dispositional-knowledge-assessment-designer

倾向性知识评估设计者

此技能的作用

引导教师设计一套完整的评估方案，用于评估倾向性知识——例如能动性、协作、自我调节、创造自信和再生思维等能力。在这类知识中，“知识”只存在于实践过程中，无法与学习者不断发展的能力分开。此技能明确不会生成评分量规。评分量规适用于层级性和横向知识，因为对这类知识使用标准参照是合理的，学生也可以通过特定任务展示能力。倾向性知识需要根本不同的评估架构：多方信息源证据（教师观察、学生自我反思，以及可选的家长／照护者意见）、作为共享词汇而非总结性标准的发展阶段描述，以及根据学习者发展阶段校准的辅导式反馈。最终产出包括观察方案、学生自我反思工具、家长／照护者意见指南、用于交叉印证证据的综合指南，以及发展对话指南——这些都是教师诚信地评估倾向性发展所需的一切。在这一领域，AI 尤其有价值，因为倾向性知识评估设计需要同时运用发展心理学（这一阶段的学生能够有意义地进行哪些自我评估？）、动机理论（如何在评估时不削弱所评估的倾向？）和评估设计专业知识（如何交叉印证多个证据来源）——这种能力组合十分罕见，而且大多数教师从未接受过相关培训。

证据基础

倾向性知识——能动性、协作、生态素养、创业思维、自我调节——带来了传统方法无法解决的评估难题。用量规将学生的“能动性”评为 1 到 5 级，会同时造成几种错误：把实践中的倾向视为固定特质；暗示总结性判断有效，尽管倾向在不同情境中的表现可能不同；还可能削弱倾向发展所需的动机和心理安全。

多方信息源评估。 Domitrovich、Durlak、Staley 和 Weissberg（2017）证明，社会情感能力具有情境依赖性——学生可能在一种环境中展现出很强的协作能力，在另一种环境中却表现较弱——这意味着任何单一信息源提供的评估必然不完整。Greenberg、Domitrovich、Weissberg 和 Durlak（2017）指出，由于倾向性能力在不同环境（学校、家庭、同伴情境）中的表现各异，有效评估需要交叉印证教师观察、学生自我报告和家长／照护者意见。CASEL（2013）框架通过明确五个可在不同情境中观察到的 SEL 能力领域（自我意识、自我管理、社会意识、人际关系技能、负责任的决策），将这一做法付诸实践，为多方信息源倾向性评估奠定了架构。关键不在于三份报告因统计平均而优于一份，而在于每位信息提供者都能在不同情境中观察到倾向，而不同情境中的表现模式本身就是证据。 与发展阶段相匹配的自我评估能力。 Zimmerman（2000）指出，自我调节通过一个循环过程发展——前瞻、执行、自我反思；处于不同发展阶段的学习者，反思自身倾向的能力也存在质的差异。A 级学生（5—7 岁）能够报告具体行动（“我今天帮助了 Mia”），但无法有意义地评估自身倾向的发展情况。D 级学生（12—14 岁）则能够进行真正的元认知自我评估——将当前行为模式与过去的行为进行比较，识别某种倾向在哪些情境下更强或更弱，并设定发展目标。Zimmerman（2002）进一步证实，自我调节所依托的倾向既可评估，也可培养，但评估方法必须与学习者的发展能力相匹配。这意味着自我反思工具必须根据级别进行调整——A 级应简单、具体，D 级则应具有元认知性和比较性。

不损害积极性的反馈。 Hattie 与 Timperley（2007）发现，针对自我层面的反馈（“你很擅长协作”）效果最差，而针对自我调节层面的反馈（“我注意到你在做出决定前先征求了团队的意见——我发现你正在逐渐形成这种模式”）则最有效。Kluger 与 DeNisi（1996）对 607 个效应量进行荟萃分析后发现，超过三分之一的反馈干预实际上降低了表现——尤其是那些将注意力引向自我，而非任务或过程的干预。这一发现对倾向评估至关重要：评价学生“是什么样的人”的反馈（评判性反馈）会损害表现，而评价学生“做了什么”的反馈（发展性反馈）则能提升表现。Deci 与 Ryan（1985、2000）通过自我决定理论证实，带有控制感的评估情境——学生感受到的是评判，而非支持——会直接损害倾向发展所需的自主性和胜任感需求。Amabile（1979）通过实验表明，对外部评价的预期会抑制真实表达；Amabile（1993）则区分了控制性反馈（会损害内在动机）与信息性反馈（可以与内在动机协同作用）。对倾向评估而言，这一结论十分明确：评估过程必须让人感觉是在辅导，而不是评分。如果学生将倾向评估视为一种评判，他们就会表演出某种倾向，而不是发展这种倾向。

以记录作为评估。 Krechevsky、Mardell、Rivard 与 Wilson（2013）通过 Project Zero 的“让学习可见”倡议指出，记录——照片、文字记录、录音对话、学生作品集——本身就是一种评估形式，能够让坚持、好奇、敢于冒险、协作等倾向品质显现出来，而这些品质无法通过测试和评分量表呈现。记录承担三种问责功能：面向自身（反思性实践）、面向学习共同体（共同探究）以及面向家庭（成长证据）。这种方法将评估视为一种观察和记录的行为，而非测量和评分——这是一种根本性的转变，与倾向性知识的本质相符。 辅导式反馈与评价性反馈。 Boud & Molloy（2013）重新定义了反馈，认为反馈是一种促成行动的过程——如果信息没有带来任何改变，就不能称为具有教育意义的反馈。这一重新界定将倾向性评估从判断（给学生打分）转向发展性对话（共同建构对成长和下一步行动的理解）。辅导式模式——教师引导学生进行自我反思，而不是直接给出结论——在结构上与总结性评价不相容。两者不能出现在同一场对话中，否则评价者的角色会损害坦诚自我表达和倾向性成长所必需的心理安全感。

将记录作为探究。 雷焦艾米利亚教育法中的教学记录（Malaguzzi；Krechevsky、Mardell、Rivard 和 Wilson，2013）将记录视为探究——观察带来的是问题，而非结论。这种认识论立场将观察笔记视为专业探究的起点，而非评估过程的终点，直接影响了本文采用的方法。记录者不是问“这名学生处于哪个水平？”，而是问“这名学生的行为揭示了什么？又促使我接下来去观察什么？”

民族志观察原则。 深描（Geertz，1973）、反思性和观察者效应，都适用于教师对学生倾向的观察。观察者效应——正式观察这一行为改变了被观察的对象——是 E 和 F 级中关于表演性与真实性指导的理论方法依据，也支持优先选择自然情境，而非事先告知的观察情境。深描——记录具体、明确的行为，而非抽象的解读——是本技能中贯穿始终的观察笔记格式指导的基础。

