📚 Year 13 WJEC Statistics: Quick Glossary & Mnemonics Guide | A-Level WJEC 统计词汇术语速记指南
Mastering statistical vocabulary is half the battle in Year 13 WJEC Statistics. This guide pairs each essential term with a clear English explanation, a direct Chinese translation, and memory tricks that link language to logic. Use it alongside your notes to lock in definitions, avoid common exam pitfalss, and build confidence for both written and data-based questions.
掌握统计学术语，是攻克 Year 13 WJEC 统计考试的关键一步。本指南为每个核心术语配上了清晰的英文解释、地道的中文翻译，以及将语言与逻辑挂钩的记忆技巧。配合你的课堂笔记一起使用，可以牢牢锁住定义、避开常见考试陷阱，并在面对文字题和数据分析题时更加自信。
1. Descriptive vs. Inferential Statistics | 描述统计与推断统计
Descriptive statistics summarise and describe the features of a data set, often using measures like mean, median, range, and standard deviation. Think of it as taking a snapshot of what the data looks like right now.
描述统计用来概括和描述数据集的特征，常用的指标包括平均数、中位数、极差和标准差。你可以想象成给当前数据拍一张快照，它只是告诉你数据“长什么样”。
Inferential statistics go a step further: they use sample data to draw conclusions or make predictions about a larger population. Every time you see a hypothesis test or a confidence interval, you are dealing with inferential thinking.
推断统计则更进一步：它利用样本数据对更大的总体进行推断或预测。每当你看见假设检验或者置信区间，你就是在运用推断思维。
Mnemonics: "Descriptive Describes, Inferential Infers." The 'D' words go together, and the 'I' words go together. You describe a dessert; you infer the ingredients.
记忆法：“描述(Descriptive)只描述(Describe)，推断(Inferential)做推断(Infer)”。两个'D'在一起，两个'I'在一起。描述一道甜品就像描述数据；推断则像根据味道猜配料。
2. Population and Sample | 总体与样本
The population is the entire set of individuals or items you want to study. It can be finite (all students in a school) or hypothetical (all possible measurements from a manufacturing process).
总体是你想研究的全部个体或项目构成的集合。它可以是有限总体（如某校所有学生），也可以是假设总体（如同一生产流程中所有可能的测量值）。
A sample is a subset of the population selected to represent it. The ultimate goal is to use the sample to say something sensible about the population.
样本是从总体中选出的、希望代表总体的一个子集。最终目标就是用样本来对总体作出合理的判断。
The word population contains 'pop' mple is just a scooped handful. The handful should taste like the whole bag if well mixed.
单词 Population 里藏着 'pop'，像一整袋爆米花；Sample 只是抓出的一把。如果搅拌均匀，这一把就能代表整袋的口味。
3. Parameter and Statistic | 参数与统计量
A parameter is a numerical summary that describes a population. Greek letters usually represent parameters: μ for population mean, σ for population standard deviation, π for population proportion.
参数是描述总体的一个数值概括。通常用希腊字母表示：μ 代表总体均值，σ 代表总体标准差，π 代表总体比例。
A statistic is a numerical summary calculated from a sample. Roman letters are used: x̄ for sample mean, s for sample standard deviation, p̂ for sample proportion. The hat (^) often signals an estimate.
统计量则是由样本计算出的数值概括。通常用罗马字母表示：x̄ 表示样本均值，s 表示样本标准差，p̂ 表示样本比例。帽子符号 (^) 常提醒这是一个估计值。
Memory link: 'Parameter → Population' both start with P; 'Statistic → Sample' both start with S. And Greek letters feel more 'universal' like a whole population.
记忆链接：'Parameter' 和 'Population' 都以 P 开头；'Statistic' 和 'Sample' 都以 S 开头。此外，希腊字母总给人一种更“普适”的感觉，正好对应整个总体。
4. Sampling Distributions and the Central Limit Theorem | 抽样分布与中心极限定理
The sampling distribution of a statistic (such as the sample mean) is the distribution of all possible values that statistic could take, across all possible samples of the same size from the same population. It is the bridge between a sample and inference.
