Mastering Experimental Data Processing & Analysis | 物理实验：实验数据整理与分析技巧

📚 Mastering Experimental Data Processing & Analysis | 物理实验：实验数据整理与分析技巧

Experimental work in A-Level Physics is not just about taking readings; it is about transforming raw numbers into meaningful conclusions. Examiners award marks not only for the accuracy of measurements but also for how systematically you record, process, and analyse data. A well-structured table, a correctly plotted graph, and a properly calculated uncertainty can often make the difference between an average and an excellent practical grade.

A-Level 物理实验不仅仅在于读取数据，更在于如何将原始数据转化为有意义的结论。考官打分时不仅关注测量是否准确，还关注你是否系统性地记录、处理和分析数据。结构清晰的表格、正确绘制的图像以及合理计算的不确定度，往往决定了实验成绩是平庸还是优秀。

1. Recording Raw Data Properly | 原始数据的规范记录

A raw data table must be clear, complete, and unambiguous. Each column should have a physical quantity heading, followed by its unit in brackets. For example, write ‘Length / cm’ rather than just ‘Length’. Every measured value must be recorded to the correct number of decimal places, consistent with the resolution of the instrument used.

原始数据表格必须清晰、完整、无歧义。每一列应包含物理量名称，并在括号内注明单位。例如，应写“长度 / cm”，而不是只写“长度”。每个测量值都要按仪器分辨率保留一致的小数位数。

Always record the instrument resolution at the top of the table, e.g. ‘Vernier caliper resolution = 0.01 cm’.

始终在表格顶部记录仪器分辨率，例如“游标卡尺分辨率 = 0.01 cm”。
Repeat measurements at least twice for each independent variable value, then calculate a mean where appropriate.

每个自变量取值至少重复测量两次，并在适当时计算平均值。
Use the same number of decimal places for all values in a column, even if the last digit is zero.

同一列中所有数值应保留相同的小数位数，即使最后一位是零也要保留。
Record raw readings before rounding; only round at the final processed stage.

先记录原始读数，不要提前四舍五入；只在最后处理阶段进行舍入。

2. Understanding Uncertainty and Error | 理解不确定度与误差

In CIE A-Level Physics, you must distinguish between random errors and systematic errors. Random errors cause readings to spread around the true value, while systematic errors shift all readings consistently in one direction. Uncertainty is the range within which the true value is expected to lie.

在 CIE A-Level 物理中，必须区分随机误差与系统误差。随机误差使读数围绕真值分散，而系统误差使所有读数朝同一方向偏移。不确定度是真实值可能落入的范围。

For a single reading with an analogue scale, the absolute uncertainty is often taken as half the smallest division.

对于模拟刻度的单次读数，绝对不确定度通常取最小分度的一半。
For a digital instrument, the absolute uncertainty is usually the smallest digit displayed, e.g. ±0.01 A for a digital ammeter.

对于数字仪表，绝对不确定度通常是最小显示位数，例如数字电流表为 ±0.01 A。
When taking multiple readings, use half the range or the standard deviation, depending on the instruction.

多次读数时，根据题目要求使用半极差或标准偏差。

Fractional uncertainty = Δx / x
Percentage uncertainty = (Δx / x) × 100%

Always state whether the uncertainty is absolute, fractional, or percentage in your answer.

在答案中务必说明不确定度是绝对、相对还是百分比形式。

3. Significant Figures and Rounding Rules | 有效数字与舍入规则

The number of significant figures in a measurement reflects its precision. A ruler reading of 2.35 cm has three significant figures, whereas 2.3 cm has only two. When you perform calculations, your final answer should not have more significant figures than the least precise value used.

测量结果的有效数字位数反映了其精度。尺子读数 2.35 cm 有三位有效数字，而 2.3 cm 只有两位。在计算中，最终答案的有效数字位数不应超过所用数据中精度最低的那个。

Do not round intermediate values; keep them in your calculator and round only the final result.

不要对中间结果进行舍入；保留在计算器中，只对最终结果舍入。
Leading zeros are not significant: 0.025 m has two significant figures.

前导零不算有效数字：0.025 m 有两位有效数字。
Trailing zeros after a decimal point are significant: 4.50 s has three significant figures.

小数点后的末尾零算有效数字：4.50 s 有三位有效数字。
Uncertainties should be given to one significant figure, e.g. 2.35 ± 0.02 cm, not 2.35 ± 0.023 cm.

不确定度通常保留一位有效数字，例如 2.35 ± 0.02 cm，而不是 2.35 ± 0.023 cm。

4. Processing Data: Derived Quantities | 数据处理：导出量的计算

Once raw data is recorded, you will often calculate derived quantities such as period, resistance, or acceleration. For example, if you measure the time for 20 oscillations, the period is T = t / 20, and its uncertainty is also divided by 20.