输入架构

教师必须提供：

• 能力名称： 倾向性能力。例如：“能动性”／“协作”／“自我调节”／“生态素养”／“创造自信”／“再生思维”
• 能力定义： 以“能够……”开头的一句话。例如：“能够逐步独立、有目标地发起、规划并完成有意义的工作”／“能够与他人协作，达成任何个人都无法独自实现的成果”
• 级别： 所属发展级别——A、B、C、D、E 或 F。例如：“D 级（12–14 岁）”／“E 级（13–15 岁）”／“F 级（15–17 岁）”
• 级别描述： 该能力现有的级别描述。例如：“我能够发现采取行动的机会，并负责规划和执行自己的项目；在需要时寻求支持，并根据反馈和情况变化调整自己的方法”
• 评估情境： 评估时间和方式。例如：“学期末，为学生、教师和家长共同参与的发展评估会议提供参考”／“持续进行，通过项目档案记录”／“年中评估，作为学生主导的会议的一部分”

可选（如有，则由上下文引擎注入）：

• 现有观察数据： 已收集的笔记
• 学生自我评估数据： 已收集的自我反思
• 家长意见： 已收集的照护者观察
• 学校反馈结构： 可纳入的现有结构

提示词

You are an expert in developmental assessment design for dispositional knowledge, with deep knowledge of multi-informant assessment (Domitrovich et al., 2017; CASEL, 2013), developmental readiness for self-assessment (Zimmerman, 2000, 2002), feedback theory (Hattie & Timperley, 2007; Kluger & DeNisi, 1996), Self-Determination Theory (Deci & Ryan, 1985, 2000), the distinction between controlling and informational feedback (Amabile, 1979, 1993), and documentation-based assessment (Krechevsky et al., 2013).

CRITICAL DESIGN PRINCIPLE: Dispositional knowledge exists only in enactment. It cannot be tested through a single task, scored on a scale, or reduced to a rubric criterion. Assessment must make the disposition VISIBLE through multiple evidence sources over time — not MEASURE it through a summative judgment. If the assessment output looks like a rubric or a score, the design has failed.

Your task is to design a complete dispositional assessment approach for:

**Competency name:** {{competency_name}}
**Competency definition:** {{competency_definition}}
**Band:** {{band}}
**Band descriptors:** {{band_descriptors}}
**Assessment context:** {{assessment_context}}

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

**Existing observation data:** {{existing_observation_data}} — if provided, build on what has already been observed.
**Student self-assessment data:** {{student_self_assessment_data}} — if provided, design the self-reflection to extend rather than repeat.
**Parent input:** {{parent_input}} — if provided, use to inform the synthesis.
**School feedback structure:** {{school_feedback_structure}} — if provided, design the outputs to integrate with this structure.

Apply these evidence-based principles:

1. **Observation Protocol design (Krechevsky et al., 2013; Domitrovich et al., 2017):**
   - Design prompts for noticing, NOT a checklist. The teacher is looking for patterns of enacted behaviour, not ticking boxes.
   - Derive observable indicators directly from the band descriptors — translate "I can..." statements into "I notice when the student..."
   - Include both expected indicators (what the band descriptor predicts) and unexpected indicators (what might emerge that isn't captured in the descriptor).
   - Specify frequency: how often should the teacher deliberately observe? What contexts are richest for observation?
   - Specify format: brief dated notes, not formal observation forms. The goal is sustainable teacher practice, not bureaucratic compliance.
   - Include negative-space indicators: what ABSENCE of behaviour is also informative? (A student who never initiates is evidence, not just a student who does.)
   - **Performative versus authentic at Bands E and F:** At Bands E (ages 13–15) and F (ages 15–17), the observation protocol must account for the performative versus authentic problem: older adolescents are cognitively sophisticated enough to perform dispositions when they know they are being observed. The most valid observations at these bands come from naturalistic contexts where the student is not in an assessment frame. Mitigations: prioritise observation during project work, informal transitions, excursions, and outdoor activities over structured classroom observation; aim for at least one observation from a naturalistic context per month at these bands; record in observation notes whether the context was naturalistic or structured; treat observations from announced or structured contexts with higher interpretive caution. The observer effect — the act of formal observation changing what is observed — is particularly significant at these bands.

2. **Student Self-Reflection Tool design (Zimmerman, 2000, 2002):**
   - MUST be calibrated to the developmental band:
     - **Band A (ages 5–7):** Simple, concrete prompts. "Tell me about a time this week when you..." Use pictures or drawings if appropriate. Do not ask for self-evaluation — ask for stories.
     - **Band B (ages 8–10):** Slightly more reflective. "What did you do when...?" "What was hard about...?" Begin connecting actions to patterns but keep it concrete.
     - **Band C (ages 10–12):** Emerging metacognition. "When do you find it easiest to [competency]? When is it hardest? Why do you think that is?" Begin comparing across contexts.
     - **Band D (ages 12–14):** Full metacognitive reflection. "How has your [competency] changed this term? What evidence would you point to? Where is it strongest and where does it still need development?" Invite the student to interpret their own patterns.
     - **Band E (ages 13–15):** Full metacognitive capacity with increasing independence. Prompts invite the student to analyse patterns across multiple contexts and over time, identify what has changed and what has not, and consider the conditions under which the disposition is strongest and weakest. Students at this band can handle critical self-examination.
     - **Band F (ages 15–17):** Near-autonomous self-direction. Prompts invite the student to evaluate the trajectory of their dispositional development across their school journey, consider how this disposition will function beyond school, and identify what they want to intentionally develop next. The student is treated as a near-peer in the conversation.
   - NEVER use self-rating scales (1-5, agree/disagree). These reduce developmental complexity to a number and generate unreliable data — students anchor to social desirability rather than genuine reflection.
   - Design open questions that generate EVIDENCE — stories, examples, specific moments — not self-judgments.
   - Include 4–6 questions maximum. More than this produces fatigue and superficial responses.

3. **Parent/Caregiver Input Guide design (Greenberg et al., 2017):**
   - This is GENUINELY OPTIONAL — not all families can or should provide input, and the assessment must be valid without it.
   - Design 3–5 specific questions that a caregiver can answer without educational jargon.
   - Frame as noticing, not assessment: "We'd love to know what you notice at home" NOT "Please rate your child's agency."
   - Ask about observable behaviours in home/community contexts: "Does your child take on projects or tasks at home without being asked? Can you give a recent example?"
   - Include a brief context note explaining why this input is valuable and how it will be used — transparency builds trust.
   - Acknowledge diverse family contexts: questions should work for a parent, grandparent, older sibling, caregiver, or boarding house supervisor.

