📚 High-Frequency Topics and Common Mistakes in CIE Year 13 Statistics | CIE Year 13 统计高频考点与易错题分析
Mastering CIE Year 13 Statistics (S2) demands more than routine calculation – it requires a sharp awareness of subtle conditions, common examiner traps, and robust interpretation of results. This article identifies the topics that consistently appear in past papers and highlights the typical mistakes that separate A* candidates from the rest.
掌握 CIE Year 13 统计 (S2) 需要的不仅仅是机械计算——它需要对各种微妙条件、常见的出题陷阱以及对结果的稳健解读保持高度敏感。本文梳理了过去试卷中反复出现的高频主题，并重点分析了那些将 A* 考生与普通考生拉开距离的典型错误。
1. The Poisson Distribution: Modelling Rare Events | 泊松分布：建模稀有事件
Students often memorise the Poisson formula P(X = x) = e⁻λ λˣ / x! but overlook the essential conditions: events must occur independently, at a constant average rate λ, and be rare within the given interval. The mean and variance are both λ. A typical mistake is applying the Poisson model to events that are not independent, such as the number of machine breakdowns when a failure increases the chance of further failures.
学生常常记住了泊松公式 P(X = x) = e⁻λ λˣ / x!，却忽略了其核心条件：事件必须独立发生、在给定区间内以恒定平均率 λ 出现，并且是稀有事件。均值和方差都等于 λ。一个典型的错误是将泊松模型用于非独立事件，例如在机器故障会增大后续故障概率的情况下计算故障次数。
Another slip-up occurs when adding Poisson probabilities; candidates sometimes add λ values directly instead of recognising that the sum of independent Poisson variables is also Poisson with parameter equal to the sum of their individual λ. Always ask: are the events truly independent and does the situation call for a fixed λ? If not, a different distribution must be used.
另一个失误出现在泊松概率的加法中；考生有时会直接将 λ 值相加，而没有认识到独立泊松变量之和仍服从泊松分布，且参数等于各个 λ 之和。一定要问自己：事件是否真正独立？情境是否要求一个固定的 λ？如果不是，就必须改用其他分布。
2. Approximating Distributions with the Normal | 正态逼近分布
The normal approximation is a high-frequency topic for both the binomial and Poisson distributions. For binomial B(n, p), the normal N(np, npq) is appropriate when np > 5 and nq > 5. A Poisson Po(λ) is well approximated by N(λ, λ) when λ > 15. The classic error is omitting the continuity correction: for a discrete variable, an interval like X ≥ 10 becomes P(X > 9.5) under the normal curve.
正态逼近是二项分布和泊松分布中的高频考点。对于二项分布 B(n, p)，当 np > 5 且 nq > 5 时，可用正态分布 N(np, npq) 近似。泊松分布 Po(λ) 在 λ > 15 时可以用 N(λ, λ) 很好地进行逼近。最经典的错误是漏掉了连续性校正：对于离散变量，诸如 X ≥ 10 的区间在正态曲线下应变为 P(X > 9.5)。
Students also mix up the conditions for the Poisson approximation to the binomial (n large, p small) and the normal approximation. You must check whether the problem asks for an exact test or an approximation, and always state the approximation condition before applying it. Missing the correction often leads to marks being lost despite correct standardisation.
学生们还会混淆泊松近似二项（n 很大，p 很小）与正态近似二项的条件。你必须判断题目要求的是精确检验还是近似检验，并在应用近似之前始终陈述其条件。漏掉校正常常导致明明标准化正确却依然丢分。
3. Continuous Random Variables: PDFs, CDFs and Properties | 连续随机变量：概率密度函数、累积分布函数与性质
A continuous random variable is described by its probability density function f(x). Many errors arise from treating f(x) as a probability directly, rather than integrating over an interval. The total area under f(x) must equal 1. When finding the median m, candidates set ∫₋∞ᵐ f(x) dx = 0.5, but often misintegrate piecewise functions.
连续随机变量由其概率密度函数 f(x) 描述。许多错误源于将 f(x) 直接当作概率，而不是在一个区间上积分。f(x) 下的总面积必须等于 1。在求中位数 m 时，考生设 ∫₋∞ᵐ f(x) dx = 0.5，但常常在分段函数上积分出错。
Another common slip is forgetting to verify that f(x) ≥ 0 for all x and that the cumulative distribution function F(x) satisfies F(x) = P(X ≤ x). In exam questions, sketching the graph can prevent errors when determining limits. Always pay attention to the domain; a PDF defined in a specific range must be integrated only over that range.
