📚 Uncertainties and Errors | 不确定度与误差
In experimental physics, every measurement carries an inherent uncertainty. No instrument is infinitely precise, and no observation can be perfectly isolated from external influences. Understanding uncertainties and errors is not about eliminating them entirely, but about quantifying and minimizing their effects so that conclusions drawn from data are reliable and meaningful. This article explores the key concepts of random and systematic errors, accuracy and precision, the mathematical treatment of uncertainty, and how to propagate uncertainties through calculations. These skills are foundational for IB Physics internal assessments and examinations, where you must evaluate experimental methodology and justify conclusions using uncertainty analysis.
在实验物理中，每一次测量都带有固有的不确定度。没有一种仪器能无限精确，也没有任何观测能够完全隔绝外界影响。理解不确定度与误差，并非要将其完全消除，而是要对其进行量化并最小化其影响，从而使由数据得出的结论既可靠又有意义。本文探讨随机误差与系统误差、准确度与精密度、不确定度的数学处理，以及如何在运算中传播不确定度等核心概念。这些技能是 IB 物理内部评估与考试的基础，你需要运用不确定度分析来评价实验方法并对结论加以论证。
1. The Nature of Measurement and Imperfections | 测量与不完美的本质
Every measurement is an approximation of the true value. The difference between a measured value and the true value is called the error. Importantly, an ‘error’ in physics does not mean a mistake; it refers to the unavoidable discrepancy arising from limitations in the measuring process. Two fundamental categories of error exist: systematic and random. Recognizing which type dominates an experiment is the first step toward improving its reliability.
每一次测量都是对真实值的一种近似。测量值与真实值之间的差异称为误差。重要的是，物理学中的“误差”并非指错误，而是指因测量过程的局限而产生的不可避免的偏差。误差分为两个基本类别：系统误差和随机误差。识别实验中哪一类误差占主导，是提高其可靠性的第一步。
2. Systematic Errors: Consistent Bias | 系统误差：持续性偏差
A systematic error causes all measurements to be shifted in the same direction by the same amount. It arises from flaws in the equipment, flawed experimental design, or personal error by the observer (such as parallax if not corrected). For example, a voltmeter that reads 0.2 V even when short-circuited introduces a zero error—all subsequent readings will be 0.2 V too high, unless subtracted. Systematic errors affect the accuracy of an experiment, but they do not affect the precision; the data points may cluster tightly, but around the wrong value. Unlike random errors, repeating measurements does not reduce a systematic error. The only remedy is to identify and eliminate its source: calibrate instruments, use a different technique, or apply a correction.
系统误差使所有测量值都向同一方向偏离相同的量。它源自仪器缺陷、实验设计不当，或观测者的个人误差（例如未纠正的视差）。例如，一台即使在短路时也显示 0.2 V 的电压表就存在零点误差——除非将其减去，否则之后的所有读数都会偏高 0.2 V。系统误差影响实验的准确度，但不影响精密度；数据点可能紧凑聚集，但却围绕着错误的值。与随机误差不同，重复测量并不能减小系统误差。唯一的解决办法是找出并消除其来源：校准仪器、采用不同的方法，或施加修正。
3. Random Errors: Fluctuations Around the Mean | 随机误差：平均值周围的波动
Random errors cause readings to scatter unpredictably about the true value. They arise from unpredictable variations such as fluctuations in environmental conditions, human judgement in reading analogue scales, or electrical noise. For instance, when timing a pendulum’s swing, reaction time may cause the measured period to be sometimes slightly too long, sometimes too short. Random errors reduce precision—the spread of measurements is wider. Crucially, the mean of many repeated measurements will approach the true value if no systematic error is present. Random errors can be reduced by taking multiple readings and averaging, or by using instruments with higher resolution.