某个统计量（比如样本均值）的抽样分布，是指从同一总体中抽取所有可能的大小相同的样本时，该统计量所有可能取值的分布。它是连接样本和推断的桥梁。
b), the sampling distribution of the sample mean will be approximately normal, no matter the shape of the original population distribution. The mean of this sampling distribution equals the population mean μ, and its standard deviation (standard error) is σ/√n.
中心极限定理指出，只要样本量足够大（常见经验法则 n ≥ 30），无论原始总体分布形状如何，样本均值的抽样分布都近似服从正态分布。这个抽样分布的均值等于总体均值 μ，其标准差（标准误）等于 σ/√n。
Think of the CLT as a 'normality factory' : feed it any weird-shaped population, draw lots of sample means, and out comes a bell curve. Mnemonic: 'CLT = Creats Life-saving Normality'.
可以把中心极限定理想象成一个“正态化工厂”：输入任意奇形怪状的总体，反复抽取样本均值，最后出来的就是一条钟形曲线。记忆口诀：'CLT = 创造救命正态'(Creates Life-saving Normality)。
5. Hypothesis Testing: Null and Alternative Hypotheses | 假设检验：原假设与备择假设
e null hypothesis (H₀) is a statement of no effect, no difference, or no change. It is the 'status quo' gests otherwise. In WJEC exams, H₀ is often written with an equals sign, e.g. H₀: μ = 100.
原假设 (H₀) 是一个声称“没有效果、没有差异、没有变化”的陈述。它是我们假定为真的“现状”，直到有证据表明情况不同。在 WJEC 考试中，H₀ 通常含有等号，如 H₀: μ = 100。
The alternative hypothesis (H₁ or Hₐ) is what we hope to establish. It can be one-sided (μ > 100 or μ < 100) or two-sided (μ ≠ 100). The choice depends on the wording of the research question.
备择假设 (H₁ 或 Hₐ) 是我们希望证实的陈述。它可以是单侧的（μ > 100 或 μ < 100），也可以是双侧的（μ ≠ 100）。选择哪一类取决于研究问题的措辞。
. Alternative = Action we aim for. The symbol H₀ looks like a zero, reminding us it is the 'zero difference' claim.
记忆窍门：Null（原假设）= Nothing new（没有新东西）。Alternative（备择假设）= 我们瞄准的 Action（行动）。符号 H₀ 很像一个零，提示它本质上是“零差异”声明。
6. Significance Level, p-value, and Decision Rules | 显著性水平、p 值与决策规则
It is the threshold we set before the test; WJEC commonly uses α = 0.05 or 0.01. If α = 5%, we are willing to accept a 5% risk of a false positive.
显著性水平 (α) 是当原假设实际为真时拒绝它的概率。它是我们在检验之前设定的门槛；WJEC 常用 α = 0.05 或 0.01。如果 α = 5%，意味着我们愿意接受 5% 的假阳性风险。
≤ α) signals that the observed result is surprising under H₀ and leads to rejection of H₀.
p 值是假定原假设为真的情况下，观察到当前检验统计量及更极端结果的概率。很小的 p 值（≤ α）意味着在原假设下出现这样的结果非常令人意外，因此我们拒绝 H₀。
A clear decision rule: If p ≤ α, reject H₀. If p > α, do not reject H₀. You never 'accept' H₀; you only say there is insufficient evidence to reject it.
清晰的决策规则：如果 p ≤ α，拒绝 H₀。如果 p > α，不能拒绝 H₀。你永远不能“接受” H₀；只能说尚无充分证据拒绝它。
Visualise α as a boundary line for a high-jump competition. The p-value is the height you actually jump. If your jump (p) clears the bar (α), you can 'reject' the record being unbeatable.
可以把 α 想象成跳高比赛的横杆，p 值是你实际跳过的高度。如果你的跳跃 (p) 没碰到横杆（即 p ≤ α），你就可以“拒绝”纪录无法打破的说法。
7. Type I and Type II Errors, and Power | 第一类错误、第二类错误与检验功效
A Type I error occurs when H₀ is true, but we mistakenly reject it. Its probability is exactly α, the significance level. Mnemonic: 'False Positive, Type I' — think of a fire alarm (positive) triggered by burnt toast when there is no real fire (H₀ true).