另一个常见的失误是忘记验证 f(x) ≥ 0 对所有 x 成立，以及累积分布函数 F(x) 满足 F(x) = P(X ≤ x)。在考试题中，画出草图可以防止在确定积分限时出错。务必留意定义域；定义在特定范围内的 PDF 只能在该范围内进行积分。
4. Linear Combinations and Variance Rules | 线性组合与方差法则
Expectation is straightforward: E(aX + bY) = aE(X) + bE(Y). Variance, however, is a pitfall. Var(aX + bY) = a²Var(X) + b²Var(Y) + 2abCov(X,Y). If X and Y are independent, Cov(X,Y) = 0, giving Var(aX + bY) = a²Var(X) + b²Var(Y). Candidates frequently forget to square the constants or mistakenly believe Var(X – Y) = Var(X) – Var(Y). In reality, for independent variables, Var(X – Y) = Var(X) + Var(Y).
期望部分很简单：E(aX + bY) = aE(X) + bE(Y)。但方差部分却是一个陷阱。Var(aX + bY) = a²Var(X) + b²Var(Y) + 2abCov(X,Y)。如果 X 和 Y 独立，则 Cov(X,Y) = 0，得到 Var(aX + bY) = a²Var(X) + b²Var(Y)。考生经常忘记将系数平方，或者错误地认为 Var(X – Y) = Var(X) – Var(Y)。实际上，对于独立变量，Var(X – Y) = Var(X) + Var(Y)。
When dealing with sums or differences of normal variables, the result remains normal, but the variance is always the sum of the individual variances when independence holds. A serious exam mistake is using the standard deviation directly instead of variance; always work with variance first, then take the square root if needed.
在处理正态变量的和与差时，结果依然服从正态，但当独立性成立时方差始终是各方差之和。考试中一个严重的错误是直接使用标准差而不是方差；一定要先处理方差，必要时再开平方根。
5. Unbiased Estimators and Sampling Distributions | 无偏估计量与抽样分布
The sample mean x̄ is an unbiased estimator of the population mean μ. For the population variance σ², the unbiased estimator is s² = Σ(xᵢ – x̄)²/(n-1). A frequent error is dividing by n instead of n-1, which gives a biased estimate. This mistake often occurs when students use their calculator’s sample variance function without understanding the difference.
样本均值 x̄ 是总体均值 μ 的无偏估计量。对于总体方差 σ²，其无偏估计量为 s² = Σ(xᵢ – x̄)²/(n-1)。一个常见的错误是除以 n 而不是 n-1，这样做会得到有偏估计。当学生在不理解差异的情况下使用计算器的样本方差功能时，这种错误时常发生。
According to the Central Limit Theorem, the sampling distribution of x̄ is approximately normal for large n, regardless of the population distribution, with mean μ and variance σ²/n. Candidates must be careful to use the correct standard error σ/√n in confidence intervals and tests. Confusing the population standard deviation with the sample standard deviation can invalidate the whole analysis.
根据中心极限定理，无论总体分布如何，当 n 很大时 x̄ 的抽样分布近似服从正态，均值为 μ，方差为 σ²/n。考生必须注意在置信区间和检验中使用正确的标准误 σ/√n。混淆总体标准差与样本标准差可能使整个分析无效。
6. Confidence Intervals for Means and Proportions | 均值和比例的置信区间
A 95% confidence interval for a population mean, when σ² is known, is x̄ ± z₀.₀₂₅ σ/√n. When σ² is unknown and estimated by s², the interval becomes x̄ ± tₙ₋₁,₀.₀₂₅ s/√n, using the t-distribution with n-1 degrees of freedom. The most damaging error is using the z-value instead of the t-value when σ is not known, especially for small sample sizes.
当 σ² 已知时，总体均值的 95% 置信区间为 x̄ ± z₀.₀₂₅ σ/√n。当 σ² 未知并由 s² 估计时，区间变为 x̄ ± tₙ₋₁,₀.₀₂₅ s/√n，此时使用自由度为 n-1 的 t 分布。最具破坏性的错误是在 σ 未知时使用 z 值而不是 t 值，尤其对于小样本而言。
For a population proportion p, the large-sample confidence interval is p̂ ± z₀.₀₂₅ √(p̂(1-p̂)/n), provided np̂ ≥ 5 and n(1-p̂) ≥ 5. Students sometimes forget to check these conditions, leading to invalid intervals. Another subtle point: the margin of error should be reported with appropriate precision, not exaggerated to many decimal places.
对于总体比例 p，大样本下的置信区间为 p̂ ± z₀.₀₂₅ √(p̂(1-p̂)/n)，前提是 np̂ ≥ 5 且 n(1-p̂) ≥ 5。学生有时会忘记检查这些条件，导致区间无效。另一个微妙之处在于，误差范围应以适当的精度报告，而不是夸大至过多的小数位。
7. Hypothesis Testing: One-Sample Tests | 假设检验：单样本检验
H₀ and H₁ must be stated clearly before any calculation. For a test concerning a population mean with unknown variance, the test statistic is t = (x̄ – μ₀)/(s/√n) compared with a t critical value. A common error is writing a two-tailed H₁ but halving the significance level when finding the critical value; the examiner expects the full α/2 in each tail.
