IB Physics: Data Analysis and Modelling Methods Masterclass | IB物理：数据分析与建模方法详解

📚 IB Physics: Data Analysis and Modelling Methods Masterclass | IB物理：数据分析与建模方法详解

Data analysis and modelling form the backbone of experimental physics in the IB Diploma Programme. Whether you are tackling the Internal Assessment (IA) or preparing for Paper 3, the ability to process raw measurements, quantify uncertainty, and construct meaningful mathematical models is essential for achieving top marks. This comprehensive guide breaks down every core technique you need, from basic error propagation to advanced curve fitting, all aligned with the IB Physics syllabus.

数据分析与建模是IB文凭课程中实验物理学的核心支柱。无论你是在应对内部评估（IA）还是备考Paper 3，处理原始测量数据、量化不确定度以及构建有意义的数学模型的能力，都是取得高分的关键。本综合指南将为你拆解所有核心技巧，从基础误差传播到高级曲线拟合，全部紧扣IB物理大纲要求。

1. Types of Uncertainty: Random vs Systematic | 不确定度的类型：随机误差与系统误差

Before diving into calculations, you must distinguish between two fundamentally different sources of error. Random uncertainty arises from unpredictable fluctuations in measurements — reading a scale, timing a pendulum, or electrical noise in a sensor. These errors scatter results around the true value and can be reduced by repeating measurements and calculating averages. Systematic uncertainty, by contrast, shifts all measurements consistently in one direction, often caused by calibration errors, zero-offsets, or faulty equipment. Repeating measurements will not reduce systematic error.

在进入计算之前，你必须区分两类本质不同的误差来源。随机不确定度源于测量中不可预测的波动——如读取刻度、为单摆计时或传感器中的电噪声。这些误差使结果围绕真值散布，可通过重复测量和计算平均值来减小。相比之下，系统不确定度会使所有测量结果一致地向某一方向偏移，通常由校准误差、零点偏移或设备故障引起。重复测量无法减小系统误差。

In your IB IA, you are expected to identify both types and state whether each can be reduced through repetition. A common assessment criterion rewards explicit discussion of how equipment quality (analogue vs digital) affects each type of uncertainty.

在IB内部评估中，你需要识别这两种误差类型，并说明每种误差是否可通过重复测量来减小。一个常见的评分标准奖励你明确讨论设备质量（模拟式与数字式）如何影响各类不确定度。

2. Absolute, Fractional and Percentage Uncertainty | 绝对、分数与百分比不确定度

Uncertainty can be expressed in three equivalent forms. Absolute uncertainty has the same unit as the measurement, such as (2.50 ± 0.05) m. Fractional uncertainty is the ratio of absolute uncertainty to the measured value, e.g. 0.05 ÷ 2.50 = 0.02. Percentage uncertainty multiplies the fractional value by 100%, giving 2% here. Converting between these forms is a fundamental skill tested throughout the IB course.

不确定度有三种等效的表示形式。绝对不确定度与测量值具有相同的单位，例如(2.50 ± 0.05) m。分数不确定度是绝对不确定度与测量值的比值，例如0.05 ÷ 2.50 = 0.02。百分比不确定度将分数值乘以100%，这里是2%。这三种形式之间的转换是贯穿IB课程的基本技能。

Fractional uncertainty = Δx / x

分数不确定度 = 绝对不确定度 / 测量值 = Δx / x

For digital instruments, the absolute uncertainty is typically taken as half the smallest digit displayed (e.g. ±0.005 A for a multimeter reading to 0.01 A). For analogue instruments, it is usually half the smallest scale division, though some examiners accept one full division. Always state your convention explicitly in your IA.

对于数字仪器，绝对不确定度通常取最小显示数字的一半（例如，读取精度为0.01 A的万用表，不确定度为±0.005 A）。对于模拟仪器，通常取最小刻度的一半，尽管有些考官接受一个完整刻度。始终在IA中明确说明你所采用的约定。

3. Combining Uncertainties in Calculations | 计算中的不确定度合成

When measurements with uncertainties are used in calculations, you must propagate the errors. The rules follow a clear hierarchy. For addition and subtraction, add absolute uncertainties: if p = a + b or p = a − b, then Δp = Δa + Δb. For multiplication and division, add fractional (or percentage) uncertainties: if p = a × b or p = a ÷ b, then Δp/p = Δa/a + Δb/b.

