📚 The Normal Distribution | 正态分布
The normal distribution is one of the most fundamental concepts in statistics. It models continuous data that clusters around a central mean, with frequencies decreasing symmetrically as values move further away. In the IB Mathematics curriculum, understanding the normal distribution is essential for both Analysis & Approaches (AA) and Applications & Interpretation (AI) courses, as it underpins hypothesis testing, confidence intervals, and the normal approximation to the binomial distribution. Its bell-shaped curve appears naturally in countless real-world measurements, from heights and test scores to measurement errors and biological traits.
正态分布是统计学中最基本的概念之一。它描述的是围绕中心均值聚集、随着偏离均值频率对称递减的连续数据。在 IB 数学课程中，无论是分析与方法（AA）还是应用与解释（AI）课程，理解正态分布都至关重要，因为它是假设检验、置信区间以及二项分布正态近似的基础。这条钟形曲线自然地出现在无数的现实测量中，从身高、考试成绩到测量误差和生物性状，无处不在。
1. Introduction to the Normal Distribution | 正态分布简介
The normal distribution, also called the Gaussian distribution, is a continuous probability distribution for a random variable X. It is defined by two parameters: the population mean mu (μ) and the population standard deviation sigma (σ). A normal random variable can take any real value, but most observations fall within a few standard deviations of the mean. In IB notation, we write X ~ N(μ, σ²) to indicate that X follows a normal distribution with variance σ². The shape is famously symmetric and bell-shaped, with the peak at the mean, and the tails extending indefinitely in both directions.
正态分布，又称高斯分布，是随机变量 X 的一种连续型概率分布。它由两个参数定义：总体均值 μ 和总体标准差 σ。正态随机变量可以取任意实数值，但大多数观测值都落在均值周围几个标准差范围内。在 IB 记法中，我们用 X ~ N(μ, σ²) 表示 X 服从方差为 σ² 的正态分布。其形状是对称的钟形曲线，峰值在均值处，两侧尾部无限延伸。
The probability density function (PDF) is f(x) = (1/(σ√2π ents are not asked to memorize or integrate this function analytically, they must interpret the area under the curve as probability. Because the total area under any probability density curve equals 1, the normal curve is a valid probability model for continuous random variables.
概率密度函数（PDF）为 f(x) = (1/(σ√2π)) e^[−(x−μ)²/(2σ²)]。虽然 IB 学生不需要记住或解析积分该函数，但必须理解曲线下面积代表概率。由于任何概率密度曲线下的总面积等于 1，正态曲线是连续型随机变量的有效概率模型。
2. Properties of the Normal Curve | 正态曲线的性质
es away from μ, the curve falls smoothly, and the rate of decline is controlled by σ. The inflection points — where the curvature changes from concave down to concave up — roperty is that the total area under the curve equals 1, representing a probability of 100%.
正态曲线具有几个关键属性，使其可预测且应用广泛。它关于均值 μ 对称，因此均值、中位数和众数三者相等。曲线呈单峰，在 x = μ 处有一个峰值。随着 x 远离 μ，曲线平滑下降，下降速率由 σ 控制。拐点——曲率由凹向下变为凹向上的点——恰好出现在 μ ± σ 处。另一个重要性质是曲线下的总面积为 1，代表概率为 100%。
r any a > 0. This symmetry is crucial for calculating probabilities using standard normal tables or technology.
此外，正态曲线渐近于水平轴，意味着它永远不会真正接触到水平轴。尾部无限接近零但不等于零，反映了极端值有可能但极其不可能的思想。由于对称性，对于任何 a > 0，都有 P(X < μ − a) = P(X > μ + a)。这种对称性对于使用标准正态表或计算器计算概率至关重要。
3. Parameters: Mean and Standard Deviation | 参数：均值与标准差
The mean μ determines the center of the distribution. Changing μ shifts the entire curve left or right along the horizontal axis without altering its shape. The standard deviation σ controls the spread or dispersion. A smaller σ results in a taller, narrower curve, indicating data are tightly clustered around the mean. A larger σ produces a flatter, wider curve, showing greater variability. In IB exam contexts, you may be given μ and σ, or you might need to estimate them from a graph or a set of data.