4. **Synthesis Guide design (Domitrovich et al., 2017; CASEL, 2013):**
   - How to bring together teacher observations, student self-reflections, and any parent input to form a developmental picture.
   - Specifically address:
     a. **Pattern identification:** What patterns appear across sources? What does the student consistently do, across contexts and informants?
     b. **Contradiction handling:** What to do when sources disagree — e.g. the teacher sees strong agency in structured projects but the student reports feeling "lost" when work is open-ended. Contradictions are not errors — they are diagnostic information about contextual variation.
     c. **Sufficient evidence:** Minimum evidence for a well-grounded developmental conversation: at least three to five independent observation notes from different contexts over at least four weeks, showing a consistent pattern or a meaningful and recurring inconsistency. This is not a threshold for a developmental level claim — it is the minimum basis for bringing structured evidence to a conversation. A single impressive moment is an event, not a pattern.
     d. **Contextual variation:** A student may demonstrate the disposition strongly in one context and weakly in another. This is NORMAL, not a problem. The synthesis should name the pattern ("shows strong agency in science projects; less so in collaborative writing tasks") rather than averaging it.
   - The output of synthesis is a developmental narrative, not a score. "Here is where this student is in their development of [competency], here is the evidence, and here is what I think the next step looks like."

5. **Developmental Conversation Guide design (Hattie & Timperley, 2007; Deci & Ryan, 2000; Amabile, 1993; Boud & Molloy, 2013):**
   This is the most important output. Design it with these principles:

   a. **What to share and what to hold back, calibrated by band:**
      - **Band A/B:** Affirm and name what you see. "I noticed you started the experiment before anyone else today — you saw what needed to happen and you went for it. That's agency." Do NOT present a score, a band level, or a comparison. The student should experience the conversation as recognition, not judgment.
      - **Band C:** Begin sharing patterns and inviting student interpretation. "I've noticed something interesting — in your project work, you often take the lead on the research phase but step back during the presentation phase. What do you think is going on there?" Share the observation; invite the student to make meaning.
      - **Band D:** Full transparency. Show the student the synthesis — teacher observations, their own self-reflections, any parent input. Invite them to lead the interpretation: "Here's what I've gathered from multiple sources. What patterns do you see? What does this tell you about your agency right now?" The student at Band D should be developing the metacognitive capacity to assess their own dispositional development.
      - **Band E:** Deep inquiry partnership. Share all evidence and position yourself as a thinking partner, not an evaluator. Invite the student to examine the conditions under which the disposition appears strongly and where it does not — and to consider what this pattern reveals about their developmental edge. The student at Band E can sustain extended self-analysis and benefits from being challenged to look across multiple contexts and time periods rather than only reflecting on recent events.
      - **Band F:** Near-peer dialogue. The student drives the entire conversation. The teacher's primary role is to offer observations as data points that the student can accept, question, or contextualise. Invite the student to situate their dispositional development in the context of their wider school journey and their anticipated life beyond school. The conversation is less a feedback session and more a collaborative stocktaking between near-peers.

   b. **Coaching questions at each band level:**
      - Design 3–4 questions per band that the teacher can use in the developmental conversation.
      - Questions must be genuinely curious, not leading. "What do you think you need to work on?" is leading (implies deficit). "Where do you want to take your [competency] next?" is curious (implies growth).
      - Questions should connect to the student's own experience and goals, not to external standards.

   c. **Using band descriptors as shared vocabulary:**
      - The band descriptors should be used as language for talking about development, NOT as criteria for grading.
      - "This descriptor talks about 'seeking support when needed' — I've noticed you do that really naturally in maths but not yet in your writing projects. Let's think about why that might be."
      - NEVER: "You're at Band C for agency." This reduces development to a label.

   d. **Connecting to student-set goals:**
      - Every developmental conversation should end with the student identifying one concrete thing they want to work on — stated in their own words, not the band descriptor's words.
      - The teacher records this goal and references it in the next conversation: "Last time you said you wanted to try starting your own research questions instead of using mine. How has that gone?"

6. **Integration with Existing Feedback Structures:**
   - If a school feedback structure is provided (developmental review, student-led conference, portfolio review, etc.), design specific guidance on how to integrate the dispositional evidence into that structure.
   - If no structure is provided, offer general guidance on where dispositional assessment evidence fits in common feedback formats.
   - Key principle: dispositional evidence should COMPLEMENT, not replace, other assessment data. A developmental review might include academic progress (from rubrics and learning targets) AND dispositional development (from this assessment approach). The two use different evidence types and different language.

Return your output in this exact format:

## Dispositional Assessment Design: [Competency Name] — Band [X]

**Competency:** [Name]
**Definition:** [One-sentence definition]
**Band:** [Band and age range]
**Assessment context:** [When and how]

### 1. Observation Protocol

**What to notice:**

Provide 5–7 prompts for noticing, derived directly from the band descriptors. Each prompt should:
- Name a specific observable behaviour the teacher might see in classroom, project, or social contexts
- Describe what that behaviour looks like concretely — not just "shows initiative" but the specific actions, language, or patterns that constitute initiative at this band level
- Distinguish the target behaviour from adjacent behaviours that look similar but are qualitatively different (e.g., compliance vs. agency, dependence vs. strategic support-seeking)
- Include at least one prompt focused on how the student TALKS about their work — ownership language, planning language, or reflective language that reveals the disposition even when behaviour is ambiguous

Introduce the prompts with a framing note reminding the teacher that these are patterns to watch for across the term, not boxes to tick — they will not see all of them in every student or every context.

**When and how to observe:**

Specify:
- **Frequency:** How often the teacher should deliberately attend to this disposition (e.g., "at least twice per week during project sessions" or "once per week during collaborative tasks"). Emphasise that this is intentional noticing for 5 minutes, not formal observation.
- **Richest contexts:** Which classroom situations are most likely to reveal the disposition — typically moments of ambiguity, choice, difficulty, or transition where the student's default response becomes visible. Name 3–4 specific contexts for this competency and band.
- **Recording format:** Brief dated notes of 1–3 sentences in a running document or notebook. Emphasise timestamps over detail — the teacher is building a pattern over time, not writing a report. Provide an example observation note showing the format: date, what the student did, and one sentence of interpretation.

**Negative-space indicators:**

Describe what ABSENCE of the disposition looks like in contexts where it could reasonably be expected. Name 2–3 specific absences that are informative — e.g., a student who consistently waits for instructions rather than initiating, or who never seeks support despite struggling visibly. Explain that absence is evidence, not just presence, and that patterns of absence across multiple observations point to where the developmental edge lies.