第一类错误发生在 H₀ 为真却被错误拒绝的时候。其发生概率正好是 α，即显著性水平。记忆法：“假阳性，第一类”——想象烤糊面包触发火警（阳性），但实际上并无火灾（H₀ 为真）。
A Type II error happens when H₀ is false, but we fail to reject it. Its probability is denoted β. This is a 'false negative': the fire is real, but the alarm stays silent.
第二类错误发生在 H₀ 为假，我们却没有拒绝它。其概率记为 β。这是“假阴性”：真着火了，警报却没有响。
The power of a test is 1 − β, the probability of correctly rejecting a false H₀. High power is desirable and can be increased by larger sample sizes or greater effect sizes.
检验的功效是 1 − β，即正确拒绝一个错误 H₀ 的概率。高功效是我们所追求的，可以通过增加样本量或增大效应量来提高。
Rhyme to recall: 'Type I is a cry when nothing went wrong. Type II is when you miss the true song.' Power is the strength to find what is hidden.
顺口溜：“第一类，虚惊一场；第二类，漏掉真相。” 功效就是发现隐藏事实的侦测力。
8. Confidence Intervals | 置信区间
A confidence interval (CI) gives a range of plausible values for an unknown population parameter, based on sample data. A 95% CI means that if we repeated the sampling process many times, about 95% of the intervals calculated would capture the true parameter.
置信区间给出了基于样本数据的未知总体参数的一个合理取值范围。95% 置信水平意味着，如果重复抽样很多次，大约 95% 计算出的区间将会包含总体真值。
istic ± (critical value) × (standard error). For a large sample mean with known σ, it is x̄ ± z* × σ/√n. If σ unknown and you use s, the t-distribution applies.
均值的置信区间通用公式为：样本统计量 ± (临界值) × (标准误)。对于大样本且 σ 已知的均值，形式为 x̄ ± z* × σ/√n。若 σ 未知并用 s 替代，则使用 t 分布。
Think of a confidence interval like a fishing net cast into the sea. The net (interval) might not always catch the fish (parameter), but a 95% confidence net is constructed so that it retains the fish in 95% of casts.
可以把置信区间想象成撒入大海的一张渔网。渔网（区间）不保证每次都能捕到鱼（参数），但一张 95% 置信网的构造方式，使得在 95% 的撒网中都能兜住那条鱼。
9. Probability Distributions at a Glance | 概率分布速览
ions: fixed n, two outcomes per trial, constant p, independence. Mean = np, variance = np(1−p).
二项分布 B(n, p) 模型化固定 n 次独立试验中成功的次数，每次试验成功概率为 p。适用条件：固定 n 次，每次只有两个结果，p 保持不变，试验间独立。均值 = np，方差 = np(1−p)。
The normal distribution N(μ, σ²) is the most important continuous distribution. Its bell-shaped curve is symmetric about μ. About 68% of values lie within 1σ of μ, 95% within 2σ, 99.7% within 3σ (the empirical rule).
正态分布 N(μ, σ²) 是最重要的连续分布。其钟形曲线关于 μ 对称。大约 68% 的取值落在 μ±1σ 内，95% 落在 μ±2σ 内，99.7% 落在 μ±3σ 内（经验法则）。
For WJEC Year 13, you also encounter the t-distribution, used when the population standard deviation is unknown and estimated by s. It has heavier tails than normal, but approaches normal as degrees of freedom increase.
WJEC Year 13 课程中还会遇到 t 分布，它在总体标准差未知并用 s 估计时使用。t 分布尾部比正态分布更厚，但随着自由度的增加，逐渐趋近正态分布。
Quick acronym for binomial checks: 'FICS' — Fixed number, Independent, Constant p, Success/failure two outcomes. For normal distribution, remember the '68-95-99.7' rule as the 'nearly all' rule.
