📚 Investigation 2: Mean and Standard Deviation of a Binomial Distribution | 探究2：二项分布的均值与标准差
A binomial distribution is one of the most fundamental discrete probability models in IB Mathematics. It describes the number of successes in a fixed number of independent trials, each with the same probability of success. In this investigation, we will explore two summary measures – the mean (expected value) and the standard deviation – and derive their formulas using different approaches. Understanding these measures helps us interpret binomial data and make predictions about real-world situations, from quality control to medical trials.
二项分布是IB数学中最基本的离散概率模型之一。它描述了在固定次数的独立试验中成功的次数，每次试验的成功概率相同。在本探究中，我们将探索两个汇总指标——均值（期望值）与标准差——并通过不同方法推导其公式。理解这些指标有助于我们解读二项数据并对现实情境做出预测，从质量控制到医学试验。
1. Introduction to Binomial Distribution | 二项分布简介
A discrete random variable X follows a binomial distribution if there are n independent trials, each trial results in either success or failure, and the probability of success p remains constant across trials. We write X ~ B(n, p). Typical examples include the number of heads when tossing a fair coin n times, or the number of defective items in a sample drawn with replacement.
如果一个离散随机变量X满足：有n次独立试验，每次试验结果只有成功或失败，且每次试验成功的概率p保持不变，那么X服从二项分布。我们记作X ~ B(n, p)。典型例子包括抛掷均匀硬币n次出现正面的次数，或有放回抽样中次品的数量。
2. Defining the Random Variable and Parameters | 定义随机变量与参数
ty mass function (PMF) gives the probability of obtaining exactly k successes (0 ≤ k ≤ n):
P(X = k) = ₙCₖ × pᵏ × (1 − p)ⁿ⁻ᵏ
参数n表示固定的试验次数，p是每次试验成功的概率。概率质量函数给出了恰好获得k次成功（0 ≤ k ≤ n）的概率：
P(X = k) = ₙCₖ × pᵏ × (1 − p)ⁿ⁻ᵏ
Here ₙCₖ (read "n choose k") is the binomial coefficient, equal to n! / (k!(n−k)!). This function allows us to compute individual probabilities and forms the basis for finding the mean and variance.
这里的ₙCₖ（读作“n选k”）是二项式系数，等于 n! / (k!(n−k)!)。该函数使我们能够计算每个单独的概率，并为求均值和方差提供了基础。
3. Mean of a Binomial Distribution: Intuitive Idea | 二项分布均值：直观理解
value E(X) = np captures. The mean is a measure of central tendency for the distribution.
如果你进行n次独立试验，每次成功概率为p，直观上你预期平均会看到大约 n × p 次成功。例如，投掷一枚均匀骰子60次并统计出现6点的次数，你预期大约有 60 × (1/6) = 10 次六点。这一直觉正是期望值 E(X) = np 所表达的含义。均值是分布中心趋势的度量。
4. Derivation Using the Definition of Expected Value | 使用期望值定义进行推导
by applying the definition E(X) = Σ k · P(X = k), summing over all possible k from 0 to n.
我们可以通过应用定义 E(X) = Σ k · P(X = k)，对所有可能的 k 从 0 到 n 求和，来正式推导均值。
Substitute the PMF: E(X) = Σk=0n · ₙCₖ pᵏ (1 − p)ⁿ⁻ᵏ. Notice that the term for k=0 vanishes. Use the identity k · ₙCₖ = n · n−1Ck−1 actoring out n and p, the remaining sum is the binomial expansion of (p + (1 − p))ⁿ⁻¹ = 1.
代入概率质量函数：E(X) = Σk=0n k · ₙCₖ pᵏ (1 − p)ⁿ⁻ᵏ。注意 k=0 的项为0。利用恒等式 k · ₙCₖ = n · n−1Ck−1 并重新索引求和。提取公因子 n 和 p 后，剩余的求和正是 (p + (1 − p))ⁿ⁻¹ = 1 的二项展开式。
e result simplifies to E(X) = np. This algebraic proof confirms our intuitive expectation.
