📚 How Schools Identify and Support Students with Special Educational Needs: What the Research Reveals | 学校如何识别与支持特殊教育需要学生：研究揭示的支持做法
Every classroom contains learners who read, remember, attend and regulate their emotions in different ways, and a minority of those learners will need something additional to, or different from, the teaching that works for the majority. The question that education research has spent forty years trying to answer is a practical one: how do schools spot those learners early and accurately, and what kinds of support actually change their outcomes rather than merely labelling them? This article sets out the identification cycle, the three-wave model of provision, and the specific support practices that research syntheses and national frameworks repeatedly identify as effective.
每一间教室里都有在阅读、记忆、注意力和情绪调节方面各不相同的学习者，其中少数人需要的是”额外的”或”不同于”主流教学的支持。教育研究用了四十年试图回答一个非常实际的问题：学校如何及早而准确地识别这些学生，什么样的支持才能真正改变他们的学习结果，而不只是给他们贴一个标签？本文将梳理识别流程、三层支持模型，以及研究综述与国家框架反复确认有效的具体支持做法。
1. Defining Special Educational Needs: Categories, Prevalence and Controversy | 界定特殊教育需要：类别、流行率与争议
In England, the SEND Code of Practice (2015) defines a child or young person as having special educational needs (SEN) if they have a learning difficulty or disability that calls for special educational provision to be made for them, meaning provision that is additional to, or different from, the provision made generally for others of the same age. The definition is deliberately relative and functional rather than medical: it is a statement about the gap between what a learner can currently do and what the ordinary curriculum and classroom demand.
在英国，2015 年《特殊教育需要与障碍实践守则》(SEND Code of Practice) 将”有特殊教育需要”界定为：学生存在学习困难或障碍，需要为其提供”额外的”或”不同于”同龄人一般性安排的特殊教育支持。这个定义刻意采用相对性与功能性而非医学式表述：它描述的是学习者当前能力与普通课程、普通课堂要求之间的差距。
Policy frameworks normally group need into four broad areas: communication and interaction (for example speech, language and communication needs, and autism); cognition and learning (moderate, severe or specific learning difficulties such as dyslexia and dyscalculia); social, emotional and mental health (anxiety, ADHD, attachment difficulties); and sensory and/or physical needs (hearing impairment, visual impairment, developmental coordination difficulties). A learner may appear in more than one area at once.
政策框架通常把需要划分为四大领域：沟通与互动（如言语、语言与沟通需要，自闭症）；认知与学习（中度、重度或特定学习困难，如阅读障碍、数学障碍）；社会、情绪与心理健康（焦虑、注意力缺陷多动障碍、依恋困难）；以及感官和／或身体需要（听觉障碍、视觉障碍、发展性协调困难）。同一名学生可能同时落在多个领域中。
Prevalence figures are instructive precisely because they are unstable. In England roughly one pupil in six is recorded as having SEN at some point in their school career, with around 4 to 5 per cent holding an Education, Health and Care Plan (EHCP), yet identification rates vary substantially between schools, regions and countries. That variation is the first clue that identification is not a neutral measurement of a child but a social and institutional process.
流行率数据之所以值得注意，恰恰是因为它并不稳定。在英格兰，大约每六名学生中就有一名在其求学阶段被记录为有特殊教育需要，其中约 4% 至 5% 持有”教育、健康与照护计划”(EHCP)；但不同学校、地区与国家之间的识别率差异相当大。这一差异提示我们：识别并不是对孩子进行的中立测量，而是一个社会性与制度性过程。
The classic tension is between the medical or deficit model, which locates the problem inside the learner, and the social model of disability, which locates it in the interaction between the learner and an environment designed for a narrow range of bodies and minds. Most contemporary frameworks adopt a hybrid position, which can be summarised as:
经典的理论张力存在于”医学／缺陷模型”与”障碍的社会模型”之间：前者把问题定位在学习者内部，后者把问题定位在学习者与”为狭窄范围的身体与心智而设计的环境”之间的互动中。当代多数框架采取折中立场，可概括为：
Need = f(learner characteristics × environmental demands × teaching quality)
需要 = f（学习者特征 × 环境要求 × 教学质量）
2. The Legal and Ethical Framework: From Segregation to Inclusion | 法律与伦理框架：从隔离走向融合
Modern identification systems are not left to individual teacher goodwill; they are legally framed. The Warnock Report (1978) introduced the term “special educational needs” into English policy and argued that up to one in five children would need some form of additional provision at some point. The Salamanca Statement (UNESCO, 1994) and Article 24 of the UN Convention on the Rights of Persons with Disabilities established inclusive education as a rights-based expectation rather than a charitable option.
