📚 Validity in A-Level Physics Experiments | A-Level 物理实验中的有效性
In AQA A-level Physics, the quality of an experiment is not judged only by how many decimal places you record. Examiners use specific terms such as validity, reliability, accuracy and precision, and each term means something different. Validity asks a fundamental question: does the experiment actually measure what it claims to measure, and can the conclusion be trusted for the stated aim?
在 A-Level 物理（AQA 等考试局）中，评价一个实验的质量不能只看你记录了多少位小数。考官会使用效度、信度、准确度和精密度等特定术语，每个术语的含义都不同。有效性提出一个根本问题：实验是否真正测量了它声称要测量的量，以及结论是否能够支持所陈述的目标？
1. What Is Validity? | 什么是有效性？
A valid physics experiment is one in which the method, instruments and conditions allow the dependent variable to change only because of changes in the independent variable. If an unmeasured variable also influences the results, the data may still look precise, but the conclusion is not valid.
一个有效的物理实验，其方法、仪器和条件使得因变量只会因为自变量的改变而改变。如果某个未测量的变量也影响了结果，那么数据可能看起来仍然精确，但结论却无效。
Validity is therefore about experimental design and control, not about taking more readings. It is a property of the whole investigation: the aim, the physical model you assume, the range you choose, the variables you hold constant, and the way you interpret the graph. A beautifully presented table of numbers cannot rescue an experiment that answers the wrong question.
因此，有效性关注的是实验设计与控制，而不是增加读数次数。它是整个探究的属性：实验目标、你假设的物理模型、选择的量程范围、保持恒定的变量，以及你解读图像的方式。一张排版精美的数据表格，也无法挽救一个"回答了错误问题"的实验。
A useful test of validity is to read your own conclusion and ask: could this result have been produced even if the physics I claim were false? If the answer is yes - because friction, heating, air resistance or a mis-set zero was doing the work - then the conclusion is not valid, no matter how small the uncertainties are.
检验有效性有一个实用方法：读一读你自己的结论，然后问：即使我声称的物理规律不成立，这个结果是否仍然可能出现？如果答案是"会"——因为摩擦、升温、空气阻力或零点未校准在起作用——那么这个结论就是无效的，无论不确定度多小。
2. Validity Compared with Reliability, Accuracy and Precision | 有效性与可靠性、准确性、精密度的比较
Students often confuse validity with reliability and accuracy. The four terms answer four different questions, and a good answer in an exam separates them clearly.
学生常常把有效性与可靠性、准确性混为一谈。这四个术语回答的是四个不同的问题，考试中高分答案会把它们清楚地区分开。
- Validity: does the experiment measure the quantity named in the aim, with only the independent variable affecting the dependent variable?
有效性：实验是否测量了目标中所述的那个量，且只有自变量影响因变量？ - Reliability: if the experiment is repeated, do you obtain consistent results? Reliability is improved by repeating readings, taking repeats and averaging, and by identifying and removing anomalies.
可靠性：重复实验时能否得到一致的结果？通过重复读数、取多次测量并求平均值、找出并剔除异常值来提高可靠性。 - Accuracy: how close a measured value is to the true value. Accuracy is affected mainly by systematic errors such as zero errors and mis-calibration.
准确性：测量值与真值的接近程度。准确性主要受零点误差、未校准等系统误差影响。 - Precision: how close repeated readings are to each other, i.e. the spread of values. Precision is affected by random errors and by the resolution of the instrument.
精密度：多次重复读数彼此之间的接近程度，即数据的离散程度。精密度受随机误差与仪器分辨率影响。
| Term | Question it answers | Improved by |
| Validity 有效性 | Am I measuring the right quantity? 我测的是不是我要测的量？ | Better design, control of variables 更好的设计、变量的控制 |
| Reliability 可靠性 | Are my results repeatable? 结果可否重复？ | Repeats, averaging 重复测量、取平均 |
| Accuracy 准确性 | Am I close to the true value? 是否接近真值？ | Calibration, correcting zero error 校准、修正零点误差 |
| Precision 精密度 | Do my repeats agree? 重复值是否一致？ | Better resolution, careful technique 更高分辨率、细致操作 |
An experiment can be precise but not valid, and it can be valid but not very precise. For example, timing a falling ball with a stopwatch may give very consistent times, yet the method may be invalid for measuring g if the ball is released by hand. Repeating that experiment a hundred times only makes it more reliable - it never makes it valid.