二项分布条件记忆缩写：'FICS' — Fixed（固定次数）, Independent（独立）, Constant p（恒定概率）, Success/failure（成败两结果）。正态分布记住“68-95-99.7”法则，想象成数据“几乎全部”落在这个范围。
10. Correlation and Regression | 相关与回归
om −1 (perfect negative) to +1 (perfect positive). r = 0 indicates no linear correlation. Correlation does not imply causation.
相关度量两个变量之间线性关系的强度和方向。皮尔逊相关系数 r 取值范围从 −1（完全负相关）到 +1（完全正相关）。r = 0 表示没有线性相关。相关不意味着因果。
ry). The least-squares regression line minimises the sum of squared residuals. Equation: ŷ = a + bx, where b is the slope and a the intercept.
回归更进一步，模型化一个变量（响应变量）如何依赖于另一个变量（解释变量）。最小二乘回归线极小化残差平方和。方程形式为 ŷ = a + bx，其中 b 是斜率，a 是截距。
Residual = observed y − predicted ŷ o for a good linear model. Any pattern suggests the linear model might not be appropriate.
残差 = 观测值 y − 预测值 ŷ。一个好的线性模型对应的残差图应在零附近随机散布。任何规律性模式都暗示线性模型可能不合适。
on, how two move together. Regression = to go back (re-gress) to the line of best fit. And 'Causation is a heavy word, correlation is just a hint.'
记忆点：Correlation（相关）= Co-relation，两个变量如何共同变化。Regression（回归）= re-gress，回到最佳拟合线上去。谨记：“因果(Causation)分量很重，相关(Correlation)只是线索。”
11. Key Notation and Symbols | 关键符号速记
| Symbol | Meaning | Memory Hint |
|---|---|---|
| μ | Population mean | M-shaped letter for Mean |
| x̄ | Sample mean | Bar over x suggests average of x's |
| σ | Population standard deviation | Sigma sounds like 'sum' of deviations |
| s | Sample standard deviation | Small s for sample spread |
| σ² | Population variance | Square of sigma, variance is 'spread squared' |
| p̂ | Sample proportion | Hat on p for estimate |
| H₀ | Null hypothesis | Zero subscript = 'zero difference' |
| H₁ | Alternative hypothesis | 1 stands for 'new idea' |
| α | Significance level | Alpha looks like a fish that you set as threshold |
| β | P(Type II error) | Beta bird hides when false H₀ is there |
| 1 − β | Power of test | 1 minus missed chance = detection strength |
| r | Pearson correlation coefficient | r for relationship |
| R² | Coefficient of determination | Squared r, proportion of variance explained |
Keep this table handy. Writing out symbols and saying their full names aloud once per revision session makes recall automatic in the exam hall.
把这张表格放在手边。每次复习时写下符号并大声说出它们的全称，就能让考场上的回忆变得自动。
12. A Pocket of Mnemonics for the Exam | 考试速记锦囊
Here is a quick-fire summary of proven mnemonics that previous WJEC high achievers used to lock in tricky concepts:
以下是 WJEC 历届高分考生用来记住棘手概念的速记口诀，简明扼要：
- CLT: 'Sample means, large enough n, turn normal — even from chaos.'
- “中心极限定理：样本均值，n 够大，从混乱中也能诞生正态。”
- Hypothesis decision: 'p low, H₀ must go. p high, let it lie.'
- “假设决策：p 值小，H₀ 要抛。p 值大，别动它。”
- Type I vs II: 'Type I sees a ghost. Type II misses the host.'
- “第一类错误：见鬼了。第二类错误：错过了主人。”
- Confidence interval: 'Point estimate plus/minus a few standard errors is the net.'
- “置信区间：点估计 ± 几倍标准误，就是那张网。”
- Correlation: 'Relationship, not rescue — r does not prove causation.'
- “相关：有关联，非拯救——r 并不证明因果。”
- Parameter vs Statistic: 'Greek for the peak (population), Roman for the run (sample).'
- “参数与统计量：希腊给总体之巅，罗马给样本之跑。”
Revisit this cheat sheet the night before your exam to prime your brain with the exact language examiners expect.
考试前夜重温这份速记清单，让你的大脑充满阅卷官所期待的标准语言。
Published by TutorHao | Statistics 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