当含有不确定度的测量值参与计算时，你必须传播这些误差。规则遵循清晰的层级。对于加法和减法，直接相加绝对不确定度：如果p = a + b或p = a − b，则Δp = Δa + Δb。对于乘法和除法，相加分数（或百分比）不确定度：如果p = a × b或p = a ÷ b，则Δp/p = Δa/a + Δb/b。

For powers and roots, multiply the fractional uncertainty by the exponent: if p = aⁿ, then Δp/p = n × Δa/a. This rule covers squares, cubes, square roots, and inverse relationships. A constant multiplier has no effect on the fractional uncertainty of the variable it multiplies.

对于幂和根，将分数不确定度乘以指数：如果p = aⁿ，则Δp/p = n × Δa/a。此规则涵盖平方、立方、平方根和反比关系。常数乘数不影响其所乘变量的分数不确定度。

p = aⁿ → Δp/p = n × (Δa/a)

In practice, keep all intermediate values unrounded until the final step. Round the final uncertainty to one significant figure, then round the measurement to match the decimal place of the uncertainty. This prevents rounding errors from masking the true precision of your result.

实践中，所有中间值保留不四舍五入，直到最后一步。最终不确定度保留一位有效数字，然后使测量值的小数位与不确定度对齐。这可以防止舍入误差掩盖你结果的真实精度。

4. Finding the Uncertainty in a Gradient and Intercept | 直线斜率和截距的不确定度

When you plot a straight-line graph, the best-fit line gives the gradient and intercept. To find their uncertainties, you must draw maximum and minimum slope lines. The maximum-slope line passes through the error bars of the first and last data points in the most extreme clockwise orientation; the minimum-slope line passes through them in the most extreme counter-clockwise orientation. The uncertainty in the gradient is half the difference between these two gradients.

当你绘制直线图时，最佳拟合线给出斜率和截距。要找到它们的不确定度，你必须画出最大斜率和最小斜率线。最大斜率线以最极端的顺时针方向穿过第一个和最后一个数据点的误差棒；最小斜率线则以最极端的逆时针方向穿过它们。斜率的不确定度等于这两条线斜率差值的一半。

Δm = (m_max − m_min) / 2

Δm = (最大斜率 − 最小斜率) / 2

Similarly, the uncertainty in the y-intercept is half the difference between the intercepts of the maximum and minimum slope lines. In recent IB examiners’ reports, candidates who draw these lines lightly in pencil and show all working receive higher marks in the “Processing” criterion. Use error bars of at least 1 mm on both axes to justify your chosen line spread.

类似地，y截距的不确定度等于最大和最小斜率线截距差值的一半。在近年IB考官报告中，那些用铅笔轻轻画出这些线并展示所有步骤的考生在”数据处理”标准中获得更高分数。在两个坐标轴上使用至少1毫米的误差棒来证明你所选直线的离散范围是合理的。

5. Linearization Techniques: From Curves to Straight Lines | 线性化技术：从曲线到直线

Many physical relationships are nonlinear, such as exponential decay, inverse-square laws, or power functions. The most powerful modelling skill in IB Physics is transforming these relationships into linear form through appropriate variable changes. A straight-line graph of y against x with gradient m and intercept c can then be used to extract physical constants.

许多物理关系是非线性的，例如指数衰减、平方反比律或幂函数。IB物理中最强大的建模技能是通过适当的变量变换将这些关系转化为线性形式。然后，以y对x绘制的斜率为m、截距为c的直线图可用于提取物理常数。

Consider the relationship T = 2π√(L/g) for a simple pendulum. Squaring both sides gives T² = (4π²/g)L, so plotting T² against L yields a straight line through the origin with gradient 4π²/g. For radioactive decay, N = N₀e^(−λt), taking natural logarithms gives ln N = ln N₀ − λt, so a plot of ln N versus t has gradient −λ and intercept ln N₀.

考虑单摆关系T = 2π√(L/g)。两边平方得T² = (4π²/g)L，因此以T²对L作图得到一条过原点的直线，斜率为4π²/g。对于放射性衰变，N = N₀e^(−λt)，取自然对数得ln N = ln N₀ − λt，因此ln N对t作图，斜率为−λ，截距为ln N₀。

Original relation	Linear form	Gradient
y = ax²	y vs x²	a
y = a√x	y vs √x	a
y = a/x	y vs 1/x	a
y = axⁿ	ln y vs ln x	n
y = Ae^(kx)	ln y vs x	k

Table 1 | 表1：常见物理关系的线性化形式

6. The Line of Best Fit and Its Physical Meaning | 最佳拟合线及其物理意义

The line of best fit is not merely a pencil stroke through scattered points — it is a visual representation of the hypothesized mathematical model. In IB Physics, you should draw the line that minimizes the total vertical distance of all points from the line. The line must pass through the centroid of the data (the point of average x and average y) when error bars are symmetric, which is a quick check of your drawing accuracy.

