📚 Random and Systematic Errors | 随机误差与系统误差
In experimental physics, no measurement is perfectly exact. Every recorded value carries an inherent uncertainty, and understanding the nature of these uncertainties is crucial for drawing valid conclusions. Errors in physics are not 'mistakes' but the unavoidable differences between a measured value and the true value. They are broadly classified into two types: random errors and systematic errors. Distinguishing between them, knowing how to minimise their impact, and being able to quantify uncertainty are skills at the heart of the IB Physics internal assessment and data-analysis questions.
在实验物理中，没有哪个测量是完全精确的。每一个记录值都携带着固有的不确定性，理解这些不确定性的本质对于得出可靠的结论至关重要。物理学中的“误差”不是指犯错，而是测量值与真实值之间无法避免的差值。误差通常分为两类：随机误差和系统误差。学会区分它们、知道如何减小它们的影响并能够量化不确定度，是 IB 物理内部评估和数据分析题目的核心技能。
1. What Are Errors in Physics? | 物理学中的误差是什么？
In the context of physical measurement, an error is the discrepancy between a measured result and the actual (true) value of the quantity being measured. It is important to stress that errors are not blunders, like reading a scale incorrectly or misrecording a digit; those are mistakes that can be avoided by careful technique. True errors arise from limitations of instruments, environmental fluctuations, or imperfect experimental procedures. A measurement is considered useful only when its associated error is estimated and reported alongside the result.
在物理测量语境中，误差是指测量结果与被测量物理量的实际（真实）值之间的差值。必须强调，误差不是失误，比如读错刻度或记错数字；这些是可以通过细心操作避免的过错。真正的误差来源于仪器的限制、环境的波动或不完善的实验步骤。只有当测量的相关误差被估计并与结果一同报告时，该测量才被认为是有用的。
2. The Nature of Random Errors | 随机误差的性质
Random errors cause measured values to scatter unpredictably about the true value. They are just as likely to make a reading too high as too low, and their magnitude and sign vary from one measurement to the next without a consistent pattern. Because random errors obey statistical laws, taking a large number of repeated readings and calculating a mean can significantly reduce their effect. However, even after averaging, a residual uncertainty remains, which can be characterised by statistical tools such as the standard deviation.
随机误差使得测量值围绕真实值不可预测地散布。它们使读数偏高的可能性和偏低的可能同样存在，其大小和正负号在不同测量中变化不定，没有固定模式。由于随机误差遵循统计规律，大量重复测量并计算平均值可以显著削弱它们的影响。但即使经过平均，残余的不确定度仍然存在，这可用标准差等统计工具加以表征。
3. Sources of Random Errors | 随机误差的来源
Random errors originate from unpredictable variations in the measurement process. Typical sources include slight fluctuations in environmental conditions (temperature, air pressure, humidity), small vibrations of the apparatus, the observer's reaction time when using a stopwatch, or electrical noise in a sensor circuit. For example, when measuring the period of a pendulum by hand, the timing error arising from human reflex is random because sometimes you start the clock slightly early and sometimes slightly late. No two trials will be identical.
随机误差来源于测量过程中无法预测的变动。典型的来源包括环境条件（温度、气压、湿度）的微小波动、仪器的轻微振动、使用秒表时观察者的反应时间，或传感器电路中的电噪声。例如，用手动方式测量单摆周期时，由人反射动作引起的计时误差是随机的，因为有时你会稍早按下秒表，有时又会稍晚。没有两次实验会是完全相同的。
4. Reducing Random Errors | 减少随机误差
The primary method for reducing random errors is to increase the number of measurements and compute the arithmetic mean. Additionally, using a higher-resolution instrument can reduce the size of random fluctuations relative to the quantity being measured. For digital sensors, signal averaging or applying a low-pass filter can smooth out high-frequency noise. In IB lab work, you are often advised to take at least five repeats for each data point and to estimate the random uncertainty by using the half-range (max–min)/2 or by calculating the standard deviation of the mean.
减少随机误差的主要方法是增加测量次数并计算算术平均值。此外，使用更高分辨率的仪器可以减小随机波动相对于被测量的大小。对于数字传感器，信号平均或采用低通滤波可以平滑高频噪声。在 IB 实验操作中，通常建议每个数据点至少重复测量五次，并使用半距法 (max–min)/2 或通过计算平均值的标准差来估计随机不确定度。
5. The Nature of Systematic Errors | 系统误差的性质
Unlike random errors, systematic errors consistently shift all measurements in the same direction. They are not reduced by simply taking more readings, because repetition does not remove the bias. A systematic error causes all data points to be either consistently too high or consistently too low. As a result, the mean of the measurements remains offset from the true value. This type of error affects the accuracy of the experiment but does not show up in the spread of repeated readings; therefore a set of data can appear very precise yet be highly inaccurate.