记录原始数据后，通常需要计算导出量，例如周期、电阻或加速度。例如，若测量 20 次振荡的总时间，则周期为 T = t / 20，其不确定度也除以 20。

T = t / 20, ΔT = Δt / 20

When adding or subtracting measurements, add absolute uncertainties. When multiplying, dividing, or using powers, add fractional or percentage uncertainties.

当测量值进行加减运算时，不确定度取绝对值相加。当进行乘除或幂运算时，不确定度取相对或百分比形式相加。

Addition/Subtraction: ΔZ = ΔA + ΔB

加/减法：ΔZ = ΔA + ΔB
Multiplication/Division: ΔZ/Z = ΔA/A + ΔB/B

乘/除法：ΔZ/Z = ΔA/A + ΔB/B
Power rule: ΔZ/Z = n × ΔA/A for Z = Aⁿ

幂规则：当 Z = Aⁿ 时，ΔZ/Z = n × ΔA/A

Always show one line of working for each processed value, so the examiner can follow your method.

每个导出量至少要写一行计算过程，以便考官理解你的方法。

5. Tabulating Processed Data | 处理数据的表格化

Processed data should be presented in a second table with clear column headings and correct units. For example, if you measured length l and time t, your processed table might include l/cm, t/s, T/s, T²/s². Use consistent significant figures throughout each column.

处理后的数据应放在第二个表格中，列标题清晰并包含正确单位。例如，若测量长度 l 和时间 t，处理表可包含 l/cm、t/s、T/s、T²/s²。每列内部有效数字要一致。

l / cm	t / s	T / s	T² / s²
20.0	17.8	0.89	0.79
30.0	21.9	1.10	1.21

Notice that T² values are rounded to two decimal places, matching the precision of T. Do not report more decimal places just because your calculator displays them.

注意 T² 值保留两位小数，与 T 的精度一致。不要因为计算器显示更多位数就写出更多小数位。

6. Choosing Axes and Linearising Data | 坐标轴选择与数据线性化

The most powerful way to analyse data is to plot a straight-line graph. This requires you to recognise the mathematical relationship between variables. For example, the period of a pendulum is related to length by T = 2π√(l/g). To make this linear, plot T² against l, because T² = (4π²/g) × l.

分析数据最有效的方法是绘制直线图。这需要你识别变量之间的数学关系。例如，单摆周期满足 T = 2π√(l/g)。为了线性化，应以 T² 对 l 作图，因为 T² = (4π²/g) × l。

Identify the independent variable on the x-axis and the dependent variable on the y-axis.

将自变量放在 x 轴，因变量放在 y 轴。
Choose linear scales so that the points are spread across at least half of each axis.

选择线性刻度，让数据点至少占据每个轴的一半以上。
Label each axis with ‘quantity/unit’, e.g. ‘l / cm’ or ‘T² / s²’.

每个坐标轴标注“物理量/单位”，例如“l / cm”或“T² / s²”。
If the relationship is exponential or power-law, plot ln y against x, or log y against log x.

若关系为指数或幂函数，可绘制 ln y 对 x，或 log y 对 log x 图像。

7. Plotting Points and Error Bars | 描点与误差棒

When drawing the graph, plot each data point with a sharp pencil as a small cross (×) or dot with a circle. Points must be plotted accurately to within half a small square. If uncertainties are known, draw error bars with length equal to 2Δx horizontally or 2Δy vertically.

作图时，用削尖的铅笔以细叉（×）或带圆圈的圆点标记每个数据点。点的位置必须准确到半小格以内。若已知不确定度，应绘制误差棒，其长度等于水平方向 2Δx 或垂直方向 2Δy。

Use a clear symbol, typically ×, because dots can be confused with grid junctions.

建议使用清晰的叉号，因为圆点容易与网格交点混淆。
Draw error bars for both x and y only if required; for many CIE experiments only the y uncertainty matters.

只有当题目要求时才同时绘制 x 和 y 误差棒；很多 CIE 实验只需考虑 y 方向不确定度。
If a point lies far from the best-fit line, check the raw data for a recording error before discarding it as an anomaly.

若某个点明显偏离最佳拟合线，先检查原始数据是否记录错误，再将其判定为异常点。

8. Drawing Best-Fit Line and Measuring Gradient | 绘制最佳拟合线与计算斜率

The line of best fit should be a single straight line that balances the points above and below it. Do not force the line through every point; instead, aim for an even distribution of residuals. The line should be drawn with a sharp pencil and a clear ruler.

最佳拟合线应是一条直线，使上下两侧的数据点数量与距离大致平衡。不要强行让直线穿过每一个点，而应使残差均匀分布。用削尖的铅笔和透明直尺绘制直线。

Gradient = (y₂ − y₁) / (x₂ − x₁)

Choose two points on the best-fit line that are far apart, not data points necessarily, and clearly mark them with coordinates.

