GCSE ALevel 辅导 独立评估 效应量 一对一
英国的课外辅导到底有没有用，英国本地已经有一整套独立评估在回答，结论不是简单的有效或者无效。同一个方法，学段不同、学科不同、谁来讲不同，得到的效应量可以差出三倍。下面把 EEF 官网的效应量口径、国家辅导计划的官方数据、以及可以照着核对的做法梳理成十节，方便 GCSE 与 A-Level 阶段的学生和家长自己判断一次辅导值不值得报。文中的数字都能在 EEF 官网、gov.uk 与 DfE 的统计发布里找到原始出处。
In England the question of whether tutoring works is answered by a body of independent evaluations rather than by opinion. The same method produces very different results depending on the age group, the subject, and who delivers the sessions. This article sets out the official effect sizes published by the Education Endowment Foundation, the Department for Education figures on the National Tutoring Programme, and the classroom conditions that make tuition more or less likely to pay off.
独立评估是谁做的：EEF 与 NFER 的分工
关于辅导有效性的英国证据，主要来自 EEF 官网的 Teaching and Learning Toolkit 以及它委托的独立评估。EEF 是英国教育部资助的独立教育研究机构，也是 What Works Network 里负责教育的一家。它把随机对照试验与准实验研究汇总成一句话结论，效应量的单位不是分数，而是额外月数。额外月数的意思，是参加干预的学生比条件相近的未参加学生平均多走了几个月的学习进度。国家辅导计划第一年的 Tuition Partners 路线，独立评估由 NFER 承担，采用准实验设计。准实验设计的做法是把参加学校与条件相近的未参加学校配对比较，而不是随机分组。之所以要这样设计，是因为辅导从来不是随机分配的，家庭与学校的选择本身就会影响结果。看清这一层，就知道这类独立评估给出的是学校层面的平均效应，不是某一个学生必然受益的保证。
In England the evidence base for tutoring comes mainly from the EEF Teaching and Learning Toolkit and the evaluations it commissions. The toolkit turns trials and quasi-experiments into a single figure expressed in months of additional progress, measured by comparing pupils who received support with similar pupils who did not. The first-year evaluation of the National Tutoring Programme Tuition Partners route was carried out by NFER using a quasi-experimental design, matching participating schools to comparable schools outside the programme. That design choice matters because tuition is never assigned at random, where family and school decisions already shape who receives it; the honest question is therefore what happened on average across comparable schools.
一对一的效应量：平均约多出五个月
EEF 官网对一对一辅导的结论是，平均约多出五个月进度，证据强度中等，成本偏高。这个数字有一个关键条件，就是辅导必须附加于正课，并且与正课明确衔接。同一页面还给出被认为最优的节奏：每次约三十分钟，每周三到五次，持续不超过十周。这组参数的指向很直接，短而密比长而稀更容易见效。很多学生以为一次补两三个小时才算认真，实际上连续三小时的注意力曲线在第二个小时就开始下滑。常见的错误是每周只上一次、一次上三小时，然后指望八周后成绩跳一个等级。
The EEF toolkit rates one-to-one tuition as having an average impact of roughly five additional months, based on moderate evidence and at a moderate to high cost. Two conditions carry most of that effect, since the tuition has to be additional to normal lessons and explicitly linked to what is taught in class. The toolkit also describes the rhythm that appears to work best, with short sessions of about thirty minutes, three to five times a week, over a period of up to ten weeks. Long, occasional blocks of tutoring feel productive to parents, but they rarely survive contact with a tired pupil in the third hour.
小学六个月与中学四个月的差距
EEF 官网在一对一那一页单独标注了一条注释：小学阶段的研究显示约多六个月，中学阶段约多四个月。这条注释对 GCSE 学生尤其要紧，因为它说明同一个方法进到中学以后效应会减弱。中学阶段的知识更依赖先备内容，缺的那一环不补上，新课再多讲一遍也用不上。A-Level 阶段的课程更集中，能直接对应的官方效应量更少，可迁移的是做法而不是那个具体月数。做法上排在前面的是诊断，先找出具体断点，再决定补什么。很多学生以为辅导就是再听一遍课，于是把课堂内容重复一遍，问题章节仍然没有解决。常见的错误是拿着整套试卷从头讲到尾，却没有停下来确认是哪一步开始听不懂的。
The EEF toolkit adds a note that matters for older pupils: studies in primary schools tend to show about six months of additional progress, while secondary studies average four. Attainment gaps at secondary level usually trace back to prerequisite knowledge, so re-teaching the current topic without repairing the missing step seldom moves a grade. For A-level students the transferable lesson is the method rather than the month count, starting with a diagnostic task that locates the exact break in understanding. Repeating a lesson a pupil has already sat through is the most common way to spend money and gain nothing.