随机误差导致读数在真实值周围不可预测地散布。它们源自不可预测的变化，例如环境条件的波动、读取模拟刻度时的人为判断，或电噪声。例如，在对摆的摆动计时时，反应时间可能导致测得的周期时而偏长，时而偏短。随机误差降低精密度——测量值的散布范围更广。关键在于，如果没有系统误差，多次重复测量的平均值将趋近于真实值。随机误差可以通过多次读数取平均，或使用分辨率更高的仪器来减小。
4. Accuracy versus Precision | 准确度与精密度
Accuracy refers to how close a measured value is to the accepted or true value. It is determined by systematic errors. Precision refers to how close repeated measurements are to each other, regardless of their proximity to the true value. It is determined by random errors and the instrument’s resolution. A useful analogy is a target: high accuracy means arrows clustered around the bullseye; high precision means arrows clustered tightly together, but possibly far from the bullseye if a systematic error exists. In lab reports, you must assess both: report the mean value (to indicate accuracy if the true value is known) and the standard deviation or half-range (to indicate precision).
准确度指测量值与公认值或真实值的接近程度，它由系统误差决定。精密度指重复测量值之间的接近程度，无论它们是否靠近真实值，它由随机误差和仪器分辨率决定。一个有用的类比是箭靶：高准确度意味着箭头密集围绕靶心；高精密度意味着箭头彼此紧密聚集，但如果存在系统误差，它们可能远离靶心。在实验报告中，你需要同时评估两者：报告平均值（若已知真实值则可表明准确度），以及标准差或半距（以表明精密度）。
5. Defining Uncertainty: Absolute, Fractional, and Percentage | 不确定度定义：绝对、相对和百分比
Uncertainty quantifies the range within which the true value is expected to lie. The absolute uncertainty (Δx) has the same units as the measurement. For a single reading on a digital instrument, Δx is typically ± the smallest scale division or the manufacturer’s stated accuracy. For an analogue scale, it is ± half the smallest division. The fractional uncertainty is Δx / x (no units) and the percentage uncertainty is (Δx / x) × 100%. These relative measures are essential when comparing the quality of different measurements or when propagating errors through multiplication and division.
不确定度量化了真实值预期所在的范围。绝对不确定度（Δx）与测量值的单位相同。对于数字仪器上的单次读数，Δx 通常为 ± 最小刻度或制造商注明的精度。对于模拟刻度，则为 ± 最小分度的一半。相对不确定度为 Δx / x（无单位），百分比不确定度为 (Δx / x) × 100%。这些相对度量在比较不同测量的质量，或在乘除法运算中传播误差时至关重要。
6. Reading Uncertainty and Instrument Limitations | 读数不确定度与仪器极限
Reading uncertainty arises from the observer’s interaction with the scale. For a ruler marked in millimetres, the reading uncertainty is typically ±0.5 mm. For a digital stopwatch displaying 0.01 s, the manufacturer’s accuracy might be ±0.1 s due to human reaction time, which dominates over the display resolution. Always consider the context: the uncertainty in a measurement of room temperature with a thermometer that has 1 °C divisions might be ±0.5 °C, but if the thermometer is not allowed to equilibrate, the resulting error can be far larger. When designing experiments, choose instruments so that the reading uncertainty is smaller than the expected random variations.
读数不确定度来自观测者与刻度尺的交互。对于毫米刻度的尺子，读数不确定度通常为 ±0.5 mm。对于显示 0.01 s 的数字秒表，由于人的反应时间，其精确度可能为 ±0.1 s，这远大于显示分辨率。始终要考虑具体情境：使用分度为 1 °C 的温度计测量室温，不确定度可能为 ±0.5 °C，但如果温度计未达到热平衡，产生的误差就会大得多。设计实验时，应选择使读数不确定度小于预期随机波动的仪器。
7. Combining Uncertainties: Addition and Subtraction | 不确定度合成：加法和减法
When two or more measurements are added or subtracted, absolute uncertainties add. If you measure the length of a table as L1 = (2.00 ± 0.01) m and an extension piece as L2 = (0.50 ± 0.01) m, the total length L = L1 + L2 = 2.50 m, with an absolute uncertainty ΔL = ΔL1 + ΔL2 = 0.02 m. Thus, L = (2.50 ± 0.02) m. The same rule applies for subtraction: if the table’s original length is 2.00 ± 0.01 m and you cut off a piece measured as 0.50 ± 0.01 m, the remaining length is 1.50 ± 0.02 m. This additive rule is simple but crucial; never subtract absolute uncertainties.