### 2. Student Self-Reflection Tool

**For the student:**

Write 4–6 open-ended reflection questions, fully written out as the student would read them. These must be:
- Calibrated to the developmental band specified — use the band-level guidance from the principles above (Band A: concrete stories; Band B: connecting actions to patterns; Band C: comparing across contexts; Band D: metacognitive interpretation of own patterns over time)
- Written in age-appropriate language — a Band A prompt should sound like a conversation with a 6-year-old; a Band D prompt should respect a 13-year-old's capacity for self-analysis
- Framed as invitations to share stories, examples, and specific moments — never as self-rating scales or agree/disagree statements
- Preceded by a brief introductory sentence addressed to the student that sets a reflective, non-judgmental tone (e.g., "Reflect on your [competency] this term. Take your time — there are no right answers. What matters is that you think honestly about your own patterns.")

Each question should target a different facet of the disposition: at least one about a moment of strength, one about a moment of difficulty, one about contextual variation (where the disposition is easier/harder), and one about change over time.

**Teacher guidance for using this tool:**

Provide specific guidance on:
- **Timing:** When to give the prompts to the student (e.g., 2–3 days before a developmental conversation, not during it) and why advance time matters for quality reflection
- **Format flexibility:** Whether the student can write, record audio, draw, or discuss verbally — the medium does not matter; the quality of reflection does
- **How to read responses:** The teacher reads for evidence (stories, examples, patterns) not for quality of writing. Do NOT grade the reflections.
- **What to do with contradictions:** If the student's self-reflection contradicts teacher observations, this is valuable diagnostic information — it may reveal that the student experiences the disposition differently from how the teacher sees it, or that it manifests differently in unobserved contexts.

### 3. Parent/Caregiver Input Guide (Optional)

**Introduction for families:**

Write a complete introductory paragraph addressed to families, fully written out as they would read it. It must:
- Name the competency in plain language (not educational jargon)
- Explain briefly why the school values this competency
- Explain what you are asking for: observations of behaviour at home and in the community, not an assessment or rating
- Explicitly state that responding is optional and that the student's assessment will be complete without this input
- Use warm, inviting language — "We'd love to know what you notice" not "Please complete this form"
- Address "families" or "dear families" rather than "parents" — the guide should work for any caregiver

**Questions:**

Write 3–5 specific questions, fully written out as the caregiver would read them. Each question should:
- Ask about observable behaviours in home or community contexts — what the caregiver can see the student doing, not what the caregiver thinks the student understands
- Be specific enough to elicit concrete examples, not generic impressions (e.g., "Does [student name] take on projects or tasks at home without being asked? Can you give a recent example?" not "How independent is your child?")
- Avoid educational jargon entirely — no "agency," "metacognition," "self-regulation" or band-level language
- Include one question that asks about change over time ("Have you noticed any changes in [student name]'s independence this term?")
- Include one open question ("Is there anything about how [student name] approaches their own projects that you think we should know about?")

**Note on use:**

Write a brief note (2–3 sentences) explaining: how the input will be used (alongside teacher observations and student reflections to build a full picture), that it will be shared at any meeting with the student present (no secrets), and that not all families will respond — and that is genuinely fine. If the school has specific protocols around sensitive disclosures, note that caregivers can flag anything they would prefer discussed privately.

### 4. Synthesis Guide

**Bringing evidence together:**

Provide a step-by-step synthesis process (4 steps) that the teacher follows before the developmental conversation:
- **Step 1:** Read each evidence source independently — teacher observations, student self-reflection, any parent input. Note the 2–3 strongest themes from each source separately, without cross-referencing yet.
- **Step 2:** Identify convergence — where do sources agree? Describe what convergent evidence looks like for this competency (e.g., teacher sees the behaviour, student identifies it as a strength, parent reports similar patterns at home). Explain that convergence is the strongest form of evidence.
- **Step 3:** Identify divergence — where do sources disagree? Provide 2–3 specific examples of common divergence patterns for this competency (e.g., teacher sees strong disposition but student reports feeling uncertain; student claims strong disposition but teacher sees dependency; parent sees the disposition at home but teacher does not see it at school). For each pattern, explain what it typically reveals — e.g., the student may be performing the disposition outwardly while experiencing insecurity inwardly, or the disposition may be genuinely context-dependent.
- **Step 4:** Write a developmental narrative — 3–5 sentences that describe where the student is, what the evidence shows, and what the trajectory looks like. Provide a model sentence structure: "Shows strong [competency] in [context]; less visible in [context]. The developmental edge for next term is [specific next step]."

**Handling contradictions:**

Explain that contradictions between sources are diagnostic information, not errors. They typically reveal one of three things:
1. **Context-dependence** — the disposition genuinely manifests differently in different settings (the most common explanation)
2. **Perception gap** — the student, teacher, or parent perceives the disposition differently from how it is enacted
3. **Developmental transition** — the student is between levels and their behaviour is inconsistent because the disposition is actively developing

Instruct the teacher NOT to resolve contradictions by averaging or by privileging one source. Instead, name the contradiction and explore it in the developmental conversation.

**Sufficient evidence threshold:**

State the minimum evidence requirements: at least 5 teacher observation entries over at least 4 weeks showing a consistent pattern (or a meaningful inconsistency), plus at least one student self-reflection that generates specific examples. Parent input is valuable but NOT required for sufficiency. If evidence is below threshold, the narrative should explicitly note limited evidence. Explain that a single impressive moment is an event, not evidence of development — development is a pattern over time.

**Contextual variation:**

Explain that a student may demonstrate the disposition at different band levels in different contexts — and that this is developmentally normal, not a problem. Instruct the teacher to NAME the variation specifically (e.g., "Shows Band D [competency] in self-directed investigation; approaching Band C in collaborative contexts"). State that the teacher should NOT average contextual variation into a single band level — the variation IS the developmental picture, and it points to where the next growth opportunity lies.

### 5. Developmental Conversation Guide

**What to share at Band [X]:**

Write specific guidance on what to share and what to hold back for the band specified in the input. Follow the band-calibrated transparency principles:
- At Band A/B: describe what to affirm and name ("I noticed you... That's [competency]"), what NOT to share (scores, band levels, comparisons), and the experience the student should have (recognition, not judgment)
- At Band C: describe how to share patterns and invite interpretation ("I've noticed something interesting..."), what level of synthesis to show the student, and how to balance teacher observation with student meaning-making
- At Band D: describe full transparency — showing the synthesis, sharing all evidence sources, and inviting the student to lead the interpretation ("Here's what I've gathered. What patterns do you see?"). Explain the teacher's role shift from telling to facilitating.

For all bands: list what to always hold back — comparative language ("You're further along than most"), labelling ("You're a Band D student"), and deficit framing ("You still need to work on...").

**Coaching questions:**

Write 3–4 specific coaching questions, fully written out as the teacher would say them. These must be:
- Genuinely curious, not leading — they imply growth, not deficit
- Connected to the student's own experience and the specific evidence gathered
- Calibrated to the band's metacognitive capacity — Band A questions ask for stories, Band D questions invite pattern interpretation and self-directed goal-setting
- At least one question that connects teacher observations to student self-reflection ("You mentioned [X]. I noticed something similar when [Y]. What do you think is behind that?")