数学两个月与读写六个月：学科差异
EEF 官网的另一个注明是，一对一在数学上的效应约两个月，在读写上约六个月。同一个干预在两个学科上差三倍，说明效应量依赖具体内容，而不是辅导这个概念本身。数学的难点在于前置链条长，一个环节断了，后面整段都推不动，所以短周期诊断式教学比泛讲更有效果。读写类提升快，一部分原因是它能在同样的时间里覆盖更多可测的子技能。给 GCSE 数学或者 A-Level 数学找辅导时，先要求对方做一次错题定位，再谈课时包。很多学生以为数学补课就是多刷题，其实不定位断点，题量只会加深同一处错误。
The same toolkit page reports that effects in mathematics appear to be substantially lower, at around two additional months, than in literacy, at around six. A threefold gap inside one intervention is a reminder that impact depends on the content being taught rather than on tutoring as a label. Mathematics has long prerequisite chains, so a single broken step blocks everything that follows, which is why short diagnostic teaching outperforms broad revision. Before buying a block of maths sessions, ask for one diagnostic paper that shows where the chain actually breaks.
小班两到五人：四个月与成本权衡
EEF 官网把小班辅导定义为一个老师或者助教带两到五名学生，平均效应约四个月，证据中等，成本低。同一页面的判断是，小班与一对一都有效，但从成本效益看值得先用小班。小班有效的前提是针对具体需求，而不是把五个水平不同的学生凑在一起讲同一份材料。诊断性评估在这里的作用，是把需求相同的学生编到一组，让四个月效应有落点。人员培训也会放大效应，受过训练的老师或者助教带同样的小组，结果通常更好。很多学生以为小班就是便宜版的私教，其实组内水平越接近，平均每人的收益越高。
The toolkit defines small group tuition as one teacher or teaching assistant working with two to five pupils, with an average impact of about four additional months on low-cost, moderate-evidence grounds. Its conclusion is that one-to-one and small group tuition both work, yet cost effectiveness argues for trying small groups first. Grouping matters more than group size, because pupils who share the same specific gap gain the most, while a mixed-ability group taught one worksheet often gains little. Training the adults who deliver the sessions is one of the few reliable ways to raise the impact.
十五小时课与三十分钟节奏：NTP 的设计口径
国家辅导计划第一年的 Tuition Partners 路线，为每名学生每科安排最多十五小时辅导，多数采取一个导师带三名学生的形式，有额外需要的学生才排一对一。辅导可以安排在正常教学时间内，也可以放在课前或者课后。当年经过质量审核的合作机构共三十三家。这套设计把剂量写得很具体，十五小时拆成每周三次、每次三十分钟，大约就是十周。反过来说，如果一门课计划每周只有一次、每次九十分钟，那就已经偏离了官方给出的最佳节奏。很多学生以为课时越长越划算，实际决定收益的是注意力分配与间隔学习。
In its first year the National Tutoring Programme Tuition Partners route offered schools up to fifteen hours of tutoring per pupil per subject, mostly delivered as one tutor with three pupils, with one-to-one reserved for pupils who needed it most. Sessions could run inside the school day or before and after it, and thirty-three quality-assured partners delivered them across England. Fifteen hours split into thirty-minute sessions three times a week comes to roughly ten weeks, which is where the toolkit's recommended pattern lands. A single ninety-minute session each week is a different intervention with a weaker evidence base.
附加而非替代：与正课衔接
EEF 官网在两处都强调，辅导要附加于正课并且与正课明确衔接，效应才站得住。衔接的意思很具体：导师知道本周课堂上要讲什么，也知道学生上一份作业错在哪里。学校要监控进度，确认这段时间的辅导确实带来了变化，而不是把课时上完就算完成。如果辅导用的方法和课堂相反，学生会面临两套解法，考试时反而更容易犹豫。常见的错误是辅导老师自成一派，完全不管学生所在考试局的评分口径。GCSE 与 A-Level 的评分细则写在各考试局官网的 specification 里，导师至少要看一遍相关的评分点，家长也可以要求看到这份对照。
The toolkit repeats one condition in both of its tutoring entries: tuition is more likely to make an impact if it is additional to normal lessons and explicitly linked to them. In practice that means the tutor knows what is being taught this week and which questions the pupil got wrong last week, rather than running a private syllabus of their own. Schools are expected to monitor progress, so that tutoring is judged by what changed in the pupil's work. A tutor working from a method that contradicts the classroom leaves the pupil holding two versions of the same procedure and a slower decision in the exam hall.
