IB IA Evaluation评估部分写作指南
IB内部评估（Internal Assessment，简称IA）是IB文凭课程中最能拉开分差的部分之一。无论是数学、物理、化学、生物还是经济学，每个科目都要求学生独立完成一项调查，并在论文中对整个过程进行评估。很多同学把大量时间花在数据收集上，却忽略了评估部分，导致实验做得不错，最终分数却不理想。这篇文章将系统讲解IA评估部分的写作方法、评分标准和常见错误，帮助你在交稿前把这一部分改到最好。
The Internal Assessment (IA) is one of the areas where IB Diploma students most often lose marks. In every subject, from mathematics and the sciences to economics, students must carry out an independent investigation and then evaluate their own work. Many students spend most of their time on data collection and treat the evaluation section as an afterthought. This article explains how the evaluation section is marked, how to structure it, and which mistakes to avoid before submission.
一、IA评估部分是什么？它在总分中占多少？| What Is the Evaluation Section and How Much Is It Worth?
IA评估是IB科学类科目（物理、化学、生物）评分标准中的一项独立准则，通常占总分的四分之一左右。以IB物理为例，新版大纲中评估准则（Evaluation）占6分，加上个人参与（Personal Engagement）2分、探索（Exploration）6分、分析（Analysis）6分、交流（Communication）4分，共24分。评估部分的6分专门考察学生能否批判性地审视自己的方法、数据和结论，而不是简单地重复实验过程。在数学IA中，对应的是反思（Reflection）准则；在经济学IA中，评估是评论文的最后一个层次。理解每个科目对评估的具体定义，是写好的第一步。
In IB science subjects such as Physics, Chemistry and Biology, Evaluation is an official marking criterion worth about a quarter of the total IA marks. In the current IB Physics syllabus, for example, Evaluation contributes 6 marks out of 24, alongside Personal Engagement (2), Exploration (6), Analysis (6) and Communication (4). This criterion tests whether you can look critically at your own methods, data and conclusions rather than simply repeating what you did. In the Mathematics IA the equivalent is the Reflection criterion, and in Economics it is the final level of the commentary. Knowing exactly what evaluation means in your subject is the first step.
二、评估与结论的区别：总结过去 vs 反思过程 | Evaluation vs Conclusion: Summarising versus Reflecting
结论（Conclusion）回答"我发现了什么"，评估（Evaluation）回答"我为什么相信这个发现，它有多可靠"。很多学生把评估写成了结论的复述，比如"我算出的重力加速度是9.72 m/s²，接近理论值9.81 m/s²"，然后就结束了。这只能算结论。真正的评估要接着分析：误差来源是什么？随机误差和系统误差分别占多少？哪个步骤对结果影响最大？换一种方法结果会不会不同？评估的本质是质疑自己的研究，而结论只是陈述研究结果。在IB评分中，能体现这种批判性反思的文字，才配得上高分档。
The conclusion answers the question "what did I find", while the evaluation answers "why do I trust this finding and how reliable is it". Many students simply restate the conclusion, writing something like "I calculated the acceleration due to gravity as 9.72 m/s squared, close to the accepted value of 9.81", and stop there. That is a conclusion, not an evaluation. A genuine evaluation continues by analysing the sources of error, separating random from systematic effects, identifying which step had the greatest impact on the result, and asking whether a different method would change the outcome. Evaluation is about questioning your own study; a conclusion merely reports it.
三、评估的三大对象：方法、数据与结果 | Three Targets: Method, Data and Results
一篇完整的评估通常覆盖三个层面。第一是方法评估：实验设计是否合理，变量控制是否到位，测量工具是否精确。第二是数据评估：数据是否稳定，有没有异常值，不确定性如何传播。第三是结果评估：最终结果与理论值、文献值的偏差有多大，这种偏差是否在误差范围内。把评估分成这三个层面来写，结构会非常清晰，也方便考官快速找到给分点。每个层面各写一到两段，篇幅配比大约为方法四成、数据三成、结果三成。
A complete evaluation typically covers three levels. The first is the method: whether the design is sound, whether variables were properly controlled, and whether the measuring instruments were precise. The second is the data: how consistent the readings were, whether outliers appeared, and how uncertainties propagated through the calculation. The third is the result: how far the final value deviates from theoretical or literature values, and whether that deviation falls within the estimated uncertainty. Organising the evaluation around these three levels keeps the structure clear and makes it easy for examiners to award marks.