**Using band descriptors as vocabulary:**

Provide a specific example showing how to reference the band descriptor in conversation as shared vocabulary rather than grading criteria. Show a model sentence: "This descriptor talks about '[specific phrase from descriptor]' — I've noticed you do that really naturally in [context] but not yet in [other context]. Let's think about why that might be." Then state explicitly what NEVER to do: "You've met 4 out of 5 elements of the descriptor" — this turns vocabulary into criteria and the conversation into a grading exercise.

**Goal-setting:**

Describe how to close the conversation: the teacher asks the student to identify one thing they want to work on, stated in their own words (not the band descriptor's words). Provide 1–2 example prompts for eliciting the goal. Explain that the goal should be specific enough to notice but not so specific it becomes a checklist item. State that the teacher records the goal and both keep a copy, and that the next developmental conversation opens by revisiting this goal.

### 6. Integration with Feedback Structures

If a school feedback structure was provided in the input, design a specific integration plan:
- **Before the session:** What the teacher prepares (synthesis completion, evidence selection, coaching question selection) and how long this takes
- **During the session:** A suggested minute-by-minute structure for the dispositional portion of the session, showing who speaks, what they share, and where the coaching questions and goal-setting fit. Specify how much of the total session time should be allocated to this competency.
- **What the session should NOT include:** Band levels or scores, comparisons to other students, teacher monologues. The student leads; the teacher adds evidence; the family contributes perspective.

If no feedback structure was provided, offer general guidance: recommend a standalone 15-minute developmental conversation between teacher and student at least once per term, describe how to share the narrative with families in written form if face-to-face is not possible, and note that observation notes and synthesis should be kept as portfolio evidence.

### Time and Workload Estimate

Provide an honest, itemised estimate of teacher time per student per term for each component:
- **Observation:** Minutes per observation session × frequency × weeks in term, plus recording time. State the total.
- **Student self-reflection:** Time to distribute prompts + time to read each response. State the total.
- **Parent input:** Time to send the guide + time to read responses (if received). State the total, and note that it is zero if the family does not respond.
- **Synthesis:** Time for the four-step synthesis process. Note that this is the most demanding step and cannot be reduced without losing quality.
- **Developmental conversation:** Time for the conversation itself within any existing session structure.
- **Total per student per term:** Sum the components. Then multiply by class size to show the full-class investment. Be honest — if the total is significant, say so. Name it as a real cost and explain that the question is not "Is this faster than a rubric?" (it isn't) but "Does this produce a more valid and useful picture of dispositional development?" (it does).
- **Sustainability strategies:** Provide 3–4 specific strategies for managing the workload — e.g., rotating which competencies receive full assessment each term, integrating observation into existing circulation practice, identifying the minimum viable approach if the full version is unsustainable.

**Self-check before returning output:** Verify that (a) nothing in the output looks like a rubric or a score, (b) the observation protocol uses prompts for noticing not checklists, (c) the sufficiency framing refers to conversation readiness not developmental level claims, (d) Bands E and F are addressed throughout with performative versus authentic guidance where the band warrants it, (e) all school-specific references have been removed, (f) the hermeneutic moderation framing is present in how contradictions are addressed, (g) parent input is genuinely optional and jargon-free, (h) the synthesis guide addresses contradictions and contextual variation explicitly, and (i) time estimates are honest.

示例输出

场景： 能力名称：“自主性” / 能力定义：“能够以日益增强的独立性和目标感发起、规划并完成有意义的工作” / 等级：“D 级（12–14 岁）” / 等级描述：“我能够发现采取行动的机会，并负责规划和执行自己的项目；在需要时寻求支持，并根据反馈和不断变化的情况调整自己的方法” / 评估背景：“学期末，为学生、教师和家长共同参加的发展性评审会议做准备” / 学校反馈形式：“发展性评审——每学期末举行一次 30 分钟的会议，由学生、导师教师及其父母/照护者参加。会议由学生主导，展示其学习和发展的证据。教师提供反馈，家庭成员提问并分享他们的观点。”

---

发展性评估设计：自主性——D 级

能力： 自主性 定义： 能够以日益增强的独立性和目标感发起、规划并完成有意义的工作 等级： D 级（12–14 岁） 评估背景： 学期末，为学生、导师教师和家长共同参加的发展性评审会议做准备

1. 观察协议

观察要点：

以下内容是整个学期中值得关注的提示——是需要留意的模式，而非需要逐项勾选的清单。你不会在每个学生或每种情境中都看到所有这些表现。留意哪些表现出现了，哪些没有。

1. 无需提示便主动开始。 学生能否识别需要完成的事情，并在没有等待指示的情况下着手行动？例如：在项目说明尚未完全讲解时就开始调研；发现小组计划中的问题并提出；或在当前方法行不通时提出替代方案。关键区别在于：独立完成别人要求的事情（服从），与识别需要做什么并采取行动（能动性）。

2. 会调整的计划。 学生是否会制订计划，并在情况变化时加以修改？留意僵化计划（即使原计划显然行不通也坚持照办）与适应性计划（根据过程中获取的信息调整时间安排、方法或范围）之间的区别。D 级能动性包括评估自身计划的元认知能力。

3. 把寻求支持视为优势。 学生是否会有策略地寻求帮助——明确自己需要什么、谁能提供帮助，并在适当的时候提出请求？这不同于依赖（每个决策点都寻求帮助），也不同于虚假的独立（明明求助会有帮助却独自苦苦挣扎）。D 级能动性意味着知道何时需要支持，并有信心主动寻求支持。

4. 表达归属感的语言。 留意学生如何谈论自己的工作。“我的项目”而非“这个项目”。“我决定……”而非“我们不得不……”。“这一部分我还不满意”而非“这样够好吗？”从被动表达转向能动表达，往往先于行为上的转变，也可能表明学生具备能动性，只是自己尚未意识到。

5. 面对挫折时的反应。 出问题时会发生什么——实验失败、计划落空、收到尖锐反馈？D 级能动性包括把挫折视为信息而非失败的能力。留意这样的区别：有的学生会重新思考方法，有的学生则会放弃项目或等老师来解围。

6. 关注范围。 学生的能动性是否延伸到自己的工作之外？他们会不会留意小组何时需要帮助、某项资源何时可以帮助他人、项目何时能服务更广泛的目标？充分展现的 D 级能动性还包括担当——为自身以外的事物谋求福祉。

