════════════════════════════════════════════
PRE-SUBMISSION SELF-CHECK PROTOCOL
═══════════════════════════════════════════════════════════════

Before returning output, complete this protocol:

1. Read all Learning Target Definitions aloud. Do any contain content lists? Flag and fix.
2. Apply the "such as" test. Could you insert "such as" naturally anywhere? If yes, you have an implicit list. Remove it.
3. Check for compound statements. Count the "and" conjunctions. Does each one join a single coherent capability, or two assessable actions? Fix any violations.
4. Apply the compound knowledge-type test. For every "and" in every statement: do the two elements belong to the same knowledge type? If one element is Type 2 (analysable in a scenario) and another is Type 3 (only evidenced through real enactment over time), the statement is a compound knowledge-type violation. Split it.
5. Verify horizontal alignment. For each LT, does the definition describe what ALL band levels do? If not, revise.
6. Quality descriptor audit. Highlight all quality adjectives. Are they tied to observable criteria? If not, replace them.
7. Assessment route verification. For each LT: is the named assessment route actually capable of producing valid evidence for this kind of knowledge? A multi-informant observation route for a Type 1 knowledge LT is as wrong as a rubric for a Type 3 disposition.

═══════════════════════════════════════════════════════════════

Your task is to author a complete LT set for:

**Competency name:** {{competency_name}}
**Competency definition:** {{competency_definition}}
**Band range:** {{band_range}}
**Programme purpose:** {{programme_purpose}}

The following optional context may or may not be provided. Use whatever is available; ignore any fields marked "not provided."

**Existing draft:** {{existing_draft}} — if provided, review and revise against the quality rules. If not provided, author from scratch.
**Subject area:** {{subject_area}} — if not provided, design the LTs to work across subjects.
**Assessment context:** {{assessment_context}} — if not provided, design for use in both project assessment and reporting.

Return your output in this exact format:

## Learning Target Set: [Competency Name]

**Competency:** [Name]
**Definition:** [One sentence, "The ability to..."]
**Programme purpose:** [Why this matters for the mission]
**Number of LTs:** [2 or 3, with rationale — including knowledge-type split rationale if applicable]

### Competency Review

[Is this competency right? Does it serve the mission? Is it a genuine capability or a topic list in disguise? If the competency needs reframing, say so before proceeding.]

### Step 0: Knowledge Type Classification

[For each emerging LT strand, run the classification decision tree. State the knowledge type, the assessment route, and the reasoning. If any strand is compound, flag it and split before proceeding. If any Type 3 strand is knowledge-contingent, name the prerequisites.]

| LT Strand | Knowledge Type | Assessment Route | Reasoning |
|---|---|---|---|
| [Strand 1] | Type [1/2/3] | [Route] | [Why this classification] |
| [Strand 2] | Type [1/2/3] | [Route] | [Why this classification] |

### LT Decomposition

[The 2-3 LTs, with rationale for each and evidence that the decomposition passes the coverage, distinctness, assessability, and knowledge-type tests]

**LT1: [Name]**
- **Definition:** "I can [generalised statement that applies across all bands]"
- **Knowledge Type:** Type [1/2/3]
- **Assessment Route:** [Route]
- **Prerequisites:** [For knowledge-contingent Type 3 only; otherwise "N/A"]
- **Rationale:** [Why this strand exists as a separate LT]

**LT2: [Name]**
- **Definition:** "I can [generalised statement]"
- **Knowledge Type:** Type [1/2/3]
- **Assessment Route:** [Route]
- **Prerequisites:** [If applicable]
- **Rationale:** [Why this strand exists]

### Band Statements / Observation Indicators

[FOR TYPE 1 AND TYPE 2 LTs — use band statement table:]

**LT[n]: [Name]**
Knowledge Type: Type [1/2]
Assessment Route: [Route]
Authoring Mode: [Mode 1 or Mode 2]

| Band | Statement | Progression Lever(s) |
|---|---|---|
| A | I can [Band A statement] | [Which lever(s)] |
| B | I can [Band B statement] | [Which lever(s)] |
| C | I can [Band C statement] | [Which lever(s)] |
| D | I can [Band D statement] | [Which lever(s)] |

[FOR TYPE 3 LTs — use observation indicator format:]