当两个或多个测量值相加或相减时，绝对不确定度相加。若测得桌子长度为 L1 = (2.00 ± 0.01) m，延伸件为 L2 = (0.50 ± 0.01) m，则总长 L = L1 + L2 = 2.50 m，绝对不确定度 ΔL = ΔL1 + ΔL2 = 0.02 m。因此 L = (2.50 ± 0.02) m。减法规则相同：若桌子原长 2.00 ± 0.01 m，切去测量为 0.50 ± 0.01 m 的一段，剩余长度为 1.50 ± 0.02 m。这一相加规则既简单又关键，切勿将绝对不确定度相减。
8. Combining Uncertainties: Multiplication, Division, and Powers | 不确定度合成：乘法、除法和幂
For multiplication and division, you add fractional or percentage uncertainties. If two quantities A ± ΔA and B ± ΔB are multiplied to obtain C = A × B, the fractional uncertainty in C is ΔC/C = ΔA/A + ΔB/B. The same holds for C = A / B. For example, to find the area of a rectangle with width w = (5.0 ± 0.2) cm and length l = (10.0 ± 0.3) cm, the area A = w × l = 50 cm². The fractional uncertainties are 0.2/5.0 = 0.04 and 0.3/10.0 = 0.03, so total fractional uncertainty = 0.07, giving ΔA = 0.07 × 50 = 3.5 cm² ≈ 4 cm² (to 1 sig. fig.); thus A = (50 ± 4) cm². For a quantity raised to a power, such as volume V = k r³, the rule is: multiply the fractional uncertainty in r by the power. If r = (2.0 ± 0.1) cm, fractional uncertainty in r is 0.05; for r³, fractional uncertainty in V becomes 3 × 0.05 = 0.15. These rules are derived from calculus but are straightforward to apply.
对于乘法和除法，需将相对或百分比不确定度相加。若两个量 A ± ΔA 和 B ± ΔB 相乘得到 C = A × B，则 C 的相对不确定度为 ΔC/C = ΔA/A + ΔB/B。对 C = A / B 同样适用。例如，求矩形面积，宽 w = (5.0 ± 0.2) cm，长 l = (10.0 ± 0.3) cm，面积 A = w × l = 50 cm²。相对不确定度分别为 0.2/5.0 = 0.04 和 0.3/10.0 = 0.03，总相对不确定度 = 0.07，所以 ΔA = 0.07 × 50 = 3.5 cm² ≈ 4 cm²（保留 1 位有效数字）；故 A = (50 ± 4) cm²。对于求幂的量，如体积 V = k r³，规则为：将 r 的相对不确定度乘以幂次。若 r = (2.0 ± 0.1) cm，r 的相对不确定度为 0.05；则 r³ 的相对不确定度为 3 × 0.05 = 0.15。这些规则源于微积分，但应用十分直接。
9. Propagating Uncertainty in Complex Equations | 复杂方程中的不确定度传播
When an equation involves multiple operations, break it down step by step using the combination rules. For example, consider the density ρ = m / V, with m = (50.0 ± 0.5) g and V = (20.0 ± 1.0) cm³. First treat the division: fractional uncertainties are 0.5/50.0 = 0.01 for mass, and 1.0/20.0 = 0.05 for volume. Sum them to get fractional uncertainty in ρ = 0.06. The calculated ρ = 50.0/20.0 = 2.50 g cm⁻³, so absolute Δρ = 0.06 × 2.50 = 0.15 g cm⁻³. Hence ρ = (2.50 ± 0.15) g cm⁻³. Always compute absolute uncertainties at the final stage, and round to 1 or 2 significant figures. Note that constants (e.g., π, g) are usually taken as having negligible uncertainty, unless you are investigating their value.