现代识别体系并不依赖教师个人的善意，而是有法律框架支撑。1978 年《沃诺克报告》把”特殊教育需要”一词引入英国政策，并提出最多五分之一的儿童在某个阶段需要某种形式的额外支持。1994 年联合国教科文组织《萨拉曼卡宣言》以及《联合国残疾人权利公约》第 24 条，则把融合教育确立为一项基于权利的期待，而非慈善式的选择。
In England, the Children and Families Act 2014 and the SEND Code of Practice 2015 set out a graduated approach, a maximum 20-week statutory assessment timetable for EHCPs, and the requirement for annual review. The Equality Act 2010 adds a duty to make reasonable adjustments and prohibits discrimination arising from disability. In the United States, the Individuals with Disabilities Education Act (IDEA) guarantees a free appropriate public education in the least restrictive environment, with an Individualised Education Program (IEP) for every eligible student.
在英格兰，2014 年《儿童与家庭法》与 2015 年《特殊教育需要与障碍实践守则》确立了阶梯式支持流程、EHCP 最长 20 周的法定评估时限，以及年度复查的要求。2010 年《平等法》进一步规定了”合理调整”义务，并禁止基于障碍的歧视。在美国，《残疾人教育法》(IDEA) 保障每位符合条件的学生在”最少限制环境”中获得免费适当的公立教育，并拥有个别化教育方案 (IEP)。
Four principles recur across jurisdictions and are worth memorising for essay questions: early identification and intervention; the participation of the learner and the family in decisions; provision that is additional to or different from ordinary teaching; and placement in the least restrictive setting consistent with good progress.
跨法域反复出现的四条原则值得为论述题背下来：及早识别与干预；学习者与家庭参与决策；提供的支持须”额外于或不同于”普通教学；以及在能取得良好进展的前提下，安置于限制最少的环境。
3. Stage 1: Universal Screening and Early Warning Signs | 阶段一：全员筛查与早期预警信号
Identification begins with universal screening: brief, low-cost measures applied to all learners rather than only to those already causing concern. Common tools include curriculum-based measurement of reading fluency (words correct per minute), phonics and decoding checks, numeracy and number-fact fluency probes, oral language screeners, spelling and writing samples, and routine vision and hearing checks.
识别始于”全员筛查”：对全体学生而非仅对已被关注的学生实施简短、低成本的测量。常用工具包括基于课程的阅读流畅度测量（每分钟正确读词数）、自然拼读与解码检测、数感与算式流畅度探测、口语语言筛查、拼写与写作样本，以及常规的视力与听力检查。
Screening is not diagnosis. Its purpose is to sort learners into “likely to need additional support” and “probably not”, accepting that some errors will occur in both directions. The quality of any screening tool is therefore described by two statistics:
筛查不等于诊断。它的目的是把学生分为”可能需要额外支持”与”大概率不需要”两类，并接受两个方向上都会出现误差。因此，任何筛查工具的质量都由两个统计指标来描述：
Sensitivity = true positives ÷ (true positives + false negatives)
特异性 = 真阴性 ÷（真阴性 + 假阳性）
A test with high sensitivity catches nearly every learner who genuinely needs help but flags many who do not; a test with high specificity does the reverse. Research on response-to-intervention systems generally recommends screening measures with sensitivity above roughly 0.90 and specificity in the region of 0.80 or better, combined with a second-stage verification for anyone who fails the first screen.