均值 μ 决定了分布的中心。改变 μ 会使整个曲线沿水平轴左右平移，但不改变形状。标准差 σ 控制着分散程度或离散度。σ 较小会产生更高、更窄的曲线，表明数据紧密聚集在均值周围。σ 较大则会产生较平、较宽的曲线，显示出更大的变异性。在 IB 考试情境中，你可能会得到 μ 和 σ，也可能需要根据图表或数据集来估计它们。
Recall that the variance is σ². In the notation X ~ variance, not the standard deviation. This is a common point of confusion. For example, X ~ lways double-check whether you are working with variance or standard deviation when inputting values into your GDC (graphical display calculator).
回想一下，方差是 σ²。在记法 X ~ N(μ, σ²) 中，第二个参数是方差，而不是标准差。这是一个常见的混淆点。例如，X ~ N(50, 16) 表示 μ = 50，σ = 4，而不是 16。在向图形计算器（GDC）输入数值时，务必仔细核对使用是方差还是标准差。
4. The Standard Normal Distribution (Z-distribution) | 标准正态分布（Z 分布）
and standard deviation σ = 1, it is called the standard normal distribution, denoted by Z ~ f standard deviations an observation x lies from its mean. The formula for the z-score is:
当正态分布的均值 μ = 0 且标准差 σ = 1 时，称其为标准正态分布，记为 Z ~ N(0, 1)。随机变量 Z 表示一个观测值 x 与其均值相差的标准差个数。z 分数的公式为：
z = (x − μ) / σ
Converting any normal variable to z-scores standardizes it, allowing us to use a single reference table (or calculator function) to find probabilities. The z-score is negative when x is below the mean, zero when x equals the mean, and positive when x is above the mean. In IB exams, you frequently need to find z first before using the inverse normal or standard normal probabilities.
将任何正态变量转换为 z 分数就实现了标准化，使我们能够使用单一的参考表（或计算器函数）来计算概率。当 x 低于均值时 z 分数为负，x 等于均值时为零，x 高于均值时为正。在 IB 考试中，你经常需要先求出 z，然后再使用逆正态或标准正态概率。
two standard deviations above the mean, which quickly tells us it lies in the upper tail.
例如，如果 IQ 分数服从 μ = 100、σ = 15 的正态分布，那么 130 分对应 z = (130−100)/15 = 2。这意味着该分数比均值高出两个标准差，这立刻告诉我们它位于上尾部。
5. Calculating Probabilities with Z-scores | 使用 Z 分数计算概率
e, P(X < a) can be found by inputting lower bound = −∞ (or a very small number like −1099 per bound = a, μ, σ. Many calculators also allow direct calculation of P(a < X < b).
要计算正态随机变量 X 落在某区间内的概率，我们对边界进行标准化，然后使用标准正态累积分布函数 Φ(z) = P(Z ≤ z)。IB 学生会使用 GDC 中的 normalcdf（或 Ncd）功能计算累积概率。例如，P(X < a) 可通过输入下界 = −∞（或一个非常小的数，如 −1099）、上界 = a、μ、σ 求得。许多计算器也支持直接计算 P(a < X < b)。
on-strict inequalities give the same probability. When using the standard normal table (if required in a paper without a calculator), you must be careful with rounding and interpretation of table values.
通常，画出钟形曲线并给所求区域涂上阴影会很有帮助。记住，由于对称性，P(Z > z) = 1 − Φ(z)。此外，对于任何连续分布，P(X = a) = 0，因此严格不等式与非严格不等式给出的概率相同。在使用标准正态表时（如果试卷不允许使用计算器），必须注意取整以及对表值的解读。
A common IB question: Given X ~ mmetry Φ(−0.5) = 1 − Φ(0.5). Technology gives the answer directly as approximately 0.5328.
一个常见的 IB 问题：已知 X ~ N(75, 100)，求 P(70 < X < 85)。这里 σ = 10。标准化：z₁ = (70−75)/10 = −0.5，z₂ = (85−75)/10 = 1.0。然后求 Φ(1.0) − Φ(−0.5)，利用对称性 Φ(−0.5) = 1 − Φ(0.5)。技术工具直接给出答案约为 0.5328。
6. Inverse Normal: Finding Values from Probabilities | 逆正态：根据概率求值
nd the value x given a left-tail probability p. In IB, this is accessed via the calculator' his is extremely useful for constructing confidence intervals and determining percentiles.
逆正态函数允许我们根据左尾概率 p 来求 x 值。在 IB 中，这通过计算器的 invNorm 功能实现。如果已知 P(X < k) = p，那么 k = invNorm(p, μ, σ)。对于标准正态分布，k = invNorm(p, 0, 1) 给出使得 P(Z < z) = p 的 z 分数。这在构建置信区间和确定百分位数时极为有用。
rm(1 − p, μ, σ). For central probabilities, such as finding the symmetric interval that contains the middle 95% of values, work out the tail probabilities: 2.5% the lower boundary is invNorm(0.025, μ, σ) and the upper boundary is invNorm(0.975, μ, σ).