四、方法评估：找出实验设计中的局限 | Evaluating the Method: Finding Weaknesses in the Design
方法评估要具体到步骤，而不是泛泛而谈。不要写"实验有误差"这种废话，而要指出：例如在测量单摆周期时，手按秒表存在反应时间误差，约0.2秒；摆角超过15度时，小角度近似失效，导致周期偏长；绳长从悬点到摆球中心的测量没有考虑球的半径。每个局限都要说明它如何影响结果：是让结果偏大还是偏小？影响程度有多大？能否通过改进消除或减小？能够把局限与结果方向联系起来的评估，在考官眼中远比"存在误差"四个字有价值。
Method evaluation must name specific steps rather than making vague statements. Writing "the experiment had errors" earns nothing. Instead, point to concrete weaknesses: for a pendulum experiment, the reaction time of the stopwatch introduces a random error of about 0.2 seconds; angles beyond 15 degrees break the small-angle approximation and lengthen the period; the string length measured from the clamp to the centre of the bob ignores the radius of the bob. For each limitation, explain how it affects the result, whether it pushes the value up or down, how large the effect is, and whether it can be reduced by a better design. Evaluations that link each limitation to the direction of the error are far more valuable to examiners.
五、数据评估：误差、异常值与不确定性 | Evaluating the Data: Errors, Outliers and Uncertainties
数据评估的核心是量化。先计算重复测量的标准差和平均值的标准误差，判断数据的分散程度是否合理。然后检查是否有异常值：如果某一次读数偏离平均值超过三个标准差，要考虑仪器故障、记录错误或条件突变等原因，并说明是否将其剔除。还要区分随机误差与系统误差：随机误差使数据围绕真实值波动，可以通过多次测量取平均来减小；系统误差使所有数据向同一方向偏移，例如未校准的电子秤每次都偏轻，平均再多也无法消除。最后，用误差传播公式把各步骤的不确定性合并，给出最终结果的不确定度，并判断理论值是否落在置信区间内。
Data evaluation is about quantification. Start by calculating the standard deviation of repeated readings and the standard error of the mean to judge whether the scatter is reasonable. Then check for outliers: if a reading deviates from the mean by more than three standard deviations, consider instrument malfunction, recording errors or changed conditions, and state whether it was excluded and why. Separate random from systematic error: random error scatters readings around the true value and can be reduced by averaging, while systematic error shifts every reading in the same direction, such as an uncalibrated balance that always reads low. Finally, combine uncertainties through error propagation to give the final value an uncertainty, and test whether the accepted value falls inside that range.
六、结果评估：与理论值和文献值对比 | Evaluating the Result: Comparing with Theoretical and Literature Values
结果评估要把自己的数值放回更大的背景下检验。如果实验结果与理论值接近，偏差在不确定度范围内，说明方法整体可靠，但仍要讨论残留的系统误差；如果偏差明显超出不确定度，就要回到方法部分找出最可能的环节，例如忽略了空气阻力、热量散失或摩擦损耗。如果条件允许，把结果与课本文献或已发表研究的数据对比，引用文献要准确，说明来源和条件差异。例如"本实验测得的反应速率常数低于文献值约15%，可能与温度控制偏差2摄氏度有关"。这种对比能让评估从自我检查上升到科学对话的高度。
Result evaluation places your value in a wider context. If the experimental result is close to the theoretical value and the difference falls within the estimated uncertainty, the method is broadly reliable, but residual systematic errors should still be discussed. If the gap clearly exceeds the uncertainty, return to the method section to identify the most likely cause, such as air resistance, heat loss or friction. Where possible, compare your value with published data from textbooks or research papers, citing the source and noting differences in conditions. For example, "the rate constant measured in this experiment is about 15 percent below the literature value, possibly because the temperature drifted by 2 degrees Celsius". This kind of comparison lifts the evaluation from self-checking to genuine scientific dialogue.
七、改进建议：具体、可行、有依据 | Writing Improvements: Specific, Feasible and Justified
改进建议是评估的落点，也是很多学生的失分点。空泛的"下次做得更仔细"没有意义，考官要看的是：改什么、怎么改、为什么有效。一个好的改进建议包含三个要素：针对的局限、具体的操作、预期的效果。例如"使用光电门计时器替代手动秒表，可将时间测量误差从0.2秒降至0.01秒，从而把周期的不确定度从5%降至0.5%"。还可以区分"可以立即实施的改进"和"受条件限制的改进"，例如学校实验室没有精密设备时，可以改用多次重复与均值处理来降低随机误差。改进建议与前面指出的局限一一对应，是最容易拿分也最容易被忽视的写作技巧。
Improvement suggestions are where many students lose marks. A vague "do better next time" carries no credit; examiners want to see what to change, how to change it, and why it would work. A strong improvement has three elements: the limitation it targets, the concrete action, and the expected effect. For example, "using a light gate instead of a hand stopwatch would reduce the timing error from 0.2 to 0.01 seconds, cutting the uncertainty in the period from 5 to 0.5 percent". It also helps to distinguish improvements that can be implemented immediately from those limited by resources, such as relying on more repeats and averaging when precision instruments are unavailable. Matching every improvement to a limitation named earlier is the most reliable way to secure these marks.
