A-Level CAIE Statistics: Common Misconceptions and Corrections | A-Level CAIE 统计：常见误区与纠正方法

📚 A-Level CAIE Statistics: Common Misconceptions and Corrections | A-Level CAIE 统计：常见误区与纠正方法

Many A-Level CAIE Statistics candidates lose marks not because they cannot calculate, but because they apply a correct formula to the wrong situation or interpret a result incorrectly. This revision guide collects the most common misconceptions in Probability & Statistics 1 and 2 and shows how to correct each one.

许多 A-Level CAIE 统计学考生丢分，不是因为不会计算，而是因为把正确的公式用在了错误的情境中，或对结果做出了错误解释。本复习指南汇总了概率与统计 1 和 2 中最常见的误区，并逐一说明纠正方法。

1. Mutually Exclusive and Independent Events | 互斥事件与独立事件

A frequent error is to treat ‘mutually exclusive’ and ‘independent’ as the same idea. Mutually exclusive means P(A ∩ B) = 0; the events cannot occur together. Independence means P(A ∩ B) = P(A)P(B); knowing that A has occurred does not change the probability of B.

一个常见错误是把 ‘互斥’ 和 ‘独立’ 当作同一概念。互斥意味着 P(A ∩ B) = 0，两个事件不可能同时发生。独立意味着 P(A ∩ B) = P(A)P(B)，即已知 A 发生不会改变 B 发生的概率。

If two events have positive probabilities, they cannot be both mutually exclusive and independent. For mutually exclusive events P(A ∩ B) = 0, but independence would require P(A)P(B) > 0, which is a contradiction. Use a Venn diagram to check whether the intersection is empty before multiplying probabilities.

如果两个事件的概率都为正，它们不可能既互斥又独立。互斥事件满足 P(A ∩ B) = 0，而独立要求 P(A)P(B) > 0，二者矛盾。在相乘概率之前，先用韦恩图检查交集是否为空。

In tree diagrams, probabilities on the second branches are often conditional. If the second event is independent of the first, the second-branch probability is the same as the marginal probability; otherwise it is a conditional probability such as P(B|A).

在树形图中，第二级分支上的概率通常是条件概率。如果第二个事件与第一个事件独立，则第二级分支概率与边缘概率相同；否则它是 P(B|A) 之类的条件概率。

2. Reversing Conditional Probabilities | 颠倒条件概率

Candidates often confuse P(A|B) with P(B|A). These are generally not equal. The correct relationship is P(A|B) = P(A ∩ B) / P(B). For example, P(disease|positive) is not the same as P(positive|disease), which is the sensitivity of a test.

考生常把 P(A|B) 与 P(B|A) 混为一谈。它们一般不相等。正确关系是 P(A|B) = P(A ∩ B) / P(B)。例如，P(患病|阳性) 不等于 P(阳性|患病)，后者才是检验的敏感度。

When a question describes selection without replacement or gives a condition, identify the denominator carefully. The denominator is the probability or frequency of the event after the vertical bar, not the total of the original sample unless the condition has no effect.

当题目描述不放回抽取或给出条件时，要仔细确定分母。分母是竖线后面那个事件的概率或频数，而不是原始样本总数，除非该条件没有影响。

Bayes’ theorem is useful for reversing conditional probabilities: P(A|B) = P(B|A)P(A) / [P(B|A)P(A) + P(B|A’)P(A’)]. Many diagnostic-test and false-positive questions reduce to this formula.

贝叶斯定理用于颠倒条件概率：P(A|B) = P(B|A)P(A) / [P(B|A)P(A) + P(B|A’)P(A’)]。许多诊断检验和假阳性问题都可以归结为该公式。

3. Discrete and Continuous Distributions | 离散分布与连续分布

For a discrete distribution such as the binomial, geometric or Poisson, P(X = x) can be positive and P(X ≤ x) is not the same as P(X < x). For example, in the binomial, P(X < 5) = P(X ≤ 4). Candidates often lose marks by using the wrong form of the inequality.

对于二项、几何或泊松等离散分布，P(X = x) 可以为正，且 P(X ≤ x) 与 P(X < x) 不相同。例如，在二项分布中，P(X < 5) = P(X ≤ 4)。考生常因使用错误的不等式形式而丢分。

For a continuous distribution such as the normal distribution, P(X = x) = 0 for any single value, so P(X ≤ k) = P(X < k). There is no need to add or subtract one; the difference only matters for discrete variables and when using the normal approximation to a discrete variable.

对于正态分布等连续分布，任意单点的概率 P(X = x) = 0，所以 P(X ≤ k) = P(X < k)。无需加一或减一；这种差别只对离散变量以及用正态近似离散变量时才有意义。

Also, a discrete random variable is represented by a probability mass function and a bar chart, while a continuous random variable has a probability density function and a smooth curve. Probability for continuous variables is represented by area under the curve, not by the height of the curve.

此外，离散随机变量用概率质量函数和条形图表示，而连续随机变量有概率密度函数和光滑曲线。连续变量的概率用曲线下面积表示，而不是用曲线高度表示。

4. Normal Approximation to the Binomial | 二项分布的正态近似

The normal approximation X ~ B(n, p) to Y ~ N(np, npq) should only be used when the distribution is sufficiently symmetric. The usual guideline in CAIE is np > 5 and nq > 5. If these conditions fail, the approximation is unreliable.

把 X ~ B(n, p) 近似为 Y ~ N(np, npq) 时，只应在分布足够对称的情况下使用。CAIE 中的通常准则是 np > 5 且 nq > 5。如果条件不满足，近似就不可靠。

Because the binomial is discrete and the normal is continuous, every boundary must be corrected by 0.5. For example, P(X ≤ 10) is approximated as P(Y < 10.5), P(X ≥ 10) as P(Y > 9.5), and P(X = 10) as P(9.5 < Y < 10.5).

由于二项分布是离散的而正态分布是连续的，每个边界都必须做 0.5 连续性校正。例如，P(X ≤ 10) 近似为 P(Y < 10.5)，P(X ≥ 10) 近似为 P(Y > 9.5)，P(X = 10) 近似为 P(9.5 < Y < 10.5)。

A common mistake is to ignore the continuity correction or to apply it in the wrong direction. Draw the interval on a number line and ask whether the endpoint should be included in the desired probability; then expand the interval by 0.5 toward the excluded side.

常见的错误是忽略连续性校正，或校正方向搞反。可以在数轴上画出区间，判断端点是否应包含在所需概率中；然后朝排除的一侧将区间扩展 0.5。

5. p-values and Significance Levels | p 值与显著性水平

The p-value is the probability of obtaining a test statistic at least as extreme as the one observed, assuming the null hypothesis H0 is true. It is not the probability that H0 is true, and it is not the probability that the alternative hypothesis is false.

p 值是在原假设 H0 为真的条件下，得到至少与观测值一样极

Published by TutorHao | A-Level 统计 Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）