一个实验可以很精密但无效，也可以有效但不太精密。例如用手按秒表给下落的小球计时，时间可能非常一致，但如果小球是"用手释放"的，那么这种方法用于测量 g 就是无效的。把这个实验重复一百次只会提高可靠性，却永远无法使其变得有效。
3. Validity and the Practical Endorsement (CPAC) | 有效性与实验技能认证（CPAC）
In AQA A-level Physics, practical skills are assessed in two ways: through a separate practical endorsement (reported as a pass or not classified) and through written papers, where practical skills and data analysis carry a significant share of the marks. The endorsement is judged against the Common Practical Assessment Criteria (CPAC), and validity sits at the centre of several of them.
在 AQA A-Level 物理中，实验技能通过两种方式评估：一是单独的实验技能认证（记作"通过"或"未分类"），二是笔试，其中实验技能与数据分析占有相当比例的分数。认证依据《通用实验评估标准》（CPAC）进行判定，而有效性正处在其中多条标准的中心位置。
- CPAC 2 - applying investigative approaches and methods: choosing the correct apparatus and using it correctly so that the measured variable is genuinely the variable in the aim.
CPAC 2——运用探究方法与手段：选择正确的仪器并正确使用，使被测变量确实是目标中的那个变量。 - CPAC 4 - making and recording observations: recording enough data, of appropriate resolution, with the correct number of significant figures and correct units.
CPAC 4——做出并记录观察：记录足够的数据，具有适当的分辨率、正确的有效数字与单位。 - CPAC 5 - researching, referencing and reporting: justifying the method and evaluating whether the conclusion is supported by the data.
CPAC 5——研究、引用与报告：论证方法的合理性，并评估结论是否得到数据的支持。
In the written papers, the practical-based questions almost always begin with a real experimental scenario, and the first marks are usually for describing the method, while the later marks are for evaluating whether the conclusion is valid. Typical wording is: "Suggest why the student's conclusion may not be valid" or "Explain how the experiment could be improved to make the conclusion more valid". Notice that the word is valid, not accurate.
在笔试中，实验类问题几乎总是从一个真实的实验情境出发：前面的分数通常考查对方法的描述，后面的分数则考查对"结论是否有效"的评价。典型的问法是："请说明为什么该学生的结论可能无效"或"请解释如何改进实验以使结论更有效"。请注意，这里用的词是"有效"，而不是"准确"。
4. Sources of Invalidity: Uncontrolled Variables | 无效性的来源：未控制的变量
The most common source of invalidity is a variable that changes during the experiment but is not the independent variable and is not measured. Such a variable is sometimes called a confounding variable, because it is confounded (mixed up) with the independent variable.
最常见的无效性来源是：某个变量在实验过程中发生变化，但它既不是自变量，也没有被测量。这种变量有时被称为混淆变量，因为它与自变量纠缠在一起、无法分离。
- Temperature: a current-carrying wire warms up, so resistance changes for a reason unrelated to the length being varied.
温度：通电导线发热，电阻发生变化，而这个变化与被改变的长度无关。 - Friction and air resistance: energy is dissipated, so the model of conservation of mechanical energy no longer holds.
摩擦与空气阻力：能量被耗散，机械能守恒模型不再成立。 - Human reaction time: a stopwatch is started and stopped by a person, so the measured time depends on reaction time rather than only on the physics.
人的反应时间：秒表由人启动和停止，所测时间取决于反应时间，而不仅仅取决于物理过程。 - tic or plastic behaviour: a spring deformed beyond its elastic limit does not obey F = kx.
弹性或塑性行为：弹簧超过弹性限度后不再遵守 F = kx。 - Parallax and alignment: reading a scale from an angle systematically shifts the reading, and a ruler not placed along the line of the quantity being measured measures the wrong length.
视差与对准：从斜角读数会系统性偏移，而直尺未沿被测长度放置则测到的是错误的长度。
A second, subtler source is a model that is applied outside its range of validity. A simple pendulum is only approximately simple harmonic for small angles, and the relationship between force and extension is only linear up to the limit of proportionality. Using the linear model beyond that range makes the physics of the interpretation wrong, even if every reading is perfect.