八、数学IA的评估：反思你的方法与证明 | Evaluation in the Math IA: Reflecting on Your Approach
数学IA（数学探究）的评估体现在反思准则中，重点不是实验误差，而是对数学方法的反思。你要问自己：这个模型在什么条件下成立？假设是否合理？模型的结果与真实情况有多大偏差，为什么？例如用微分方程模拟人口增长时，没有考虑资源上限的模型在长期会高估人口，引入逻辑斯蒂模型后拟合度显著提高。还要评估方法的普适性：这个结论能否推广到其他情形？计算过程中有没有近似或简化，它们如何影响最终答案的精度？如果使用图形计算器或软件辅助计算，也要说明其局限。数学IA的评估不需要实验数据，但需要同样深刻的批判性思考。
In the Mathematics IA, evaluation appears in the Reflection criterion and focuses on your mathematical approach rather than experimental error. Ask yourself: under what conditions does this model hold, are the assumptions reasonable, and how far does the model deviate from reality? For example, a simple exponential model of population growth ignores resource limits and overestimates the population in the long run, while the logistic model fits the data far better. You should also assess the generality of your method: can the conclusion be extended to other situations? Note any approximations or simplifications and how they affect the precision of the final answer. If you used technology such as a graphing calculator or software, discuss its limitations too. The Math IA needs no experimental data, but it demands the same critical depth.
九、科学IA的评估：从实验细节到结论信度 | Evaluation in Science IAs: From Experimental Detail to Reliability
科学类IA（物理、化学、生物）的评估准则要求把方法、数据、结果三部分串联起来，最终落脚在结论是否可信上。物理实验要关注测量精度、系统误差和误差传播；化学实验要关注反应条件控制、产物纯度、滴定终点判断等细节；生物实验要关注样本量、对照设置、生物变量的天然波动。生物实验常出现的一个问题是样本量太小，例如只测量了三次，导致统计结论不可靠，评估时就要明确提出增大样本量的具体方案。无论哪个学科，考官都希望看到"证据-推理-结论"的链条：先摆出数据证据，再推理误差来源，最后对结论的可信度给出明确判断。
In science IAs, the Evaluation criterion asks you to connect method, data and results into a judgement about whether the conclusion can be trusted. Physics experiments focus on measurement precision, systematic error and error propagation; chemistry on reaction conditions, product purity and end-point detection in titrations; biology on sample size, control groups and natural biological variation. A common weakness in biology is a tiny sample size, such as only three repeats, which makes statistical claims unreliable; the evaluation should propose a concrete plan to increase the sample. In every science, examiners reward an evidence-reasoning-conclusion chain: present the data evidence, reason about the sources of error, and then state clearly how confident you are in the conclusion.
十、评估部分的评分标准与高分特征 | Marking Criteria and What a Top-Band Evaluation Looks Like
了解评分标准才能对号入座地拿分。以IB物理评估准则为例，1至2分档只要求简单指出局限和改进；3至4分档要求评估方法并对改进作出一般性描述；5至6分档（满分档）要求系统地评估所有相关方面，包括方法、数据、结果，改进建议具体且与局限明确对应，并讨论数据的可靠性和结果的合理性。高分特征可以归纳为四点：具体（提到真实数值和具体步骤）、量化（用不确定度说话）、系统（方法、数据、结果全覆盖）、关联（改进与局限一一对应）。写作时对照这四点自查，评估部分的分数会明显提升。
Knowing the mark descriptors lets you target the top band directly. In the IB Physics Evaluation criterion, for example, 1 to 2 marks require only simple limitations and suggestions; 3 to 4 marks require an evaluation of the method with general improvements; 5 to 6 marks, the top band, require a systematic evaluation of all relevant aspects including method, data and results, with specific improvements clearly linked to the limitations identified, plus a discussion of data reliability and the reasonableness of the result. Top-band writing can be summarised in four words: specific, quantified, systematic and linked. Check your draft against these four features before submission.
十一、常见错误与交稿前检查清单 | Common Mistakes and a Pre-Submission Checklist
最后总结评估部分最常见的五个错误：把评估写成结论的复述；只说有误差却不指明来源和方向；改进建议空泛，无法操作；忽略数据层面的评估，只谈方法；评估内容与实验记录不一致，考官一眼就能识破。交稿前可以按清单自查：是否区分了随机误差和系统误差？是否给出不确定度并判断理论值是否落在范围内？是否每个局限都对应一个具体改进？是否提到数据异常值及处理理由？是否说明结论的可信程度？是否使用了准确的专业术语？把这六项全部打勾，你的评估部分就达到了高分档的水准。
Finally, the five most common mistakes: restating the conclusion as the evaluation; saying "there were errors" without identifying sources or directions; offering vague improvements that cannot be acted on; evaluating only the method while ignoring the data; and writing an evaluation that contradicts the lab record, which examiners spot immediately. Before submission, run through this checklist: have you separated random and systematic error? Have you quoted an uncertainty and tested whether the accepted value falls within it? Does every limitation have a matching improvement? Did you explain any outliers and how you handled them? Have you stated how confident you are in the conclusion? Did you use accurate technical terms? If you can tick all six boxes, your evaluation is at top-band standard.
需要一对一辅导？
16621398022
同微信 | 免费咨询课程规划
访问 aleveler.com 了解更多
更多咨询请联系16621398022（同微信）
或截图后用「扫一扫 → 相册」选择本图