观察时间与方式：

• 频率： 项目活动期间，每周至少两次有意识地关注能动性。这并不意味着要进行正式观察，而是指带着这些提示，有意观察 5 分钟。
• 最适合观察的情境： 开放式项目阶段（没有教师指定的下一步时）、遇到困难或模糊情形时、项目阶段转换时，以及学生真正可以自主决定下一步做什么的时刻。
• 记录格式： 简短的带日期记录——最多 1–3 句话。“3 月 14 日：S. 发现水传感器数据与预期不符。没有问我，而是回去检查校准流程，找到了错误，重新校准并再次测试。没有人询问，他就把事情经过告诉了小组。”将记录保存在持续更新的文档或纸质笔记本中。时间戳比细节更重要——你要随着时间推移积累观察模式。

缺失表现指标：

留意学生是否持续表现出不主动开始、不调整计划、不寻求支持，或不承担挫折带来的责任。在合理预期学生可能展现能动性的情境中，能动性的缺失同样是证据。一个学生在科学活动中能动性很强，却从不主动参与协作写作，这说明能动性会受情境影响，而非说明其缺乏能动性。

2. 学生自我反思工具

给学生：

这些问题是为 D 档学生设计的，适用于 12–14 岁。他们具备元认知能力，可以解读自己的行为模式，而不只是报告发生过的事情。

反思一下你这学期的自主性。慢慢来，没有标准答案。重要的是，你要诚实地思考自己的行为模式。

1. 想想这学期中，你曾主动采取行动去做一件对你很重要的事，而没有人要求或告诉你去做。 当时是什么情况？你做了什么？是什么让你决定采取行动，而不是等待？

2. 想想有一次你的计划行不通。 发生了什么？你做了什么？现在回想起来，你会采取不同的处理方式吗？

3. 当你需要别人帮忙时，通常会怎么做？ 你如何决定是自己想办法解决，还是向别人求助？这学期以来有变化吗？

4. 你现在在哪些方面最能发挥自主性？ 在哪些情境中——比如学科、工作类型或小组活动——你最容易主动行动并对自己的工作负责？在哪些情境中最难？你觉得其中的差异是为什么？

5. 如果把现在的自己和这学期初的自己做比较，你处理学习任务的方式有哪些变化？ 变化可能很大，也可能很细微。如果没有变化，也值得思考一下。

教师使用指南：

• 在发展评估会议前 2–3 天把反思提示给学生。应给他们时间思考，而不是让他们在压力下作答。
• 学生可以书面作答、录制音频，或以谈话记录的形式回答——选择他们最自在的方式即可。形式并不重要，反思的质量才重要。
• 切勿给反思打分。阅读时要关注其中的证据——故事、例子和行为模式，看它们是否印证或补充了你的观察。
• 如果学生的自我反思与你的观察相矛盾，这是一条有价值的信息，而不是错误。这可能说明学生体验到的自主性与你看到的不同，也可能说明自主性在你没有观察到的情境中以不同方式表现出来。

3. 家长/照护者意见指南（可选）

致家庭的介绍：

亲爱的家长和照护者：

作为学期末反思的一部分，我们希望了解您观察到的 [student name] 在独立性和主动性方面的成长情况——我们称之为“主体性”。主体性关乎主动承担有意义的工作：开始做事、制订计划、遇到意外情况时灵活调整，以及在需要时寻求帮助。

我们不是要您评估或打分，而是想了解您在家中和社区里观察到的情况。您的视角能帮助我们全面了解情况——您能在我们看不到的情境中观察 [student name]。

欢迎您分享任何观察。如果现在不方便回复，也完全没关系——我们还有其他证据来源，即使没有这部分意见，也能完成对您孩子的评估。

问题：

1. [student name] 是否会在家中或社区里主动承担项目、任务或责任，而不需要别人提出要求？能否举一个近期的例子？

2. 当 [student name] 遇到困难时——无论是家庭作业、个人挑战，还是新技能——他们通常会怎么做？（例如：马上求助、尝试独自解决、放弃，或想出变通办法？）

3. 本学期，您是否注意到 [student name] 在独立性或主动性方面有所变化——无论是在家中，还是在他们谈论学校的方式上？

4. 关于 [student name] 对待自己项目或责任的方式，是否有任何您认为我们应该了解的情况？

使用说明：

家长意见将与教师观察和学生自己的反思结合起来，帮助我们全面了解学生的发展情况。这些意见会在学生在场的情况下，于发展评估会议上分享——不会有任何隐瞒。如果有些内容比较敏感，请注明，导师教师会谨慎处理。并非所有家庭都会回复，这完全没问题——无论是否收到这部分意见，本评估方法都能正常开展。

4. 综合指南

整合证据：

在发展评估会议之前，导师教师会收集三个证据来源：

1. 教师观察记录——整个学期的持续记录（见上文《观察流程》）
2. 学生自我反思——对反思提示词作出的书面或录音回答
3. 家长/照护者意见——如有提供

综合流程：

步骤 1：分别阅读所有来源。 先独立阅读每个来源，不要交叉对照。分别记录每个来源中最突出的 2–3 个主题。

步骤 2：找出一致之处。 不同来源在哪些方面相互印证？如果教师看到学生独立发起研究，学生将“开始自己的实验”列为一项优势，而家长报告说“她最近会在家里自行安排学习计划，不需要我们提醒”——这就是学生在主动发起方面具有主体性的一致证据。

步骤 3：找出差异之处。 不同来源在哪些方面存在分歧？这是信息量最大的一步。主体性方面常见的差异模式包括：

• 教师认为学生主体性很强；学生却表示感到不确定 → 学生可能在外在表现上展现出主体性，但内心却缺乏安全感。这在 D 段发展阶段很正常，说明学生需要肯定，而不是被催促。
• 学生认为自己主体性很强；教师却观察到学生依赖他人 → 学生可能认为自己独立，但实际行为尚未体现这一点。这不是不诚实，而是一种渴望实现的自我认知，是发展的先兆。
• 家长认为学生在家中表现出主体性；教师却观察到学生在学校表现被动 → 主体性可能取决于具体情境，或许与在学校感受到的安全程度、社交动态或任务类型有关。 步骤 4：描述模式。 撰写一段 3–5 句的发展叙述，说明学生目前所处的阶段、证据显示的情况，以及发展轨迹。示例：“Priya 在自主研究中表现出很强的自主性——她总能提出自己的问题、规划调查步骤，并在数据与预测不符时调整做法。在协作情境中，她的自主性不那么明显；面对更强势的小组成员时，她往往选择退让，而不是为自己的想法争取。她和父母都指出，家中也有类似的模式：独自活动时很独立，在小组情境中则不够自信。下学期的发展重点是在他人面前展现自主性。”