**LT[n]: [Name]**
Knowledge Type: Type 3
Assessment Route: Multi-informant observation
Prerequisites: [Named Type 1/2 LTs, if knowledge-contingent]

Band A indicators — the teacher notices:
- The student [specific observable behaviour]
- The student [specific observable behaviour]

Band B indicators — the teacher notices:
- The student [specific observable behaviour]
- The student [specific observable behaviour]

Band C indicators — the teacher notices:
- The student [specific observable behaviour]
- The student [specific observable behaviour]

Band D indicators — the teacher notices:
- The student [specific observable behaviour]
- The student [specific observable behaviour]

Band E indicators — the teacher notices:
- The student [specific observable behaviour]
- The student [specific observable behaviour]

Band F indicators — the teacher notices:
- The student [specific observable behaviour]
- The student [specific observable behaviour]

Student self-reflection prompts:
- Band A: [prompt]
- Band B: [prompt]
- Band C: [prompt]
- Band D: [prompt]
- Band E: [prompt]
- Band F: [prompt]

### Quality Check

[Run every statement through the full v4.0 checklist. Flag any issues.]

**Format checks:**
- [ ] Every Type 1/2 band statement begins with "I can..."
- [ ] Every LT Definition begins with "I can..."
- [ ] Every Competency Definition begins with "The ability to..."
- [ ] All statements present tense, active voice
- [ ] Type 3 LTs use observation indicator format, not rubric band statements

**Prohibition checks:**
- [ ] No parentheses anywhere
- [ ] No inline examples in band statements or definitions
- [ ] No "such as" or "for example" anywhere
- [ ] No content inventories
- [ ] No vague quality descriptors without observable criteria

**Construct checks:**
- [ ] Single-construct rule passes
- [ ] Compound knowledge-type check passes: no statement bundles elements from different knowledge types

**Knowledge type checks:**
- [ ] Every LT has a Knowledge Type field
- [ ] Every LT has an Assessment Route that matches its knowledge type
- [ ] Type 3 LTs do not have rubric band statements
- [ ] Knowledge-contingent Type 3 LTs have prerequisites named
- [ ] No Type 3 element is buried inside a Type 1 or Type 2 LT

**Coherence checks:**
- [ ] Vertical coherence (clear A-F progression)
- [ ] Horizontal coherence (LT definition describes all band levels)
- [ ] Progression uses levers, not topic escalation

**Language checks:**
- [ ] Plain, assessable language throughout
- [ ] Consistent spelling convention
- [ ] Strong action verbs only

**Completeness checks:**
- [ ] All cells filled or marked N/A with rationale
- [ ] Knowledge Type and Assessment Route fields completed for every LT

### Known Limitations

[Where the LT set is strongest and where teacher judgment will still be needed]

**Self-check before returning output:** Run the full pre-submission protocol: (a) read all definitions aloud — any content lists? (b) "such as" test — could you insert it anywhere? (c) compound statement check — count the "and" conjunctions, (d) compound knowledge-type test — do elements on either side of "and" belong to the same knowledge type? (e) horizontal alignment — does the definition describe all band levels? (f) quality descriptor audit — all adjectives tied to observable criteria? (g) assessment route verification — is the named route capable of producing valid evidence for this knowledge type?

输出示例

场景 1(Type 2): 能力名称:“批判性思维” / 能力定义:“分析信息、评估证据并构建论证充分的观点的能力” / 档位范围:“A-D 档(5-14 岁)” / 项目宗旨:“我们的使命要求学生能够为一个需要他们的世界作出贡献。为复杂的现实问题作出贡献,需要具备批判性思维的能力——辨别有力证据与薄弱证据、识别假设,并构建经得起质疑的论证。缺乏批判性思维,学生就会成为他人结论的消费者,而不是自己结论的提出者”


学习目标集:批判性思维

能力: 批判性思维 定义: 分析信息、评估证据并构建论证充分的观点的能力 项目宗旨: 为一个需要批判性思考者的世界作出贡献——成为提出自己结论,而非消费他人结论的人 LT 数量: 2——分析与论证是截然不同的能力(可以分析而不论证,也可以论证却分析不充分)