高灵敏度的工具能捕捉到几乎所有确实需要帮助的学生，但也会误报许多并不需要的人；高特异性的工具则相反。回应性干预 (RTI) 研究通常建议：筛查工具的灵敏度应在 0.90 以上，特异度约为 0.80 或更高，并对初筛未通过者进行第二轮确认性评估。
Alongside formal screening, schools rely on structured teacher referral. Warning signs that research and guidance consistently highlight include a plateau in progress despite good teaching, a growing gap with peers, avoidance of reading aloud or written tasks, frequent unexplained absence, sudden changes in behaviour or mood, difficulty following multi-step instructions, and persistent difficulty with peer relationships.
除正式筛查之外，学校还依赖结构化的教师转介。研究与指引反复强调的预警信号包括：在教学质量良好的情况下进展仍然停滞；与同伴的差距不断扩大；回避朗读或书写任务；频繁且原因不明的缺课；行为或情绪的突然变化；难以完成多步骤指令；以及与同伴关系持续困难。
4. Stage 2: The Graduated Approach: Assess, Plan, Do, Review | 阶段二：阶梯式方法：评估、计划、实施、复查
The graduated approach is the operational heart of SEN practice. Rather than treating support as a single decision, it treats it as a cycle that is repeated, with increasing intensity and specialist involvement, until the learner makes adequate progress. The four stages are assess, plan, do and review, and a full cycle should normally be completed at least three times a year.
阶梯式方法是特殊教育实践的运行核心。它不把支持看作一次性决定，而是一个不断重复的循环，随需要提高强度并引入专业人员，直到学生取得足够进展。四个环节是”评估、计划、实施、复查”，一个完整循环通常每年至少完成三次。
| Stage | 环节 | Key actions | 关键动作 | Evidence to keep | 需保留的证据 |
| Assess | 评估 | Class teacher and SENCO build a clear analysis of need using assessment data, observation, work samples and pupil/parent views | 班级教师与 SEN 协调员结合测评数据、观察、作业样本与学生／家长意见，形成对需要的清晰分析 | Baseline scores, dated observations, pupil and parent statements | 基线成绩、带日期的观察记录、学生与家长陈述 |
| Plan | 计划 | Agree 2-4 measurable targets, the intervention, who delivers it, and the review date | 商定 2 至 4 个可测量目标、干预方式、实施者与复查日期 | Written plan or IEP with target dates | 写明目标日期的书面计划或 IEP |
| Do | 实施 | Class teacher remains responsible for delivery and daily adaptation; support staff carry out targeted work | 班级教师仍对教学实施与日常调整负责；助教执行针对性教学 | Attendance and fidelity records, session logs | 出勤与执行保真度记录、课时日志 |
| Review | 复查 | Compare progress against baseline, decide to continue, change or escalate | 将进展与基线比较，决定继续、调整或升级支持 | Post-intervention scores, revised plan | 干预后成绩、修订后的计划 |
A common failure mode is a “review” that only reports attendance at sessions rather than learning outcomes. Research on intervention effectiveness is clear that the review must be a data comparison: what was the starting point, what was expected, and what was achieved.
常见的失败模式是”复查”只汇报学生出席了几节课，而不汇报学习结果。关于干预效果的研究十分明确：复查必须是数据比较——起点是什么、预期是多少、实际达成为何。
5. Assessment Tools: Standardised Tests, Observation and Dynamic Assessment | 评估工具：标准化测验、观察与动态评估
Standardised, norm-referenced tests allow a school to compare a learner with a large representative sample of the same age. Their results are usually reported as a standard score with a mean of 100 and a standard deviation of 15, or as a percentile rank. The conversion formula is:
标准化常模参照测验让学校可以把学生与同年龄的大样本群体比较。其结果通常以平均数为 100、标准差为 15 的标准分，或以百分等级呈现。换算公式为：
Standard score = 15 × (raw score − mean) ÷ standard deviation + 100
标准分 = 15 ×（原始分 − 平均分）÷ 标准差 + 100
Interpretation requires caution. Every test score carries measurement error, estimated by the standard error of measurement:
解读时必须谨慎。每一个测验分数都带有测量误差，可用测量标准误估计：
SEM = SD × √(1 − rₓₓ)
where SD is the standard deviation and rₓₓ is the test’s reliability. A standard score of 85 is not meaningfully different from 88 when the confidence band overlaps, and single-score decisions are therefore poor practice.