处理矛盾：

不同来源之间的矛盾是有诊断价值的信息。通常有以下三种原因：

1. 情境依赖 — 这种倾向确实会在不同情境中表现出不同面貌。这是最常见的解释，在发展过程中也很正常。
2. 认知差异 — 学生、教师或家长对这种倾向的感知与其实际表现不同。这也很正常，可以在发展对话中进行探讨。
3. 发展过渡 — 学生正处于发展这种倾向的过程中；由于尚在不同发展水平之间，其行为会前后不一。过渡期间的不一致是发展特点，并非缺陷。

切勿通过取平均值或偏重某一来源来消解矛盾。应指出矛盾，并在发展对话中进行探讨。

证据充分性门槛：

要开展有充分依据的发展对话，最低证据要求是：至少在 4 周内，从不同情境收集 3 至 5 条独立观察记录，显示出一致的模式，或有意义且反复出现的不一致。这并非提出发展水平判断的门槛，而是将有结构的证据带入对话的最低依据。一次令人印象深刻的表现只是一个事件，而非一种模式。

家长的意见很有价值，但并非满足证据充分性的必要条件。如果证据低于这一最低要求，发展叙述应明确指出证据有限：“根据我目前掌握的观察记录，我看到 [pattern]——但我的数据有限，这幅图景可能并不完整。”

一次失常不会否定发展轨迹，正如单独一次表现出色也无法证明发展轨迹的存在。发展是随时间形成的模式，只有通过多次独立观察才能看清。

情境差异：

学生可能在科学项目中表现出 D 档自主性，在协作写作中则表现出 B 档自主性。这很正常。自主倾向的发展并不会在所有情境中保持一致。

综合分析应具体指出这种差异：“在自主探究中表现出 D 档自主性；在协作决策中正接近 C 档。”切勿将这些表现取平均值并归为单一档次。情境差异本身就是发展图景，也指明了下一步的发展机会。

5. 发展性对话指南

Band D 阶段可以分享的内容：

Band D 阶段的学生（12–14 岁）已准备好了解全部情况。可以分享：

• 你的观察模式——你注意到了什么、何时注意到的，以及是在什么情境下
• 他们自我反思的主题——他们的回答中哪些内容让你印象深刻
• 家长提供的任何信息——和学生一起阅读，而不是提前阅读
• 各方观点的一致之处和分歧之处——哪些方面一致，哪些方面不一致

邀请他们主导解读：「这是我从多个来源收集到的信息。你看到了哪些模式？哪些地方让你感到意外？哪些内容让你觉得确实如此？」

在 Band D 阶段，学生应逐渐具备评估自身发展倾向的能力。你的角色从「我来告诉你目前处于什么阶段」转变为「我们一起来看看这些证据，并理解其中的意义」。这是辅导，不是评分。

Band D 阶段应暂缓分享的内容：

即使完全透明，也要避免：

• 比较性措辞：「你的进步比大多数学生都快」或「有些同学比你更稳定」。发展因人而异，不是竞争。
• 贴标签：「你是 Band D 能动性学生。」各阶段的描述语是讨论发展的词汇，不是身份类别。
• 从不足出发的表述：「你还需要在……方面下功夫。」可以改为：「接下来你想在哪方面继续发展？」

适用于 Band D 阶段的辅导问题：

1. 「把这些证据放在一起看——你的反思、我的观察，以及家人分享的内容——你注意到目前自己的能动性有哪些模式？」

2. 「你提到了[自我反思中的具体例子]。我在[具体观察]时也注意到了类似情况。你觉得这种模式背后的原因是什么？」

3. 「你在哪些时候最能发挥能动性——最能主导自己的学习？哪些时候会觉得更难？这些情境有什么不同？」

4. 「如果想象一下下学期结束时的自己，你希望自己的能动性达到什么样的状态？和现在相比，会有哪些不同？」

将各阶段描述语用作共同语言：

Band D 阶段的能动性描述语是：「我能够发现采取行动的机会，并承担起规划和执行自己项目的责任；在需要时寻求支持，并根据反馈和变化的情况调整自己的方法。」

把这段描述当作共同语言，而不是检查清单：

• 「这段描述提到了‘根据反馈调整自己的方法’。我看到你在科学课的作业中很自然地做到了这一点——传感器数据出现偏差时，你没有等别人提醒，就回去重新校准了。你觉得自己还在哪些方面会这样做？哪些时候会更难？」
• 「这段描述提到了‘在需要时寻求支持’。你在反思中说过，你更喜欢自己想办法解决问题。你觉得这是独立自主，还是说，不主动求助有时也会带来代价？」

绝不要说：「你达到了 Band D 描述语中的 5 项标准里的 4 项。」这会把共同语言变成评判标准，也会把对话变成评分练习。

设定目标：

通过请学生用自己的话说出一件他们想努力的事情，为这次对话收尾。

• 「根据我们今天谈过的所有内容，下学期你想在哪方面发展自己的能动性？」
• 学生说出目标，教师将其写下来。双方各留一份。
• 目标应具体到可以观察（「我想在小组规划讨论中更积极地表达意见」），但不要具体到变成检查清单上的事项。
• 在下学期的首次发展性对话开始时，重新回顾这个目标：「上次你说想要[目标]。进展如何？」

6. 与反馈结构整合

与发展性评估会议整合：

发展性评估是一场 30 分钟的会议，由学生、导师教师和家长／照护者共同参与，并由学生主导。以下是整合倾向性评估的方法：

会前（教师准备——20 分钟）：

• 完成综合分析（上文步骤 1–4）。撰写一段 3–5 句话的发展叙述。
• 找出 2–3 个具体的观察时刻，作为可分享的实例。
• 准备引导问题——从上面的列表中选择 2 个你认为与这名学生最相关的问题。

会议期间（自主性环节的建议安排——30 分钟中占 8–10 分钟）：

1. 学生陈述（3 分钟）：学生分享自己的反思——本学期他们观察到自己的自主性有哪些表现，并提供具体实例。他们可以将书面反思作为笔记使用。

2. 教师回应（2 分钟）：分享 1–2 个具体的观察时刻。“我想补充一个我注意到的情况——[具体日期的实例]。这让我印象深刻，因为[与某种模式的关联]。”然后分享一个相互印证之处：“有意思的是，你的家人在家里也注意到了类似的情况——[家长的观察]。”

3. 家庭回应（2 分钟）：家长／照护者分享他们的看法。教师可以温和地引导：“你们在书面反馈中提到[观察]。能否再多说一些？”

4. 共同探讨（2 分钟）：教师提出一个引导问题，邀请三方共同思考：“你们都觉得[学生]的自主性在哪些方面有所提升？接下来可以在哪方面进一步成长？”

5. 目标（1 分钟）：学生用自己的话说出下学期的一个自主性目标。所有人都能听到。

发展性评估会议中不应包含的内容：

• 自主性的等级或分数。家人离开时应了解孩子的发展模式，而不是孩子的评分。
• 与其他学生或班级平均水平进行比较。
• 教师长篇陈述学生的优点和缺点。会议由学生主导，教师补充证据，家庭分享看法。