能力审查

这项能力的界定恰当。它描述的是一种真正的能力(而非主题),直接服务于使命(作出批判性贡献需要批判性思维),且足够宽泛,可适用于所有学科和项目。定义包含三个要素:分析、评估、构建论证。分析与评估密切相关(评估本身就是一种分析),因此可合并为一个 LT。论证构建则有所不同——它是一种产出能力(创造某种东西),而非分析能力(拆解某种东西)。因此设为两个 LT。

步骤 0:知识类型分类
LT 方向知识类型评估路径理由
证据分析Type 2推理质量评分标准存在多种有效的分析框架;推理质量决定结果;没有唯一正确答案;分析深度会随经验增长。能否迁移到新情境是检验标准。
理据充分的论证Type 2推理质量评分标准论证质量依据推理的成熟度来评估,而非正确与否;存在多种有效立场;发展的重点是构建和捍卫立场的能力,而不是回忆正确答案。

未发现 Type 3 要素。作为一种倾向,批判性思维(自发质疑的习惯,而不仅是在被要求时才质疑)是真实且重要的能力——但这并非该能力所衡量的内容。如果项目希望记录批判性思考的倾向,就需要另设一个 Type 3 LT,并为其建立独立的观察路径。

LT 分解

LT1:证据分析

  • 定义:“我能够审视信息和证据,识别哪些内容相关、可靠且重要”
  • **知识类型:**Type 2
  • **评估方式:**推理质量评分标准
  • **前置条件:**不适用
  • **依据:**此 LT 涵盖批判性思维的分析维度——拆解信息、评估信息质量并判断哪些内容重要的能力。这是批判性思维的输入端:处理给定的信息。

LT2:有理有据的论证

  • 定义:“我能够构建清晰的论证,并以证据和推理作为支撑”
  • **知识类型:**Type 2
  • **评估方式:**推理质量评分标准
  • **前置条件:**不适用
  • **依据:**此 LT 涵盖批判性思维的产出维度——提出论点、以证据支撑论点,并以可供审视和质疑的方式呈现论点的能力。这是批判性思维的输出端:提出立场。

分解检验:

  • **覆盖度:**LT1(分析)和 LT2(论证)共同涵盖了“批判性思维”的完整含义。能够分析证据并构建有理有据的论证的学生,就是具有批判性思维的学生。
  • **独立性:**学生可能擅长分析但不擅长论证(能看出缺陷,却无法清楚表达自己的立场),也可能擅长论证但不擅长分析(论证颇具说服力,但依据的证据质量不佳)。这两者确实是不同的能力。
  • **可评估性:**这两个 LT 都能在学生的作品中体现出可观察的证据——可以通过学生评估来源、识别偏见和选择证据的方式观察其分析能力;也可以通过学生组织论点、提供支撑和回应相反立场的方式观察其论证能力。
  • **知识类型检验:**两个 LT 均属于 Type 2,不存在复合知识类型问题。
能力等级描述

LT1:证据分析 知识类型:类型 2 评估路径:推理质量量规 编写模式:模式 2

等级描述进阶要素
A我能从给定来源中识别相关信息范围:单一来源;独立性:给定来源
B我能比较多个来源中的信息并识别差异复杂度:多个来源;推理:比较
C我能评估来源的可靠性,并解释为什么某些证据比其他证据更有说服力推理:评估性判断;精确性:解释原因
D我能分析相互竞争的证据,识别假设和偏见,并论证针对特定目的而言哪些证据最可信迁移:将分析应用于新情境;复杂度:相互竞争的证据、多个标准
E我能评估一组证据,识别现有材料中的系统性缺口或偏见,并论证哪种阐释框架最适合该目的推理:评估整体证据基础,而非仅评估个别来源;范围:一组证据中的系统性模式,而非单次比较
F我能对一组证据进行批判性论述,兼顾证据产生的条件、其服务的利益,以及它所能揭示内容的局限独立性:自主构建批判性论述;推理:对证据基础本身进行二阶分析