当给定的概率是右尾概率时，进行转换：若 P(X > k) = p，则 P(X < k) = 1 − p，再应用 invNorm(1 − p, μ, σ)。对于中心概率，例如求包含中间 95% 值的对称区间，需计算出尾部概率：每尾 2.5%，因此下边界为 invNorm(0.025, μ, σ)，上边界为 invNorm(0.975, μ, σ)。
to un-standardize: x = μ + zσ. Always check your calculator manual and IB question style.
注意，某些较旧型号计算器的逆正态功能只接受标准正态分布的概率，因此需要你自己反标准化：x = μ + zσ。务必查阅计算器手册和 IB 的题目风格。
7. The Empirical Rule (68–95–99.7) | 经验法则（68–95–99.7%）
For any normal distribution, approximately 68% of the data falls within one standard deviation of the mean (μ ± 1σ), about 95% falls within two standard deviations (μ ± 2σ), and about 99.7% falls within three standard deviations (μ ± 3σ). This empirical rule, also known as the 68–95– imately normal by comparing these percentages, or to use the rule to describe data spread.
对于任何正态分布，大约 68% 的数据落在均值的一个标准差范围内（μ ± 1σ），约 95% 落在两个标准差范围内（μ ± 2σ），约 99.7% 落在三个标准差范围内（μ ± 3σ）。这条经验法则，也称为 68–95–99.7 法则，可以在不用计算器的情况下提供一个快速心算估计。IB 学生可能会被要求通过比较这些百分比来检验数据集是否近似正态，或使用该法则描述数据的分布情况。
ghts of adult women are normally distributed with μ = 165 cm and σ = 6 cm, then roughly 95% f women have heights between 165 − 12 = 153 cm and 165 + 12 = 177 cm. A height outside 153–177 cm would be considered unusual, occurring in only about 5% of the population.
例如，如果成年女性的身高服从 μ = 165 cm、σ = 6 cm 的正态分布，那么大约 95% 的女性身高在 165 − 12 = 153 cm 与 165 + 12 = 177 cm 之间。身高在 153–177 cm 之外会被视为不寻常，仅约 5% 的人口会出现。
e precise probabilities can be found with the standard normal table: μ ± 1σ captures 68.27%, μ ± 2σ captures 95.45%, and μ ± 3σ captures 99.73%. The empirical rule rounds these to the memorable 68–95–99.7 values, which are sufficient for many qualitative interpretations.
更精确的概率可以通过标准正态表得到：μ ± 1σ 包含 68.27%，μ ± 2σ 包含 95.45%，μ ± 3σ 包含 99.73%。经验法则将这些数字四舍五入为易于记忆的 68–95–99.7，对许多定性解释来说已经足够。
8. Normal Approximation to the Binomial Distribution | 二项分布的正态近似
se conditions. If met, the binomial random variable X can be approximated by a normal distribution with μ = np and σ = √npq.
当二项分布 B(n, p) 试验次数较多且 p 不太接近 0 或 1 时，其形状类似于正态曲线。使用正态近似的经验法则为：np > 5（有时为 np ≥ 5）且 nq > 5，其中 q = 1 − p。在 IB 中，可能需要你陈述并验证这些条件。若满足条件，则可用 μ = np、σ = √npq 的正态分布来近似二项随机变量 X。
d k − 0.5. Forgetting the continuity correction is a common error that leads to inaccurate results, especially for moderate n.
必须应用连续性校正，因为二项分布是离散的而正态分布是连续的。要近似 P(X = k)，使用正态模型中的 P(k − 0.5 < X < k + 0.5)。对于 P(X ≤ k)，使用上界为 k + 0.5 的 normalcdf。对于 P(X ≥ k)，使用下界为 k − 0.5。忘记连续性校正是常见错误，尤其在 n 不大时会导致结果不准确。
Example: Suppose X ~ Here np = 30 > 5, nq = 70 > 5, so normal approximation is suitable with μ = 30, σ = √21 ≈ 4.583. To estimate P(X ≤ 25), we use normalcdf(−1099, 25.5, 30, 4.583) ≈ 0.162, which closely matches the exact binomial probability.