结果简化为 E(X) = np。这个代数证明证实了我们的直观预期。
5. Derivation Using Linearity of Expectation (Sum of Bernoulli Trials) | 利用期望的线性性质推导（伯努利试验之和）
A more elegant method recognises that a binomial random variable can be written as the sum of n independent Bernoulli random variables. Let Xi = 1 if the i-th trial is a success, and 0 otherwise. Then X = X1 + X2 + ... + Xn.
更简洁的方法是将二项随机变量表示为n个独立伯努利随机变量之和。设第i次试验成功时 Xi = 1，否则为0。则有 X = X1 + X2 + ... + Xn。
Each Xi has the same Bernoulli distribution with E(Xi) = p. By the linearity of expectation, regardless of dependence, E(X) = Σ E(Xi) = n × p.
每个 Xi 服从相同的伯努利分布，且 E(Xi) = p。根据期望的线性性质，无论独立性如何，都有 E(X) = Σ E(Xi) = n × p。
This approach is particularly powerful because it avoids messy summations and works for any collection of random variables. It is the preferred method in IB examinations when justifying the mean formula.
这种方法特别有效，因为它避免了繁复的求和，且对任何随机变量的集合都成立。这是IB考试中证明均值公式的首选方法。
6. Variance of a Binomial Distribution: Formula and Meaning | 二项分布方差：定义与公式
While the mean tells us the expected number of successes, the variance measures the spread or variability around that mean. For a binomial distribution, the variance is given by:
Var(X) = np(1 − p)
均值告诉我们预期的成功次数，而方差衡量围绕该均值的离散程度或变异性。对于二项分布，方差由下式给出：
Var(X) = np(1 − p)
fixed n, the variance is maximized when p = 0.5, reflecting the greatest uncertainty (a 50‐ nce). It shrinks to zero as p approaches 0 or 1, where the outcome becomes almost certain.
方差同时依赖于n和p。注意对于固定的n，当p = 0.5时方差最大，体现了最大的不确定性（50%对50%的机会）。当p趋近于0或1时，方差缩小至零，此时结果几乎是确定的。
7. Derivation of Variance Using the Bernoulli Sum Approach | 使用伯努利之和法推导方差
Using the same decomposition X = Σ Xi, where the Xi nt random variables, the variance of the sum is the sum of the variances: Var(X) = Σ Var(Xi).
使用相同的表示 X = Σ Xi，其中 Xi 是独立的伯努利变量，我们可以计算 Var(X)。对于独立的随机变量，和的方差等于方差之和：Var(X) = Σ Var(Xi)。
A Bernoulli variable has variance Var(Xi) = p(1 − p) because E(Xi n is both rigorous and straightforward, relying on independence which is a defining assumption of the binomial model.
伯努利变量的方差为 Var(Xi) = p(1 − p)，因为 E(Xi²) = p，故方差 = p − p² = p(1 − p)。因此，Var(X) = n × p(1 − p)。该推导既严谨又直接，依赖于二项模型的基本假设——独立性。
8. Derivation Using the Definition Var(X) = E(X²) − [E(X)]² | 使用定义 Var(X) = E(X²) − [E(X)]² 推导
anipulating similar identities as in the mean derivation, we obtain E(X²) = n(n−1)p² + np.
为了完整性，我们也可以直接从概率质量函数推导方差。我们需要计算 E(X²) = Σ k² · P(X = k)。通过与均值推导类似的恒等式变换，可得 E(X²) = n(n−1)p² + np。
− p). Although more algebraically involved, this method demonstrates the internal consistency of the formulas.
然后，利用 Var(X) = E(X²) − (np)²，得到 np + n(n−1)p² − n²p² = np − np² = np(1 − p)。尽管代数步骤更多，但该方法展示了公式的内在一致性。
9. Standard Deviation and Its Interpretation | 标准差及其解释
The standard deviation is simply the square root of the variance:
σ = √(np(1 − p))
标准差即方差的平方根：
σ = √(np(1 − p))
The standard deviation gives a measure of spread in the same units as X (number of successes). It is used to construct typical intervals. For a binomial distribution, roughly 68% of outcomes lie within one standard deviation of the mean when n is large and p is not too extreme. In IB problems, you may be asked to compare the actual probability of X being within one σ of the mean with the empirical rule.