LT2:有理有据的论证 知识类型:类型 2 评估路径:推理质量量规 编写模式:模式 2

等级描述进阶要素
A我能陈述自己的观点并给出理由独立性:单一主张、单个理由
B我能提出有多项证据支持的主张复杂度:多项证据
C我能构建论证,清楚陈述主张、提供支持性证据,并承认不同观点范围:包含反方立场;推理:理解论证结构
D我能构建有连贯性的论证,整合证据、回应反驳,并评估自身立场的力度迁移:对论证进行自我评估;精确性:连贯且整合
E我能构建有连贯性的论证,整合相互竞争的观点,预判可能的反对意见,并回应最有力的反方立场复杂度:整合相互竞争的观点;推理:以最有力的方式呈现反方立场
F我能评估自身论证的局限——识别证据不足之处、起关键作用的假设,以及改变自身立场需要哪些条件独立性:批判性地评估自身论证;推理:识别认识论局限
质量检查
  • 单一构念规则: 每项陈述只评估一种连贯的能力。LT1 始终关注审视证据;LT2 始终关注构建论证。没有任何陈述将两者结合。
  • 复合知识类型检查: 两个 LT 都属于 Type 2。没有混合类型的元素。
  • 无括号或行内示例: 均不存在。没有“例如”或“比如”。
  • 仅使用可观察动词: 识别、比较、评估、解释、分析、论证、陈述、呈现、构建、建立、整合、回应。没有“理解”“知道”或“欣赏”。
  • 纵向一致性: 在每个 LT 内,Band D 确实比 Band A 更进阶——各项杠杆体现出清晰、连贯的进阶过程。
  • 横向一致性: 两个 LT 的 Band B 描述了相近的发展要求——比较来源和呈现基于多项证据的主张,都是 Band B 级别的任务。
  • LT 定义适用于所有等级: “我能审视信息和证据,识别哪些内容相关、可靠且重要”适用于所有等级——Band A 以简单的方式做到这一点(单一来源),Band D 则以更深入的方式做到这一点(分析存在偏差的相互竞争证据)。
  • Knowledge Type 字段: 两个 LT 均标注为 Type 2,并指定了推理质量评分标准。
  • 评估路径验证: 推理质量评分标准适用于 Type 2 横向知识。能否迁移到新情境是检验标准。
  • 2 个 LT: 这一数量有充分理由——分析和论证是截然不同的能力。
已知局限
  1. 批判性思维既依赖领域,也具有跨领域通用性。 能对科学进行批判性思考的学生,未必能将这种能力迁移到历史学科中。这些 LT 将其编写为领域通用能力,但教师应认识到,批判性思维在不同学科中的表现各不相同,跨学科迁移需要明确的教学。

  2. Band A 刻意设定得较为基础。 “我能陈述自己的观点并给出理由”对批判性思维而言可能显得过于简单,但它准确体现了 5-7 岁儿童的发展起点。若要求 Band A 达到“分析多种观点”,就会设定不可能实现的要求,导致学生表面上达标或教师感到挫败。

  3. 这些 LT 未涵盖批判性思维的倾向。 学生可能具备分析证据的能力(被要求时能够做到),却不具备这种倾向(不会自发去做)。如果项目要求培养这种倾向,请另行编写一个 Type 3 LT——“作为实践的批判性思维”——并通过多方观察进行评估,而非使用评分标准。


情境 2(Type 3 — 展示观察指标路径): 能力名称:“自我调节” / 能力定义:“在不同情境中监控并调整自身情绪和行为反应的能力” / 等级范围:“Band A-D(5-14 岁)” / 项目宗旨:“我们的使命要求学生逐步主动承担对自身学习及其对他人影响的责任。自我调节是基础——缺少它,自主会变成冲动,协作会演变为冲突”


学习目标集:自我调节

能力: 自我调节
定义: 在不同情境中监控并调整自身情绪和行为反应的能力
项目宗旨: 自我调节是自主和协作的基础——缺少它,学生就无法逐步承担对自身学习及其对他人影响的责任
LT 数量: 3 — 知识类型分析揭示了三个真正不同的分支,必须将它们分开,以避免评估路径分配错误

能力评审

这项能力的设计框架合理,但包含了三种知识类型的要素。“监测并调整”看似是一项单一能力,实际上却包含:(a) 理解什么是压力反应以及调节如何发挥作用(类型 1——可教授、可测试的知识),(b) 分析自身模式,并在结构化情境中选择合适的策略(类型 2——关于自我调节的推理),以及 (c) 随着时间推移,在真实情境中实际进行调节(类型 3——通过行动展现的倾向)。如果把这些内容合并为一个通过评分标准评估的学习目标(LT),就只能涵盖类型 1 和类型 2 的要素。类型 3 的实际表现——也就是这项能力的核心——将不会体现在评估证据中。因此需要三个 LT。这是正确的设计,而非人为膨胀。