示例：假设 X ~ B(100, 0.3)。此时 np = 30 > 5，nq = 70 > 5，因此正态近似合适，μ = 30，σ = √21 ≈ 4.583。要估计 P(X ≤ 25)，我们使用 normalcdf(−1099, 25.5, 30, 4.583) ≈ 0.162，这与精确的二项概率非常接近。
9. Testing for Normality (Q-Q Plots and Goodness-of-fit) | 正态性检验（Q-Q 图与拟合优度）
In data analysis, we often need to assess whether a sample comes from a normal population. IB students may be introduced to graphical methods such as histograms with a superimposed normal curve, and quantile-quantile (Q-Q) plots. In a normal Q-Q plot, the sample quantiles are plotted against the theoretical quantiles of a normal distribution. If the points lie approximately along a straight diagonal line, the data are consistent with normality. Systematic curvature or outliers indicate departure from normality.
在数据分析中，我们经常需要评估样本是否来自正态总体。IB 学生可能会接触到图形化方法，例如叠加正态曲线的直方图，以及分位数-分位数图（Q-Q 图）。在正态 Q-Q 图中，样本分位数被绘制在正态分布的理论分位数上。如果点大致沿一条对角线分布，则数据与正态性一致。系统性的弯曲或离群点表明偏离正态性。
The chi-squared goodness-of-fit test can also be used to test for normality, though it is more common for discrete distributions. To use it, you group the continuous data into intervals and compare observed frequencies with expected frequencies calculated from the normal distribution using estimated (or known) μ and σ. The null hypothesis is that the data follow a normal distribution. A large chi-squared statistic and a small p-value lead to rejection of normality.
卡方拟合优度检验也可用于检验正态性，尽管它更常见于离散分布。使用时，需将连续数据分组为区间，再将观测频率与基于估计（或已知）μ 和 σ 计算出的期望频率进行比较。原假设是数据服从正态分布。较大的卡方统计量和较小的 p 值会导致拒绝正态性。
IB exams may ask you to comment on normality based on given plots or descriptive statistics. For instance, if the sample mean and median are very different, or the skewness is far from zero, normality may be questionable. Recognising the assumptions behind statistical tests is an important skill.
IB 考试可能会要求你根据给出的图或描述性统计量评论正态性。例如，如果样本均值和中位数差异很大，或偏度远离零，正态性可能存疑。识别统计检验背后的假设是一项重要技能。
10. Real-world Applications and IB Exam Tips | 实际应用与 IB 考试技巧
The normal distribution models many natural and social phenomena. In biology, it describes traits like height, blood pressure, and leaf length. In quality control, manufacturers use normal distribution to monitor product weights and tolerances. In finance, log-returns are often assumed normal. In IB internal assessments, you might collect data and test for normality before applying a t-test or creating confidence intervals. Always state the assumption that the underlying population is normally distributed when necessary.
正态分布模拟了许多自然和社会现象。在生物学中，它描述身高、血压、叶长等性状。在质量控制中，制造商使用正态分布监控产品重量和公差。在金融领域，对数收益率常被假定为正态分布。在 IB 内部评估中，你可能会收集数据并在应用 t 检验或创建置信区间前检验正态性。必要时，务必陈述总体服从正态分布的假设。
Exam tip: Always draw a sketch to visualise the problem before computing. Identify what is given — probabilities, bounds, mean, standard deviation — and what is asked. Use the correct calculator syntax: normalcdf for probability, invNorm for quantile. Watch for wording such as "at least", "exceeds", "between", which guide the tail areas. If using the standard normal table, round z-scores to two decimal places as required. And never forget the continuity correction when approximating a binomial distribution.
考试技巧：在计算前始终画出草图来可视化问题。确定已知条件——概率、边界、均值、标准差——以及所求。使用正确的计算器语法：normalcdf 用于概率，invNorm 用于分位数。留意“至少”、“超过”、“介于”等用语，这引导着尾部区域的选取。如果使用标准正态表，按要求将 z 分数四舍五入到两位小数。在近似二项分布时，永远不要忘记连续性校正。
Finally, be aware that technology can sometimes give slightly different answers due to rounding or syntax. In IB, candidates are expected to show working, including the standardisation step (z-score) and the reverse standardisation step when interpreting inverse normal results. Clear communication of method is as important as the final answer.
最后，要注意由于舍入或语法不同，技术工具有时会给出略微不同的答案。在 IB 中，考生应写出解题步骤，包括标准化步骤（z 分数）以及在解释逆正态结果时的反标准化步骤。清晰地表达方法与得出最终答案同等重要。
Published by TutorHao | IB Mathematics 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