在任何计算之前，必须清晰陈述 H₀ 和 H₁。对于未知方差的总体均值检验，检验统计量为 t = (x̄ – μ₀)/(s/√n)，并将其与 t 临界值比较。一个常见错误是写下了双尾 H₁，却在查找临界值时将显著性水平减半；考官期望的是每尾使用 α/2。
Many candidates mix up the p-value approach and the critical region method. If p-value < α, reject H₀. Directly comparing the test statistic with the critical value is also valid, but do not confuse the two in your conclusion. Also, remember that the test statistic is computed assuming H₀ is true; using sample data incorrectly in the standard error formula leads to a wrong conclusion.
许多考生混淆了 p 值法与拒绝域法。如果 p 值 < α，则拒绝 H₀。直接将检验统计量与临界值比较也是可行的，但不要在结论中将两者搞混。另外，请记住检验统计量是在假定 H₀ 为真的条件下计算的；在标准误公式中误用样本数据会导致错误结论。
8. Hypothesis Testing: Two-Sample t-Tests and Normal Tests | 双样本检验：t 检验与正态检验
When comparing two population means with unknown but assumed equal variances, the pooled variance sₚ² = [(n₁-1)s₁² + (n₂-1)s₂²] / (n₁+n₂-2) is used. The test statistic t = (x̄₁ – x̄₂) / sₚ√(1/n₁ + 1/n₂) follows a t-distribution with n₁+n₂-2 df. Forgetting to check the assumption of equal variances can be penalised; sometimes the question provides an F-test for equality.
在比较两个总体均值且未知但假设方差相等时，需使用合并方差 sₚ² = [(n₁-1)s₁² + (n₂-1)s₂²]/(n₁+n₂-2)。检验统计量 t = (x̄₁ – x̄₂) / sₚ√(1/n₁ + 1/n₂) 服从自由度为 n₁+n₂-2 的 t 分布。忘记检验方差相等的假设可能被扣分；有时题目会提供关于方差相等性的 F 检验。
If variances are not assumed equal, the Welch–Satterthwaite approximation is examined less frequently at CIE but may appear. A typical blunder is subtracting the sample means in the wrong order or failing to specify the correct alternative hypothesis. Set up H₁ to match the problem’s context, usually one-tailed if evidence suggests a direction.
如果不假设方差相等，Welch-Satterthwaite 近似在 CIE 中出现频率较低，但可能会考。一个典型的错误是颠倒了样本均值相减的顺序，或者未能指定正确的备择假设。设定 H₁ 要与问题背景匹配，如果证据表明具有方向性，通常采用单尾检验。
9. Type I and Type II Errors and Power | 第一类错误、第二类错误与功效
A Type I error occurs when H₀ is true but is rejected; its probability is α, the significance level. A Type II error occurs when H₀ is false but is not rejected; its probability is β. The power of a test is 1 – β, the probability of correctly rejecting a false H₀. Students often define these correctly but stumble when calculating β under a specific alternative value.
当 H₀ 为真却被拒绝时，发生第一类错误，其概率为 α，即显著性水平。当 H₀ 为假却没有被拒绝时，发生第二类错误，其概率为 β。检验的功效是 1 – β，即正确拒绝一个错误 H₀ 的概率。学生们往往能够正确定义，但在具体备择值下计算 β 时却遇到困难。
The typical calculation involves finding the critical region under H₀, then computing the probability of falling inside that region assuming the true parameter takes the alternative value. A frequent pitfall is failing to adjust the standard error for the alternative distribution, or using the wrong tail when the test is one-tailed. Carefully draw distributions for H₀ and H₁ to avoid confusion.
典型的计算需要先找出 H₀ 下的拒绝域，再假设真实参数取备择值，计算落入该区域的概率。常见的陷阱是未能根据备择假设分布调整标准误，或者在单尾检验中用错了尾部。仔细画出 H₀ 和 H₁ 的分布图可以避免混淆。
10. Chi-Squared Tests for Goodness of Fit and Independence | 卡方检验：拟合优度与独立性
The chi-squared statistic χ² = Σ (O – E)² / E is a staple of S2. For a goodness-of-fit test, degrees of freedom = number of classes – number of estimated parameters – 1. For a contingency table, df = (r-1)(c-1). A common mistake is including cells with very small expected frequencies without combining them; the rule of thumb is that no E should be less than 5.
卡方统计量 χ² = Σ (O – E)² / E 是 S2 的主打内容。对于拟合优度检验，自由度 = 类别数 – 估计的参数个数 – 1。对于列联表，df = (r-1)(c-1)。一个常见的错误是在期望频数很小的单元格存在时没有合并它们；经验法则认为所有 E 都不应小于 5。
When assessing independence, the hypotheses must be stated in words: H₀: no association between the two variables, H₁: some association. The expected frequencies are calculated assuming independence: E = (row total × column total) / grand total. A serious error is comparing
Published by TutorHao | Year 13 统计 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