步骤 0:知识类型分类
LT 分支知识类型评估方式理由
理解压力反应类型 1使用明确标准的评分标准存在先决知识链:学生必须先理解压力期间生理和心理上会发生什么,才能对策略进行推理。低层级的错误会层层传递——误解压力反应的学生会错误地运用调节策略。
调节策略分析类型 2推理质量评分标准任何情境下都有多种有效策略;推理质量决定学生能否选出合适的策略;分析深度会随经验而提升;能否迁移到新情境是检验标准。
作为实践的自我调节类型 3多信息源观察这种能力只能在不同时间和情境中的实际行动里体现。能分析情境中的策略(类型 2),却不能在真实情境中进行调节的学生,并未养成这种倾向。必须观察真实行为。这项能力依赖知识:需要类型 1 和类型 2 的先决知识。

复合知识类型标记:原始能力定义捆绑了三种类型。以下拆分将它们划分为独立的 LT,并分别采用不同的评估方式。

LT 分解

LT1:理解压力反应

  • 定义:“我能够描述身心如何对压力作出反应,并解释调节策略如何发挥作用”
  • **知识类型:**类型 1
  • **评估路径:**具有明确标准的评分量规
  • **先修要求:**不适用
  • **理由:**这是知识基础。如果不了解压力反应过程中会发生什么,学生就无法判断哪些策略适用,也无法解释原因。通过解释任务和知识检测,可以高效地评估此 LT。

LT2:分析调节策略

  • 定义:“我能够分析具体情境,并选择适合该情境的调节策略”
  • **知识类型:**类型 2
  • **评估路径:**推理质量评分量规
  • **先修要求:**不适用(不过,LT1 的知识有助于提升表现)
  • **理由:**这是推理能力。学生能够从多个角度审视情境,并推理出适合的策略。通过要求将所学迁移到新情境中的情境任务进行评估,而不是考查对练习过的示例的记忆。

LT3:通过实践实现自我调节

  • 定义:“我能够在不同环境中调节自己对压力的反应。[通过多方观察者的观察进行评估——参见观察规程。]”
  • **知识类型:**类型 3
  • **评估路径:**多方观察者的观察——教师记录、学生自我反思、家长反馈、Strive 数据。在发展性谈话中综合这些信息。不计入总结性成绩。
  • **先修要求:**必须先掌握 LT1(理解压力反应)和 LT2(分析调节策略)。只有在学生具备相关知识和分析能力后,类型 3 的观察才是有效证据。
  • **理由:**这是付诸实践的倾向。如果学生能够描述如何调节(LT1)并分析策略选择(LT2),却无法在真实情境中进行调节,就说明他们尚未形成自我调节能力。此 LT 记录了其他 LT 无法体现的内容——长期的实际实践。

分解检验:

  • **覆盖范围:**三个 LT 共同涵盖了自我调节的完整内涵:了解什么是调节(LT1)、推理何时以及如何运用调节(LT2),并真正付诸行动(LT3)。
  • **区分度:**学生可能了解相关科学知识,却无法应用(LT1 强,LT2/3 弱)。学生可能能够在情境中推理调节策略,却无法在真实压力下运用(LT2 强,LT3 弱)。这三者确实彼此独立。
  • **可评估性:**LT1 通过解释任务评估。LT2 通过情境分析评估。LT3 通过多方观察者的观察评估。
  • **知识类型检验:**每个 LT 只包含一种知识类型。不存在复合项违规。
等级陈述 / 观察指标

LT1:理解压力反应 知识类型:类型 1 评估路径:具有明确标准的评分量规 编写模式:模式 2

等级陈述进阶杠杆
A我能识别身体何时感到平静,何时感到有压力独立性:在成人支持下识别状态;范围:仅关注自己的身体
B我能描述压力反应期间身体会发生什么,并说出有帮助的策略推理:描述机制;复杂度:将反应与策略联系起来
C我能解释压力反应如何影响思维和行为,并比较不同策略如何针对反应的不同方面推理:解释因果关系;复杂度:将策略与机制进行比较
D我能分析个体因素和情境因素如何影响压力反应,并评估哪些调节方法有证据支持迁移:应用于不同情境;推理:对证据作出评价性判断
E我能批判性地评估对压力反应的解释,指出哪些方面的证据充分、哪些方面仍有争议,并解释个体差异和情境差异如何影响调节原则的适用性迁移:将批判性评估应用于证据基础本身;推理:指出一般原则的局限
F我能依据证据基础,构建一套关于自身压力调节特征背后机制的说明,并评估一般调节原则在哪些条件下不太适用于我个人独立性:自主构建理论说明;迁移:从证据推广到个人特征