第二个更隐蔽的来源是把模型用在了其适用范围之外。单摆只有在小角度下才近似做简谐运动，力与伸长量之间也只有到弹性限度为止才呈线性关系。一旦超出这个范围仍使用线性模型，那么即便每一个读数都完美，解读所用的物理也是错的。
5. Instrument Choice, Range and Resolution | 仪器的选择、量程与分辨率
Validity is partly a question of instrument choice. An instrument must be able to measure the quantity you claim to be measuring, across the whole range of the independent variable, with a resolution that makes the expected change detectable.
有效性在一定程度上是仪器选择的问题。仪器必须能够测量你声称在测的量，覆盖自变量的整个取值范围，并且其分辨率足以让预期变化可被检测出来。
Percentage uncertainty is a good way to argue about this in an exam:
百分不确定度是考试中论证这一点的好办法：
percentage uncertainty = (uncertainty ÷ measured value) × 100%
mm is used to measure a 100 cm length, the percentage uncertainty is (0.001 ÷ 1.00) × 100% = 0.1% he same ruler is used to measure a 2 mm change in length, the percentage uncertainty is 50% ling microscope or a vernier scale reduces the uncertainty and strengthens the conclusion.
如果用分度值为 1 mm 的 30 cm 直尺测量 100 cm 的长度，百分不确定度为 (0.001 ÷ 1.00) × 100% = 0.1%。若用同一把尺测量 2 mm 的长度变化，百分不确定度则为 50%，大到实验无法支撑任何精确结论。改用移动显微镜或游标卡尺可降低不确定度，从而增强结论的说服力。
Note carefully: a better instrument improves precision and may improve accuracy, but it does not automatically make the experiment valid. Choosing a micrometer to measure the diameter of a wire does not help if the wire is being heated by the current, because the physical condition under which resistance is measured is still wrong.
请特别注意：更好的仪器能提高精密度，也可能提高准确性，但并不会自动使实验变得有效。即便你换用千分尺测量导线直径，只要导线仍被电流加热，实验就依然无效，因为测量电阻时的物理条件本身是错的。
6. The Hooke's Law Example: Where Validity Breaks Down | 胡克定律示例：有效性在哪里失效
Consider the classic experiment: hang masses on a spring and measure the extension. The aim is to test whether the extension is proportional to the force, and then to determine the spring constant k. The relationship claimed is:
考虑经典实验：在弹簧下挂砝码并测量伸长量。目标是检验伸长量是否与力成正比，进而测定弹簧劲度系数 k。所声称的关系为：
F = k x
Here F is the force in newtons, x is the extension in metres, and k is the spring constant in N m⁻¹. The graph of F against x is a straight line through the origin with gradient k - but only while the spring obeys Hooke's law.
这里 F 是力（单位 N），x 是伸长量（单位 m），k 是劲度系数（单位 N m⁻¹）。以 F 对 x 作图得到一条过原点的直线，斜率即为 k——但仅在弹簧遵守胡克定律时成立。
If you investigate the extension of a spring against force but allow the spring to exceed its limit of proportionality, the extension is no longer measuring the simple linear relationship stated in Hooke's law, so the conclusion becomes invalid. A graph that curves at the top end is evidence that the spring has been permanently deformed. Any value of k calculated from the curved region is meaningless, because the physical law being tested no longer applies there.
如果你研究弹簧的伸长量与力的关系，但让弹簧超过了弹性限度，伸长量就不再反映胡克定律所述的简单线性关系，因此结论变得无效。图像末端出现弯折，说明弹簧已发生永久形变。从弯曲区域算出的任何 k 值都是没有意义的，因为被检验的物理规律在该区域已不再适用。
Other validity problems in this experiment include: using rusty or bent springs, not waiting for the spring to settle before reading, reading the scale from an angle (parallax), and forgetting to record the mass of the hanger itself. Each of these can shift the data, but the key point is the same - the experiment no longer measures the quantity named in the aim.