其中 SD 为标准差，rₓₓ 为测验信度。当置信区间重叠时，标准分 85 与 88 之间并没有实质差异，因此凭单一分数做决定是糟糕的做法。
Three further tools complement standardised testing. Systematic observation using a coding schedule captures attention, on-task behaviour and peer interaction in real classrooms. Error analysis examines the pattern of mistakes in reading or mathematics, distinguishing, for example, a decoding difficulty from a comprehension difficulty or a calculation error from a conceptual misunderstanding. Dynamic assessment measures how much a learner improves with structured prompts and scaffolding, giving a direct estimate of learning potential rather than accumulated attainment.
另有三类工具可与标准化测验互补。使用编码表的系统观察，能在真实课堂中记录注意力、任务专注行为与同伴互动。错误分析检视阅读或数学中的错误模式，例如区分解码困难与理解困难，或计算失误与概念误解。动态评估则测量学习者在结构化提示与支架下能提升多少，直接估计学习潜能而非已有的学业积累。
Finally, curriculum-based measurement provides short, repeated probes that are sensitive to small changes over weeks, which makes them ideal for monitoring progress during an intervention rather than for labelling a learner.
最后，基于课程的测量提供简短、可重复的探测，对几周内的小幅变化也敏感，因此非常适合在干预过程中监测进展，而不是用来给人贴标签。
6. What Research Says About Effective Identification | 研究揭示的有效识别做法
The most consistent finding in the identification literature is that no single source of evidence is sufficient. Universal screening alone produces false positives; teacher judgement alone is vulnerable to bias and to variation in classroom expectations. Combining screening data with structured teacher ratings, and then verifying with a second-stage assessment, produces substantially better accuracy than any one method.
识别研究中最一致的结论是：任何单一证据来源都不足够。仅靠全员筛查会产生假阳性；仅靠教师判断则易受偏见与课堂期望差异影响。把筛查数据与结构化教师评定结合，再进行第二轮确认评估，其准确性显著优于任何单一方法。
Research on disproportionality adds an important caution. Certain groups are identified at rates that cannot be explained by underlying prevalence alone: in England, for example, pupils eligible for free school meals, certain ethnic groups, and looked-after children are over-represented in some SEN categories; in the United States, English learners have historically been over-identified in some categories and under-identified in others. School systems are therefore advised to audit their own identification data by group, year and category.
关于”比例失衡”的研究提供了一项重要警示。某些群体的识别率无法仅由基础流行率解释：例如在英格兰，领取免费校餐的学生、某些族裔群体与受照护儿童在某些 SEN 类别中比例过高；在美国，英语学习者在部分类别中历史上被过度识别，在另一些类别中则被识别不足。因此建议学校按群体、年级与类别审核自身的识别数据。
| Practice | 做法 | Why research supports it | 研究支持的理由 |
| Multi-stage screening with verification | 多阶段筛查并设置确认环节 | Reduces false positives and avoids unnecessary labelling | 减少假阳性，避免不必要的标签化 |
| Curriculum-based measurement over time | 长期使用基于课程的测量 | Detects rate of learning, not just current level | 检测学习速率，而不仅是当前水平 |
| Assessment in more than one setting | 在多个情境中评估 | Separates context-specific difficulty from general difficulty | 区分
Published by TutorHao | Exam Prep Revision Series | aleveler.com 更多咨询请联系16621398022（同微信） CommentsMore posts |
屏轩国际教育cambridge primary/secondary checkpoint, cat4, ukiset,ukcat,igcse,alevel,PAT,STEP,MAT, ibdp,ap,ssat,sat,sat2课程辅导，国外大学本科硕士研究生博士课程论文辅导