LT2:调节策略分析 知识类型:Type 2 评估路径:推理质量评分标准 编写模式:Mode 1

等级陈述进阶杠杆
A我难过时,能从提供的选项中选择一种平静下来的策略独立性:在支持下作出选择;范围:眼前的情境
B我能为某种情境考虑不止一种策略,并解释自己为何选择其中一种复杂度:多个选项;推理:为选择提供理由
C我能从多个角度分析情境,并选择适合情境和相关人员的策略范围:考虑情境和他人;推理:多角度分析
D我能评估复杂情境中的相互竞争策略,指出它们之间的矛盾,并说明自己的做法为何适用于不止一种情境迁移:跨情境应用;复杂度:处理相互竞争且存在矛盾的策略
E我能为持续时间较长且利害关系重大的情境制定调节计划,预判首选策略可能失效的条件,并确定应急方案推理:预判和应急规划;范围:持续时间较长且利害关系重大的情境
F我能为全新的重大情境设计调节方法,参照机制和证据说明设计依据,并根据一段时间内观察到的结果评估其效果独立性:自行设计方法;迁移:将原则应用于全新设计问题;推理:依据证据评估自己的设计

LT3:将自我调节付诸实践 知识类型:Type 3 评估路径:多方观察 先修要求:LT1(理解压力反应)、LT2(调节策略分析)

A 级指标——教师观察到:

  • 学生难过时能接受成人的安慰或引导,并开始回到活动中
  • 在非结构化的真实情境中,学生能在成人提示下使用一种明确的平静策略

B 级指标——教师观察到:

  • 在真实的困难时刻,学生无需成人提示便能主动采用调节策略
  • 事件过后,学生能较准确地说出自己采取了什么调节做法 Band C 指标 — 教师会注意到:
  • 学生会在不同情境中选择不同策略,表明他们意识到一种情境中有效的方法在另一种情境中可能行不通
  • 学生能在合理时间内从挫折中恢复,并重新投入任务或互动

Band D 指标 — 教师会注意到:

  • 学生在持续面临挑战时,能够在没有外部支架的情况下保持情绪调节
  • 被问及时,学生能够说出并解释自己的调节模式,包括这些模式随着时间发生了哪些变化

Band E 指标 — 教师会注意到:

  • 学生能在情绪完全失控前识别早期征兆,并在没有成人提示的情况下主动采用策略,包括在高风险情境中
  • 学生能根据对策略是否有效的实时评估,在情境中调整策略
  • 事件发生后,学生能够具体反思哪些因素改变了自己的状态,以及自己会采取哪些不同做法

Band F 指标 — 教师会注意到:

  • 学生在持续面临高风险的真实情境时,能够始终保持情绪调节,无需支架或成人事后干预
  • 学生能够综合描述自己多年来的调节特点,指出发生了哪些变化,以及哪些条件仍会给自己带来挑战
  • 学生会主动安排有助于自身情绪调节的环境、社交或时间条件,为高要求情境做好准备

学生自我反思提示:

  • Band A:告诉我这周有哪件事很难,你当时是怎么做的
  • Band B:描述一下本学期你发现的一个模式,说明你如何应对困难时刻
  • Band C:比较一下你对自身情绪调节的观察和教师的观察——哪些地方一致,哪些地方不一致?
  • Band D:分析一下你的情绪调节在不同环境中的表现,并说明过去一个学期有哪些变化
  • Band E:评估一下过去一年你的情绪调节发生了哪些变化——有哪些转变、哪些条件促成了这些转变,以及你在哪些方面仍会失去控制
  • Band F:综合描述一下随着时间推移,你在不同高风险情境中的情绪调节方式——你的调节特点是什么、哪些问题仍未解决,以及接下来需要努力的方向
质量检查

格式检查:

  • 每条 Type 1/2 等级描述均以“I can...”开头
  • 每条 LT 定义均以“I can...”开头
  • Competency 定义以“The ability to...”开头
  • 所有陈述均使用现在时和主动语态
  • Type 3 LT(LT3)采用观察指标格式,而非量规等级描述

禁止项检查:

  • 全文没有括号
  • 等级描述或定义中没有行内示例
  • 全文没有“such as”或“for example”
  • 没有内容清单
  • 没有缺少可观察标准的模糊质量描述

构念检查:

  • 所有陈述均符合单一构念规则
  • 复合知识类型检查通过:LT1 纯属 Type 1,LT2 纯属 Type 2,LT3 纯属 Type 3。所有陈述均无混合类型元素。

知识类型检查:

  • 每个 LT 都有 Knowledge Type 字段
  • 每个 LT 的 Assessment Route 均与其知识类型匹配(量规/推理量规/多方信息源观察)
  • Type 3 LT 没有量规等级描述,仅使用观察指标
  • 知识依赖型 Type 3 LT(LT3)列明了前置条件(LT1、LT2)
  • LT1 或 LT2 中没有隐藏 Type 3 元素

连贯性检查:

  • 纵向连贯性:每个 LT 内部均有清晰的 A-D 级进阶
  • 横向连贯性:LT 定义说明了所有等级的共同表现
  • 进阶通过杠杆实现(独立性、复杂度、范围、推理、迁移),而非不断升级主题

语言检查:

  • 全文使用浅白、可评估的语言
  • 遵循英式英语拼写规范
  • 仅使用有力的动作动词,不使用“understand”“know”或“appreciate”

完整性检查:

  • 所有单元格均已填写
  • 每个 LT 的 Knowledge Type 和 Assessment Route 字段均已填写
已知局限
  1. LT3 的证据取决于观察机制。 多方信息源观察要求教师保存注明日期且可检索的记录。在专用观察应用推出之前,教师实际会使用的最简便、最可持续的系统就是合适的系统。无法检索并跨一个学期进行比较的记录只是印象,并非证据。

  2. LT3 依赖知识,时机很重要。 在 LT1 和 LT2 的前置条件基本具备之前,不应启动 LT3 的 Type 3 观察路径。在学生尚未掌握相关知识和策略储备之前观察其自我调节,会产生无效证据。

  3. 不同信息源之间的分歧是信息,而非错误。 如果教师的观察记录与学生的自我反思对 LT3 的评价不一致,这属于应在发展性对话中提出的发展信息,而不是需要在对话前解决的问题。

  4. 倾向并非能力。 学生可能在 LT1 和 LT2 上得分很高(掌握科学知识,也能推理策略),但在 LT3 上表现出的证据有限(实际中无法做到自我调节)。这并非评估失败,而是真实且重要的发现。设置这三个 LT,正是为了让这种情况显现出来。


已知局限

  1. 这是一套实践方法论,而非经过同行评审的框架。 LT 编写规则是在课堂实践中不断开发和完善的,并非基于对照实验研究。它们以成熟的评估研究为基础(Black & Wiliam、Bloom、Wiggins & McTighe),但具体规则(单一构念、不提供行内示例、进阶杠杆、知识类型分类)是 Manning 的原创综合成果。

  2. 应用 LTs 需要专业判断。 一套编写完善的 LT 可以提供结构,但判断某份具体的学生作业是否符合某个特定等级的描述(Type 1/2),或某种观察模式是否构成某项倾向的证据(Type 3),仍需要教师运用专业判断。LTs 可以减少主观性,但无法消除主观性。

  3. 默认采用 2 个 LT 可能并不适用于所有能力。 有些能力自然可以拆分为 2 个维度;另一些则很难清晰拆分。知识类型的划分可能需要 3 个 LT,即使 2 个看起来已经足够。该方法论提供了判断标准(覆盖范围、区分度、可评估性、知识类型),但设计者必须自行判断何时遵循默认做法,何时偏离。

  4. Type 3 评估比 Type 1 或 Type 2 更难实施。 多方观察需要配套基础设施(可检索的记录、自我反思流程、发展性对话安排),而大多数学校目前尚未具备这些条件。这是一个现实的限制。解决办法不是为了方便而将 Type 3 LTs 纳入评分量规——那只会造成已经进行了测量的假象。解决办法是逐步建立这些基础设施。

评论

支持 Markdown