谁来讲：助教、导师与结构化方案
EEF 官网对助教承担的一对一辅导给了一条限制：助教有经验、受过训练并且得到支持时，效果尤其明显，例如执行一套结构化的干预方案。结构化方案的标志是有固定顺序、有记录、有阶段性检查，而不是每次现场发挥。同一页面也提醒，由助教或者小班承担同样内容时，平均效应会比一对一低一些，但成本效益可能更好。这条提醒对预算有限的家庭有实际意义，把有限的课时用在有方案、有记录的年轻老师身上，通常比零散地上高价课更稳。很多学生以为老师头衔越大越好，其实方案与记录比头衔更能决定十周之后的结果。
The toolkit flags a boundary condition for teaching assistants: the interventions they deliver tend to work well when they are experienced, well trained and supported, for example when following a structured programme. Structure means a fixed sequence, a record of what was covered and checkpoints where progress is reviewed, rather than improvising each session. The same page notes that instruction delivered by assistants or in small groups shows smaller average effects than one-to-one tuition, while often offering better value for money. For a family paying per hour, a planned and recorded sequence beats an expensive but improvised one.
国家辅导计划四年：覆盖率与补贴退坡
按 gov.uk 公布的口径，国家辅导计划从 2020 到 2021 学年开始，到 2023 到 2024 学年结束，覆盖英格兰公办小学与中学。补贴比例从 2021 到 2022 学年的 75%，原计划降到 2023 到 2024 学年的 25%，2023 年 5 月 23 日 gov.uk 宣布改为 50%。2023 到 2024 学年主流学校每名符合条件的学生至少拿到 67.50 英镑，上一学年是 162 英镑，落差接近六成。同一学年主流学校每小时每名学生上限 18 英镑，其中教育部补贴 9 英镑。四年总投入超过十亿英镑，2023 年 5 月的官方口径是已开课超过三百万门。DfE 统计发布显示参与率从 2021 到 2022 学年的 87.4%，降到 2022 到 2023 学年的 73.8%，再降到 2023 到 2024 学年的 59.4%。这组数字说明补贴退坡直接压低了学校的使用意愿，也解释了为什么 2024 年以后学校的辅导更依赖家庭自己付费。
According to gov.uk, the National Tutoring Programme ran from the 2020/21 academic year to 2023/24 and covered state-funded primary and secondary schools in England. The subsidy tapered from 75% in 2021/22 towards a planned 25% in 2023/24, until the Department for Education announced on 23 May 2023 that it would fund half of tutoring costs instead. Allocations per pupil premium pupil fell from £162 in 2022/23 to a minimum of £67.50 in 2023/24, with a cap of £18 per pupil per hour, of which the DfE covered £9 in mainstream schools. Department for Education statistics show participation falling from 87.4% of schools in 2021/22 to 73.8% in 2022/23 and 59.4% in 2023/24. The programme was backed by over £1 billion across four years, and by May 2023 more than three million courses had started.
判断一次辅导是否有效：可以核对的信号
判断一次辅导值不值得继续，可以先看有没有诊断环节，也就是用错的题目定位断点，而不是直接排课。接着看目标能不能写出来，比如两个月内把二次函数配方这类题型稳定做对，而不是笼统地说提高数学。再看辅导与课堂是否对齐，导师是否知道学生所在考试局的评分口径与最近几次作业。周期上，短而密的设计通常比长而稀更容易验证效果，两到三周就该有一次可见的变化。如果六到八周过去，错题类型没有减少，需要调整的是方案，不是课时数量。很多学生以为报了班就等于提分，真正决定结果的是诊断、对齐与复盘这三件事有没有真的发生。
To judge whether a tutoring arrangement is working, start with the diagnostic step, which means locating the break in understanding from wrong answers rather than booking hours. A useful arrangement then states a measurable target, such as consistently solving quadratic equations by completing the square within two months, instead of the vague aim of improving at maths. Alignment comes next, and that means the tutor knows the mark scheme of the pupil's exam board and has seen the last few pieces of homework. If six to eight weeks pass without any change in the pattern of errors, the plan needs rebuilding rather than more hours.
海外直订Higher Mathematics for AQA GCSE 高等数学AQA GCSE¥645.0📱 去京东购买京东精选需要一对一辅导？Need One-on-One Tutoring?
A-Level / GCSE / IB / KS2-KS3 全科辅导，按考试局评分口径设计课程
16621398022 同微信
Visit aleveler.com
或截图后用「扫一扫 → 相册」选择本图