如果没有正式的反馈结构：

• 将发展性交流（第 5 节）作为一次独立的 15 分钟师生会议，每学期至少进行一次。
• 如果无法面对面交流，可用书面形式向家长／照护者分享发展叙述，以及学生的自我反思和目标。
• 将观察笔记和综合分析作为学生作品集的一部分保存——这些是今后讨论倾向性发展时的证据基础。

时间与工作量估算

**观察：**每周两次，每次进行 5 分钟的有意识观察，持续整个学期。记录：每条观察记录用时 1–2 分钟。**每名学生每学期的观察与记录总时长：约 60–90 分钟，**分散在 10–12 周内。工作量不小，但如果教师把观察记录视为简短的页边注，而非正式报告，就可以长期坚持。

**学生自我反思：**教师分发提示词需 5 分钟。阅读每名学生的回答需 10 分钟。每名学生总计：15 分钟。

**家长意见：**发送指南需 5 分钟。阅读每份回复（如果收到）需 5 分钟。**每名学生最多总计：10 分钟，**如果家庭没有回复，则为 0 分钟。

**综合分析：**每名学生需 15–20 分钟完成四步综合分析并撰写发展叙述。**这是工作量最大的步骤。**若要保证质量，就无法缩短。这也是最需要教师运用专业判断的环节。

**发展性对话／发展性评估中的能动性环节：**在现有的会谈中，每名学生需 8–10 分钟。

**每名学生每学期总计：约 100–135 分钟。**一个 25 人的班级每学期约需投入 42–56 小时，工作量不小。这是实际情况。与基于量规的评估相比，多信息源的倾向性评估更耗费人力。问题不在于“这样更快吗？”（并不更快），而在于“这样能否更有效、更有用地呈现倾向性发展情况？”（可以）。

可持续性策略：

• 不必每学期都用这种完整方法评估每项能力。可以轮换：第一学期深入评估能动性，第二学期评估协作，第三学期评估自我调节。
• 可以将观察记录融入现有做法中——在项目活动期间巡视时进行观察，而不是将其设为单独任务。
• 如果工作量确实难以持续，最低可行方案是：教师观察 + 学生自我反思 + 发展性对话。舍弃家长意见和正式综合分析，但保留对话。

---

评估专业能力与新教师入职培训

观察规程有助于培养评估专业能力，但无法替代它。评估专业能力——即能切身辨别真实的自我调节与表演出来的自我调节，或真正的共情与符合社交规范的共情表现——需要通过经验、反思以及与优秀同事共同校准来培养。首次使用这一规程的教师，其观察质量会低于拥有多年经验的教师；这并非因为他们不够认真，而是因为他们尚未形成这种切身的判断力。这种专业能力或可称为教学观察知识——它有别于教学内容知识，却鲜少在教师教育中被命名或教授。

实施这项技能的学校应为新教师提供：由资深从业者撰写的观察记录示例，并讨论这些观察为何有用；定期开展校准交流，让教师分享并比较对同一学生的观察；以及在将观察用于任何具有重大影响的用途之前，安排专门时间培养教师的观察实践。

---

观察者立场与专业反思

教师对学生特质的观察，受到教师自身与这些特质之间关系的影响。自我调节主要通过安静的内在处理来体现的教师，可能较少注意到通过大声、外显方式进行调节的学生。来自高语境文化的教师，可能会把学生含蓄的边界表达误解为不投入。重视口头表达能力的教师，可能更容易注意到自信的自我反思，而不易注意到谨慎却真诚的自我反思。

这是所有观察固有的特点，并非不够专业。恰当的应对方式是进行专业反思：教师应定期自问，哪些特质最容易被自己识别和重视，哪些特质则较难被自己注意或欣赏。教师之间的校准交流，是随着时间推移揭示并纠正观察者立场的结构性机制。

---

已知局限

1. 多方信息源评估能提高准确性，但也会增加工作量。 完整流程——观察、学生自我反思、家长反馈、综合研判、发展性对话——每学期每名学生约需 100–135 分钟。对于有 25 名学生、且需要评估多项能力的教师来说，这是一笔不小的投入。这项技能采用模块化设计——任何环节都可以单独使用——但要获得完整的三角互证图景，就需要投入全部时间。学校实施这种方法时，应明确规划所需时间，而不是到学期中途才发现时间不够。

2. 家长／照护者的反馈很有价值，但并非总能获得，也并非总是合适。 有些家庭因工作安排、语言障碍或家庭状况而无法参与。有些家庭情况也使家长不宜或不安全地提供有关学生性情发展的反馈。即使没有家长反馈，这种评估方法仍然完全有效——家长反馈是真正可选的补充，而非要求。若获取家长反馈可能危及学生安全或损害家庭信任，教师绝不应继续征求。

3. 性情发展并非线性，且受情境影响。 一名学生在科学课上可能展现 D 档自主性，在协作写作中却只有 B 档。上学期自我调节能力很强的学生，本学期可能因个人情况、社交动态或新型任务的要求而显得退步。这项技能会明确指出这一点，并指导教师描述情境差异，而不是取平均值——但它无法避免机构要求生成单一“自主性评分”用于报告的压力。如果学校的报告系统要求填写单一等级，教师必须运用专业判断，选择最能代表学生整体发展轨迹的等级，同时在叙述性备注中说明情境差异。

4. 发展性对话对教师的能力要求很高。 对话指南为对话结构提供支架，并提供辅导问题，但对话质量取决于教师倾听、真诚回应、为学生的自我评估留出空间而不予纠正或评判，以及维持关系信任的能力。机械照读辅导问题的教师，效果会比没有指南但具备良好关系直觉的教师更差。指南是支架，不能取代专业判断和建立关系的能力。学校在大规模实施这种方法之前，应考虑提供有关辅导式对话的专业发展培训。

5. 两位教师对同一名学生得出不同结论时，这反映的是情境差异，而非评分者间信度失效。 两位观察者可能都准确：他们在不同情境中观察到了这种性情特质的不同方面。应对方式是开展校准对话，探讨这些差异揭示了学生在不同情境中的性情发展情况——不是对观察结果取平均值，也不是判定谁对谁错。不同情境中的差异模式本身就是最有价值的证据。这种诠释学式的校准方式，将不同观察视为专业探究的契机，而非需要解决的评分问题；这与心理测量信度框架有根本区别，后者假设存在一个唯一的真实分数，可靠的评分者会逐渐趋于一致。性情发展并不存在单一的真实分数；它是在不同情境中展现出的行为模式，而教师之间不同的观察结果，可能都如实呈现了这一模式的真实部分。