标准差以与X相同的单位（成功次数）度量离散程度。它用于构建典型区间。对于二项分布，当n较大且p不极端时，大约68%的结果落在均值的一个标准差范围内。在IB问题中，你可能被要求比较X在均值±1σ范围内的实际概率与经验法则。
10. Worked Examples | 实例计算
Example 1: 20 × 0.5 = 10. The variance is Var(X) = 20 × 0.5 × 0.5 = 5, so the standard deviation σ = √5 ≈ 2.236. We would expect about 10 heads, with a typical spread of about 2.2 heads.
示例1：抛掷一枚均匀硬币20次，设X为正面朝上的次数。则n = 20, p = 0.5。均值 E(X) = 20 × 0.5 = 10。方差 Var(X) = 20 × 0.5 × 0.5 = 5，故标准差 σ = √5 ≈ 2.236。我们预期约10次正面，典型离散程度约2.2次。
Example 2: In a multiple‐ σ ≈ 1.677. The low expected score and moderate spread highlight the risk of pure guessing.
示例2：在一项有15道题、每题四个选项的选择题考试中，一名学生完全凭猜测作答。猜对的概率 p = 0.25。预期得分均值 = 15 × 0.25 = 3.75。方差 = 15 × 0.25 × 0.75 = 2.8125，σ ≈ 1.677。较低的期望分数和中等的离散程度突显了纯粹猜测的风险。
These calculations can be quickly verified using a GDC's binomial functions, but understanding the formulas allows you to predict centre and spread without computing the full distribution.
这些计算可以用图形计算器的二项函数快速验证，但理解公式可让你无需计算整个分布即可预测中心与离散程度。
11. Investigation: Exploring Mean and SD with Technology | 探究：用技术探索均值与标准差
n and sample standard deviation, and compare them to the theoretical values μ = 12 and σ ≈ √30×0.4×0.6 = √7 d p. You will observe that as the number of simulations increases, the sample statistics converge to the theoretical parameters. This empirical approach reinforces the derivation and highlights the law of large numbers.
为加深理解，请利用技术模拟二项实验。例如，使用计算器或Python从B(30, 0.4)分布生成500个样本。计算样本均值和样本标准差，并与理论值 μ = 12 和 σ ≈ √30×0.4×0.6 = √7.2 ≈ 2.683 进行比较。尝试不同的 n 和 p。你会发现随着模拟次数的增加，样本统计量收敛于理论参数。这种实证方法巩固了推导，同时也突显了大数定律。
der which conditions (e.g., np ≥ 5 and n(1−p) ≥ 5) does the normal approximation work well? Such exploration is excellent preparation for the IB Internal Assessment.
进一步探究：绘制模拟分布图，并叠加具有相同均值和标准差的正态曲线。在哪些条件下（例如 np ≥ 5 且 n(1−p) ≥ 5）正态近似效果较好？这类探索是IB内部评估的绝佳准备。
12. Common Misconceptions and Exam Tips | 常见误区与考试技巧
Misconception 1: he proportion, then E(Y) = p, but E(X) = np. Always check what the random variable counts.
误区1：混淆样本比例的均值与成功次数的均值。如果 Y = X/n 表示比例，则 E(Y) = p，而 E(X) = np。务必检查随机变量统计的是什么。
Misconception 2: Assuming the standard deviation is always √np. Remember the factor (1−p). Only when p = 0.5 does np(1−p) = n/4, giving σ = √n/2.
误区2：认为标准差总是 √np。请记住因子 (1−p)。仅当 p = 0.5 时，np(1−p) = n/4，得到 σ = √n/2。
Exam Tip: When asked to find the mean and standard deviation of a binomial distribution, write the formulas explicitly and substitute values. If the question involves 'within one standard deviation of the mean', compute the interval (μ − σ, μ + σ) and sum the probabilities for integer values within that interval. Show your reasoning clearly to earn full credit.
考试技巧：当题目要求计算二项分布的均值和标准差时，明确写出公式并代入数值。如果问题涉及“均值的一个标准差范围内”，计算区间 (μ − σ, μ + σ) 并将该区间内所有整数取值的概率求和。清晰地展示推理过程以获得满分。
Published by TutorHao | 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