当方程涉及多种运算时，可使用合成规则逐步分解。例如，考虑密度 ρ = m / V，其中 m = (50.0 ± 0.5) g，V = (20.0 ± 1.0) cm³。首先处理除法：质量的相对不确定度为 0.5/50.0 = 0.01，体积为 1.0/20.0 = 0.05。两者相加得 ρ 的相对不确定度为 0.06。计算得 ρ = 50.0/20.0 = 2.50 g cm⁻³，故绝对 Δρ = 0.06 × 2.50 = 0.15 g cm⁻³。因此 ρ = (2.50 ± 0.15) g cm⁻³。应在最后一步计算绝对不确定度，并四舍五入至 1 或 2 位有效数字。注意常数（如 π、g）通常视为不确定度可忽略，除非你正在研究它们的值。
10. Repeated Measurements and Standard Deviation | 重复测量与标准差
When several independent readings of the same quantity are taken, the best estimate of the true value is the arithmetic mean, x̄. The spread is quantified by the standard deviation of the sample (s). For N measurements, s = √[ Σ(xi – x̄)² / (N – 1) ]. In IB Physics, the absolute uncertainty in the mean is often taken as the standard error, s / √N, or more simply as half the range (max – min)/2. The standard error decreases with the square root of the number of measurements, so taking 100 readings instead of 10 reduces random uncertainty by a factor of √10 ≈ 3.16. Always present final results as x̄ ± uncertainty, state the confidence level (usually 68%), and show calculations clearly.
当对同一量进行多次独立读数时，真实值的最佳估计值为算术平均值 x̄。散布程度由样本标准差 (s) 来量化。对于 N 次测量，s = √[ Σ(xi – x̄)² / (N – 1) ]。在 IB 物理中，平均值的绝对不确定度通常取为标准误差 s / √N，或更简单地取半距 (max – min)/2。标准误差随测量次数的平方根而减小，因此进行 100 次而非 10 次读数可将随机不确定度减小 √10 ≈ 3.16 倍。呈现最终结果时应始终写成 x̄ ± 不确定度，注明置信度（通常为 68%），并清晰展示计算过程。
11. Graphical Analysis: Error Bars and Best-Fit Lines | 图形分析：误差棒与最佳拟合线
Plotting data on a graph allows visual assessment of trends and uncertainties. For each data point, draw vertical and/or horizontal error bars representing the absolute uncertainty in that variable. Then fit the best straight line that passes through as many error bars as possible, not necessarily through the origin. Additionally, draw maximum and minimum gradient lines (lines of worst fit) that still pass through all error bars. The uncertainty in the gradient is given by (gradient of best-fit line – gradient of worst-fit line) or the half-range of the two extreme gradients. Similarly, the intercept uncertainty is determined from extreme intercepts. This graphical method yields reliable uncertainties for derived quantities such as acceleration or resistivity.
将数据绘制成图可直观地评估趋势和不确定度。为每个数据点绘制垂直和/或水平误差棒，以表示该变量的绝对不确定度。然后拟合最佳直线，尽可能使其穿过尽可能多的误差棒，不一定过原点。此外，绘制仍穿过所有误差棒的最大和最小梯度线（最差拟合线）。梯度的不确定度由（最佳拟合线梯度 – 最差拟合线梯度）或两个极端梯度的半距给出。同样，截距不确定度由极端截距确定。这种图形方法可为加速度或电阻率等导出量提供可靠的不确定度。
12. Practical Strategies to Minimize Uncertainties | 减小不确定度的实用策略
Good experimental design starts with anticipating sources of uncertainty. Use instruments with finer resolution, take multiple readings, and control environmental variables (temperature, draughts, vibration). Measure larger quantities: timing 20 swings of a pendulum rather than 1 reduces the fractional uncertainty in the period. Align readings to avoid parallax, and check for zero errors before and after measurements. When measuring a thin wire’s diameter, a micrometer gives far better precision than a ruler. Finally, always critically evaluate your uncertainty estimates: are they realistic? Have you accounted for reaction time, the spread of data, and instrument precision? This reflective process is at the heart of the IB Physics investigation and is essential for achieving high marks in the personal engagement and evaluation criteria.
好的实验设计从预见不确定度来源开始。使用分辨率更高的仪器，进行多次读数，并控制环境变量（温度、气流、振动）。测量较大规模：对摆的 20 次摆动计时而不是 1 次，可减小周期的相对不确定度。对齐读数以避免视差，并在测量前后检查零点误差。测量细导线直径时，千分尺远比尺子精确。最后，始终批判性地评估你的不确定度估算：它们是否合理？你是否考虑了反应时间、数据散布和仪器精度？这一反思过程是 IB 物理探究的核心，也是获得个人参与和评价标准高分的必要环节。
Published by TutorHao | Physics Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
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