与随机误差不同，系统误差会使所有测量值一致地朝同一个方向偏移。仅靠增加读数次数无法减小系统误差，因为重复并不能消除这种偏差。系统误差导致所有数据点要么一致偏高，要么一致偏低。结果是，测量值的平均值依然偏离真实值。这类误差影响实验的准确度，但不会在重复读数的离散度中表现出来；因此，一组数据可能看起来非常精密，却很不准确。
6. Sources of Systematic Errors | 系统误差的来源
Systematic errors can arise from faulty calibration of instruments, such as a balance that reads 0.2 g high when nothing is on the pan (zero error). A misaligned metre rule, a thermometer with a cracked capillary, or a voltmeter with a non-zero offset are typical examples. They may also stem from the experimental technique itself, for instance a parallax error when reading a liquid meniscus from an angle, or heat loss to the surroundings that is not accounted for in a calorimetry experiment. Often, these errors are constant throughout the experiment, but they can also be proportional to the measured value (scale error).
系统误差可能来源于仪器校准不当，例如空盘时仍显示 0.2 克的天平（零点误差）。一根未对齐的米尺、毛细管破裂的温度计或具有非零偏移的电压表都是典型的例子。系统误差也可能源于实验手法本身，比如从某一角度读取液体弯月面时产生的视差，或者量热实验中未考虑的向环境散失的热量。这些误差通常在整次实验中保持恒定，但也可能与测量值成比例（标度误差）。
7. Identifying Systematic Errors | 识别系统误差
Detecting systematic errors demands a critical comparison with an accepted reference value or the use of an alternative measurement method. In IB Physics, one practical approach is to perform a 'zero check' before each experiment — for example, recording the reading of an ammeter with no current flowing. Plotting a graph can also be revealing: if the best-fit line does not pass through the origin when theory predicts it should, a systematic zero error is likely present. Furthermore, measuring the same quantity with two independent instruments and finding a consistent discrepancy points towards a bias in one of them.
探测系统误差要求与公认的参考值进行批判性比较，或使用另一种测量方法。在 IB 物理中，一个实用的方法是在每次实验前进行“零点检查”——例如，在没有电流流过时记录电流表的读数。绘制图像也很有揭示性：如果最佳拟合线并不经过理论预测应当经过的原点，就很可能存在系统性的零点误差。此外，用两种独立的仪器测量同一个量并发现一致的偏差，往往表明其中一台仪器存在偏倚。
8. Reducing Systematic Errors | 减少系统误差
The correction of systematic errors requires careful experimental design and calibration. Instruments should be calibrated against known standards before use, and zero errors should be subtracted from all readings. Parallax errors can be avoided by reading scales perpendicularly and by using mirrors placed behind the pointer. In calorimetry, surrounding the apparatus with insulation and applying a cooling correction curve can compensate for heat exchange with the environment. Whenever possible, the experimental procedure should be altered so that the bias cancels out — for instance, by interchanging components or by using a differential method that measures the difference between two quantities rather than an absolute value.
修正系统误差需要谨慎的实验设计和校准。仪器在使用前应按已知标准进行校准，零点误差应从所有读数中扣除。通过垂直读取刻度并在指针后方放置镜子，可以避免视差。在量热学中，用绝热材料包裹装置并采用冷却修正曲线，可以补偿与外界的热交换。只要有可能，就应该改变实验步骤以使偏差相互抵消——例如，交换组件，或采用差分方法测量两个量之间的差值而非绝对值。
9. Comparing Random and Systematic Errors | 随机误差与系统误差的比较
The table below summarises the key differences between the two classes of error.
下表总结了两类误差的主要区别。
| Aspect | 方面 | Random Error | 随机误差 | Systematic Error | 系统误差 |
|---|---|---|
| Effect on readings | 对读数的影响 | Scatter about true value | 围绕真实值散布 | All shifted in same direction | 整体向同一方向偏移 |
| Reduced by averaging? | 平均化能减小吗？ | Yes | 是 | No | 否 |
| Affects precision or accuracy? | 影响精密还是准确？ | Precision (spread of data) | 精密度（数据离散度） | Accuracy (closeness to true value) | 准确度（接近真实值） |
| Visible in replicate measurements? | 重复测量中可见吗？ | Yes, as fluctuations | 是，表现为波动 | Not by repeating alone | 单独重复无法发现 |
| Statistical treatment | 统计处理 | Standard deviation, error bars | 标准差，误差棒 | Calibration, zero correction | 校准，零点修正 |
Understanding this comparison allows an experimenter to decide where to invest effort: sometimes improving technique to remove bias is more urgent than carrying out hundreds of repeats that only reduce random scatter.