在拟合线上选择两个相距较远的点（不一定是数据点），并标出坐标。
Use a large triangle on the graph to show the rise and run, and label the chosen points.

在图上画出足够大的直角三角形以表示纵坐标差和横坐标差，并标注所选点。
When calculating the gradient, do not round intermediate differences; report the final gradient to 2 or 3 significant figures.

计算斜率时，不要对中间差值四舍五入；最终斜率保留两到三位有效数字。

9. Using the Gradient and Intercept | 斜率与截距的应用

For a straight line y = mx + c, the gradient m and intercept c contain physical information. In the pendulum example, if T² is plotted against l, the gradient is m = 4π²/g, so g = 4π²/m. The intercept is expected to be zero, but a small non-zero intercept may indicate a systematic error in timing or measurement.

对于直线 y = mx + c，斜率 m 和截距 c 包含物理信息。在单摆例子中，若以 T² 对 l 作图，斜率为 m = 4π²/g，因此 g = 4π²/m。截距预期为零，但若截距不为零，则可能提示计时或测量中存在系统误差。

Read the intercept directly from the y-axis where the line crosses x = 0, if x = 0 is visible on the graph.

若 x = 0 在坐标轴范围内，直接从 y 轴读取截距。
If the graph does not include x = 0, calculate the intercept using the equation y = mx + c with one point on the line.

若图像范围不包含 x = 0，则用线上某点代入 y = mx + c 计算截距。
Use the gradient uncertainty from maximum and minimum slope lines to express the final result with an uncertainty.

利用最大和最小斜率线计算斜率的不确定度，并给出带不确定度的最终结果。

10. Maximum and Minimum Gradient Lines | 最大与最小斜率线

To estimate the uncertainty in a gradient, draw two additional lines: one with the steepest possible slope that still passes through most error bars, and one with the shallowest possible slope. The uncertainty in the gradient is half the difference between these two gradient values.

为估算斜率的不确定度，可再画两条辅助线：一条尽可能陡但仍穿过大部分误差棒，另一条尽可能平缓。斜率不确定度等于两条斜率差的一半。

Δm = (m_max − m_min) / 2

For example, if m_max = 4.2 s²/m and m_min = 3.8 s²/m, then m = 4.0 ± 0.2 s²/m. Then use this uncertainty to calculate the final percentage uncertainty in g.

例如，若 m_max = 4.2 s²/m，m_min = 3.8 s²/m，则 m = 4.0 ± 0.2 s²/m。接着利用该不确定度计算 g 的百分比不确定度。

11. Identifying Anomalies and Evaluating the Experiment | 识别异常点与评估实验

An anomalous point is one that does not fit the general trend and cannot be explained by the stated uncertainties. When you identify an anomaly, circle it on the graph and exclude it from the line of best fit, but do not remove it from the data table.

异常点是不符合总体趋势且无法用所述不确定度解释的点。识别出异常点后，在图上圈出它，并在绘制最佳拟合线时排除它，但不要把它从数据表中删除。

Always explain possible sources of error, e.g. reaction time, parallax error, heat loss, or air resistance.

务必解释可能的误差来源，例如反应时间、视差、热量散失或空气阻力。
Suggest at least one realistic improvement, such as using a data logger or repeating with smaller intervals.

提出至少一条实际可行的改进建议，例如使用数据采集器或以更小间隔重复实验。
Compare your experimental result with the accepted value and discuss whether the difference is within the calculated uncertainty.

将实验所得结果与公认值比较，并讨论差异是否在计算出的不确定度范围内。

12. Final Checklist for Practical Exams | 实验考试最终检查清单

Before submitting your practical paper, quickly review the following points. They cover the core skills that CIE examiners consistently reward.

提交实验试卷前，快速检查以下几点。它们涵盖了 CIE 考官一贯认可的核心技能。

Checklist Item	Details
Units	Every heading and axis has a unit.
Significant figures	Consistent within each column.
Graph	Pencil, labelled axes, sensible scales, best-fit line.
Gradient calculation	Large triangle, two line points, correct formula.
Uncertainty	Half-range or resolution, propagated correctly.
Conclusion	Relates gradient/intercept to the physical quantity.

Mastering these data-processing techniques will not only improve your practical marks but also deepen your understanding of how physics models are tested in the laboratory. Practice with past papers, plot your graphs neatly, and always keep uncertainties in mind.

掌握这些数据处理技巧不仅能提升你的实验分数，还能加深你对物理模型如何在实验室中被验证的理解。多练习历年真题、规范作图，并始终将不确定度放在心上。

Published by TutorHao | Physics Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）

Mastering Experimental Data Processing & Analysis | 物理实验：实验数据整理与分析技巧