该实验的其他有效性问题包括：使用生锈或弯曲的弹簧、读数前未等弹簧稳定、从斜角读数（视差），以及忘记记录挂钩本身的质量。这些问题都会使数据发生偏移，但关键点是一致的——实验已不再测量目标中所声称的那个量。
7. Systematic Error, Random Error and Validity | 系统误差、随机误差与有效性
Random errors scatter readings around a mean value. They reduce precision but do not, on their own, destroy validity, because averaging repeated readings reduces their effect. Systematic errors, by contrast, shift every reading in the same direction, and their effect is not removed by repeating the experiment.
随机误差使读数围绕平均值散布。它们降低精密度，但本身并不破坏有效性，因为重复测量取平均可以减弱其影响。相反，系统误差使每一次读数都朝同一方向偏移，其影响不会因重复实验而消除。
A classic systematic error is a zero error, where the instrument does not read zero when it should. A stopwatch started slightly late, an ammeter reading 0.02 A when no current flows, or a micrometer with a nonzero zero reading all produce this effect. A zero error in a measured quantity makes the absolute value inaccurate.
典型的系统误差是零点误差，即仪器在应当读零时并未读零。秒表启动稍晚、电流为零时电流计却显示 0.02 A、千分尺零点不为零，都会造成这种效应。被测物理量存在零点误差，会使其绝对值不准确。
However, there is an important nuance that earns high marks. If a systematic offset affects only the dependent variable by a constant amount, the shape of the graph may be unchanged. A constant zero error in the extension of a spring shifts the line up or down but does not change its gradient, so the value of k can still be valid even though each individual extension reading is inaccurate. Conversely, if the offset depends on the reading, or if the independent variable is affected, the gradient itself is distorted and the conclusion is no longer valid. Validity depends on the structure of the error, not merely on its presence.
不过有一个能拿高分的重要细节：如果系统偏差只是以恒定数值影响因变量，图像的形状可能不变。弹簧伸长量上的恒定零点误差会使直线上下平移，却不改变斜率，因此即便每个伸长量读数都不准确，k 的数值仍可能是有效的。反之，如果偏差随读数变化，或者自变量也受到影响，斜率本身就被扭曲，结论便不再有效。有效性取决于误差的结构，而不仅仅是误差是否存在。
8. Improving Validity: Design, Control and Fair-Test Logic | 提高有效性：设计、控制与公平实验逻辑
Improving validity means changing what is measured or how it is measured, not simply collecting more data. In an exam answer, the strongest structure is: identify the uncontrolled variable, explain how it affects the dependent variable, state how it should be controlled, and explain how this changes the conclusion.
提高有效性意味着改变"测什么"或"怎么测"，而不仅仅是收集更多数据。在考试作答中，最强的结构是：指出未控制的变量，解释它如何影响因变量，说明应如何控制它，并解释这种改动如何改变结论。
- Control the variables: keep temperature constant by using a small current for a short time; use a vacuum tube or a light gate for free-fall measurements; use a fiducial marker or a clamp to ensure repeatable release.
控制变量：用短时间小电流保持温度恒定；用真空管或光电门测量自由落体；用基准标记或夹具保证释放条件可重复。 - angles below about 10° so that sin θ ≈ θ and the small-angle approximation remains valid.
选择正确的范围：只在模型成立的范围内取数据。对单摆，角度应小于约 10°，这样 sin θ ≈ θ，小角度近似仍然成立。 - Measure the correct quantity: measure the extension from the loaded position to the reference position, and subtract the zero-error reading.
测量正确的量：从加载位置到参考位置测量伸长量，并扣除零点误差读数。 - Use a direct method where possible: measure a single period over many oscillations, or use a light gate, rather than relying on hand timing.
尽可能采用直接方法：测量多次全振动所对应的总时间再求单个周期，或使用光电门，而不是依赖手动计时。 - Justify the conclusion: state the physical relationship being tested and check whether the graph supports it, not whether the numbers look tidy.
为结论广告 / Sponsored牛津国际 AQA 考试局Oxford International AS & A Level Physics 物理教材…物理 备考教辅 · 正版现货 · 与本文内容相关查看详情 / View on Taobao参考价 ¥598.0更多咨询请联系16621398022（同微信）
备考资讯每日更新，关注不错过考纲变化 · 科目分析 · 学习方法
在微信里长按二维码识别关注
或截图后用「扫一扫 → 相册」选择本图公众号：tutorhao