了解了这种比较后，实验者便可以决定应当将精力投向何处：有时候，消除偏差的技术改进，远比进行几百次只能减少随机散布的重复测量更为紧迫。
10. Accuracy and Precision | 准确度与精密度
Accuracy describes how close a measured value is to the accepted 'true' value, whereas precision refers to how close a set of repeated measurements are to each other. A highly precise experiment may have very small random errors, yet still be inaccurate because of a systematic offset. Ideally, IB Physics students should design investigations that are both accurate and precise. The concepts are often visualised by a target: accurate shots cluster near the bullseye, precise shots cluster tightly together (even if far from the centre), and a good experiment delivers a tight cluster right on the bullseye.
准确度描述的是测量值接近公认“真实”值的程度，而精密度指的是一组重复测量值彼此接近的程度。一个精密度很高的实验可能拥有很小的随机误差，但由于系统偏移仍然不准确。理想情况下，IB 物理学生设计的探究应该是既准确又精密的。这些概念常常用靶子来形象化：准确的射击簇拥在靶心附近，精密的射击紧密地扎在一起（即使远离中心），而一项好的实验会提供一个紧紧聚在靶心上的集群。
11. Error Propagation and Combining Uncertainties | 误差传递与不确定度的合成
ultiplied or divided, relative (percentage) uncertainties are added. The IB Physics data booklet provides the necessary formulas. A typical calculation for a derived quantity Q = a + b has an absolute uncertainty ΔQ = √Δa² + Δb² when uncertainties are independent. For Q = k × aⁿ, the relative uncertainty ΔQ/Q = |n| lying these rules correctly tells the reader how confident you can be in the final result.
当最终结果是由几个各有不确定度的测量量计算得到时，这些不确定度必须按传递规则合成。对于相加或相减的量，绝对不确定度以平方和根式相加，或简单按最差情况的线性求和。对于相乘或相除的量，相对（百分比）不确定度相加。IB 物理数据手册给出了必要的公式。一个典型的计算是，对于导出量 Q = a + b，当不确定度相互独立时，其绝对不确定度 ΔQ = √Δa² + Δb²。对于 Q = k × aⁿ，相对不确定度 ΔQ/Q = |n| × Δa/a。正确地应用这些规则可以告诉读者你对最终结果可以有多大的信心。
the volume V = (4/3)πr³ has a relative uncertainty ΔV/V = 3 × (0.05/5.20) ≈ 0.0288 or 2.9% ow a modest measurement uncertainty can be magnified when a quantity is raised to a power.
例如，如果测得球体半径 r = (5.20 ± 0.05) cm，体积 V = (4/3)πr³ 的相对不确定度 ΔV/V = 3 × (0.05/5.20) ≈ 0.0288 即 2.9%。因此绝对不确定度 ΔV = V × 0.0288。这说明，当量被乘方时，一个适度的测量不确定度可能被放大。
12. Practical Tips for IB Physics Investigations | IB 物理探究的实用建议
To excel in the internal assessment, always distinguish between random and systematic errors in your methodology. State clearly how you minimised each type: for random uncertainty, specify the number of trials and the statistical method used (e.g. 'the half-range was taken' or 'the standard deviation of the mean was calculated'). For systematic errors, discuss calibration steps (e.g. 'the force sensor was zeroed before each data-collection run') and justify why you believe no hidden bias remains. When plotting graphs, include both vertical and horizontal error bars where appropriate, and draw a best-fit line together with maximum and minimum slope lines to estimate the uncertainty in the gradient.
要在内部评估中表现出色，务必在你的方法中区分随机误差和系统误差。清楚说明你是如何最小化每一类误差的：对于随机不确定度，指明重复次数和使用的统计方法（如“采用半距法”或“计算了平均值的标准差”）。对于系统误差，讨论校准步骤（如“每次数据采集前力传感器都回零”），并证明为何确信没有隐藏偏差。绘制图表时，在适当时同时添加竖直和水平误差棒，并画出最佳拟合线以及最大、最小斜率线，以估计梯度中的不确定度。
ures, and the measured value rounded to the same decimal place. Use clear language such as ' ± 0.06) V, which agrees with the accepted value of 4.80 V within experimental uncertainty'. This demonstrates a mature understanding of how errors frame a scientific conclusion.
再者，始终以一致的格式呈现不确定度：绝对不确定度保留一到两位有效数字，测量值则四舍五入到相同小数位。使用清晰的语言，比如“测量结果为 (4.82 ± 0.06) V，在实验不确定度范围内与公认值 4.80 V 相符”。这展示了你对于误差如何框定科学结论的成熟理解。
Published by TutorHao | Physics Revision Series | aleveler.com
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