最佳拟合线不仅仅是穿过散点的一笔铅笔线条——它是所假设数学模型的可视化表示。在IB物理中，你应当画出使所有数据点到线的总垂直距离最小的直线。当误差棒对称时，直线必须通过数据质心（平均x和平均y对应的点），这是检查作图准确性的快速方法。

The physical meaning of gradient and intercept must be linked directly to the theory being tested. If you plot voltage against current for a resistor, the gradient is the resistance R. If you plot the stopping potential versus frequency in the photoelectric effect, the gradient is the Planck constant divided by the elementary charge, and the intercept gives the work function divided by the elementary charge. Always state these meanings explicitly in your conclusion.

斜率和截距的物理意义必须直接与被检验的理论相联系。如果你绘制电阻中电压对电流的图，斜率就是电阻R。如果你在光电效应中绘制遏止电压对频率的图，斜率是普朗克常数除以元电荷，截距给出逸出功除以元电荷。务必在结论中明确陈述这些意义。

7. Pearson Correlation Coefficient and Coefficient of Determination | 皮尔逊相关系数与决定系数

The Pearson correlation coefficient, r, quantifies the strength and direction of a linear relationship between two variables, ranging from −1 to +1. An r value near +1 or −1 indicates a strong linear correlation; a value near zero indicates no linear correlation. The coefficient of determination, r², represents the proportion of variance in the dependent variable that is predictable from the independent variable.

皮尔逊相关系数r量化两个变量之间线性关系的强度和方向，取值范围从−1到+1。r值接近+1或−1表示强线性相关；接近零表示无线性相关。决定系数r²表示因变量的方差中可由自变量预测的比例。

r = Σ((xᵢ − x̄)(yᵢ − ȳ)) / √(Σ(xᵢ − x̄)² · Σ(yᵢ − ȳ)²)

In your IB Physics IA, reporting r or r² from your calculator or spreadsheet software demonstrates quantitative rigour. However, note that a high correlation does not prove causation — the relationship may be confounded by a third variable. This critical thinking is often rewarded in the “Evaluation” criterion.

在IB物理IA中，报告计算器或电子表格软件给出的r或r²值体现了定量严谨性。但需注意，高相关性并不证明因果关系——关系可能受到第三个变量的混杂影响。这种批判性思维在”评估”标准中常获加分。

8. Residual Analysis: Evaluating Model Fit | 残差分析：评估模型拟合质量

Residuals are the vertical differences between each data point and the predicted value from the best-fit line. Plotting residuals against the independent variable reveals whether a linear model is appropriate. If residuals are randomly scattered around zero with no discernible pattern, the linear model is a good fit. If residuals form a U-shape or curve, a nonlinear relationship is present, indicating that linearization was incomplete or incorrect.

残差是每个数据点与最佳拟合线预测值之间的垂直差异。将残差对自变量作图可以揭示线性模型是否合适。如果残差在零附近随机散布且无可辨别的模式，则线性模型拟合良好。如果残差形成U形或曲线形状，则存在非线性关系，表明线性化不完整或不正确。

In the IB context, residual analysis is rarely required explicitly, but understanding it helps you diagnose why a graph deviates from the expected line. Common causes of systematic residual patterns include unaccounted damping, friction, thermal expansion, or a calibration offset that varies with the measured quantity.

在IB背景下，残差分析很少被明确要求，但理解它有助于你诊断为何图形偏离预期直线。系统性残差模式的常见原因包括未考虑的阻尼、摩擦、热膨胀或随被测量变化的校准偏移。

9. Propagation of Uncertainty in Graphical Models | 图形模型中的不确定度传播

Once you have determined the gradient and intercept from a graph, you may need to propagate their uncertainties into a derived physical quantity. For example, if you determine the acceleration due to gravity g from the gradient m of a T² versus L graph using g = 4π²/m, then the fractional uncertainty in g equals the fractional uncertainty in m.

一旦你确定了图形的斜率和截距，你可能需要将它们的不确定度传播到派生物理量中。例如，如果你利用T²对L图的斜率m通过g = 4π²/m确定重力加速度g，则g的分数不确定度等于m的分数不确定度。

Δg/g = Δm/m

For more complex dependencies, combine uncertainties using the rules from Section 3. If a calculated quantity involves both a gradient and an intercept, treat them as independent variables and sum their fractional contributions in quadrature if using advanced statistics, or simply add them linearly for a conservative estimate, which is acceptable at SL and HL.

对于更复杂的依赖关系，使用第3节的规则合成不确定度。如果计算量同时涉及斜率和截距，将它们视为独立变量。使用高级统计时可按平方和开根号的方式合成，或为保守估计直接线性相加——这在SL和HL中都是可接受的。

10. Log-Log Plots and Power Laws | 双对数图与幂律关系

Many physical laws take the form y = kxⁿ, where n is a constant exponent. Taking logarithms of both sides yields log y = log k + n log x. A plot of log y against log x is a straight line with slope n and intercept log k. This technique is essential for identifying the exponent in relationships where n is unknown, such as the period of a planet versus its orbital radius.

许多物理定律采取y = kxⁿ的形式，其中n是常数指数。对两边取对数得到log y = log k + n log x。以log y对log x作图得到斜率为n、截距为log k的直线。该技术对于在n未知的关系中确定指数至关重要，例如行星公转周期与其轨道半径的关系。

Whether you use base-10 or natural logarithms does not affect the slope — both yield the same n value. However, the intercept differs depending on the base. State clearly which logarithm you used so that your intercept can be correctly interpreted. In a log-log plot, error bars that are symmetric in y become asymmetric in log y, so careful propagation is required.

无论使用以10为底还是自然对数，都不会影响斜率——两者给出的n值相同。然而，截距因底数不同而不同。明确说明你使用的对数类型，以便正确解释截距。在双对数图中，y方向对称的误差棒在log y中变成不对称的，因此需要仔细传播。

11. Using Spreadsheets and Graphing Calculators Effectively | 有效使用电子表格与绘图计算器

Modern IB assessment allows the use of graphing calculators and spreadsheet software for data analysis. You should be proficient in using linear regression functions, computing Pearson’s r, and generating residual plots. In Excel or Google Sheets, the LINEST function returns regression statistics including slope, intercept, and their standard errors; the RSQ function returns r².

现代IB评估允许使用绘图计算器和电子表格软件进行数据分析。你应当熟练使用线性回归函数、计算皮尔逊r以及生成残差图。在Excel或Google Sheets中，LINEST函数返回回归统计量，包括斜率、截距及其标准误差；RSQ函数返回r²。

In your IA report, do not simply paste raw software output. Instead, present the regression equation, quote the r² value, and translate the parameters into physical quantities. Explaining your processing steps clearly — rather than relying solely on software — demonstrates the conceptual understanding that examiners reward.

在IA报告中，不要简单粘贴软件原始输出。相反，呈现回归方程，引用r²值，并将参数转化为物理量。清晰解释你的处理步骤——而非仅依赖软件——展示了考官所欣赏的概念理解能力。

12. Model Evaluation: Limitations, Assumptions and Improvements | 模型评估：局限、假设与改进

A mathematical model of experimental data is always an idealization. Evaluating its limitations requires you to compare the theoretical assumptions with the actual experimental conditions. For instance, a pendulum model assumes small-angle approximation (sin θ ≈ θ), a point-mass bob, and negligible air resistance. If your experiment used large amplitudes or a non-spherical bob, the model—and therefore the extracted value of g—will be systematically biased.

实验数据的数学模型总是一种理想化。评估其局限需要你将理论假设与实际实验条件进行比较。例如，单摆模型假设小角度近似（sin θ ≈ θ）、质点摆锤和可忽略的空气阻力。如果你的实验使用大摆角或非球形摆锤，模型——以及由此提取的g值——将存在系统性偏差。

To improve your model, consider extending it: include damping corrections, use the moment of inertia for a physical pendulum, or plot a dimensionless quantity that removes scale effects. A strong IA conclusion acknowledges the model’s boundaries, quantifies the impact of each assumption on the final uncertainty, and proposes concrete, actionable improvements.

为改进模型，可考虑扩展它：加入阻尼修正、对物理摆使用转动惯量、或绘制消除尺度效应的无量纲量。一份出色的IA结论应承认模型的边界，量化每个假设对最终不确定度的影响，并提出具体、可操作的改进建议。

Published by TutorHao | Physics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

IB Physics: Data Analysis and Modelling Methods Masterclass | IB物理：数据分析与建模方法详解