📚 Finding Probabilities in the t-Distribution | 在t分布中求概率
In many real-world statistical problems, the population standard deviation σ is unknown and the sample size is small. In such cases, the normal distribution cannot be directly applied to the sample mean. Instead, the t-distribution (Student's t-distribution) provides a more accurate model for finding probabilities and constructing confidence intervals for the population mean. This article explains how to find probabilities associated with the t-distribution, a crucial skill for IB and Cambridge A-Level Mathematics.
在许多实际统计问题中，总体标准差σ未知且样本量较小。此时，样本均值的分布不能直接使用正态分布。取而代之的是t分布（学生t分布），它能更准确地为总体均值的概率计算和置信区间构建提供模型。本文将解释如何求与t分布相关的概率，这是IB和剑桥A-Level数学中的关键技能。
1. Understanding the t-Distribution | 理解t分布
mple size n is less than 30 and the population standard deviation σ is estimated by the sample standard deviation s. Its exact shape depends on the degrees of freedom, ν = n - 1.
t分布是一族连续概率分布，它对称且呈钟形，与标准正态分布相似，但尾部更厚。这意味着它更容易产生远离均值的值。当样本量n小于30且总体标准差σ由样本标准差s估计时，就会使用t分布。其确切形状取决于自由度ν = n - 1。
e differences become negligible, and many textbooks allow the use of the normal distribution as an approximation. However, the IB curriculum expects you to use the t-distribution table or your GDC whenever σ is unknown, regardless of sample size, unless stated otherwise.
随着ν增大，t分布趋近于标准正态分布N(0, 1)。事实上，当ν ≥ 30时，差异变得可忽略，许多教科书允许使用正态分布近似。但是，IB课程期望你在σ未知时始终使用t分布表或图形计算器，不论样本量大小，除非另有说明。
2. Degrees of Freedom | 自由度
t reflects the number of independent pieces of information that go into the estimation of the standard deviation. When we use s instead of σ, we lose one degree of freedom because the sample mean itself is used in calculating s. Hence, the t-distribution's shape adjusts to account for the additional uncertainty.
自由度ν（或df）由样本量决定。对于单样本均值，ν = n - 1。这一概念反映了用于估计标准差的独立信息数量。当我们用s代替σ时，由于计算s时用到了样本均值本身，我们损失了一个自由度。因此，t分布的形状会根据额外的不确定性进行调整。
Always verify the correct ν before looking up probabilities. Using n instead of n–1 is a common mistake that leads to narrower tails and understated probabilities.
在查概率之前务必核实正确的ν。使用n而非n–1是一个常见错误，它会使得尾部变窄、概率被低估。
3. The t-Distribution Table | t分布表
value of t such that P(T > t) = α for given ν. For instance, with ν = 10 and α = 0.05 (one‑ ibution with 10 df, the probability that the test statistic exceeds 1.812 is exactly 0.05.
考试中提供的t分布表（例如IB公式手册）给出了不同自由度下选定右侧尾部概率α（常为0.10, 0.05, 0.025, 0.01, 0.005）对应的临界值t*。有些表还显示双尾概率。每个条目是满足P(T > t) = α的t值。例如，当ν = 10且α = 0.05（单尾）时，t* = 1.812。这意味着在自由度为10的t分布下，检验统计量超过1.812的概率恰为0.05。
To find the probability of |T| in value, you would look up the absolute value in the table and read the corresponding two‑tail α. For example, if t = 2.228 and ν = 10, the two‑ail probability α = 0.05 because 2.228 is the critical value for α/2 = 0.025 in each tail.
要查找|T|超过某一值的概率，你可以在表中查找该绝对值，并读出对应的双尾α。例如，若t = 2.228且ν = 10，则双尾概率α = 0.05，因为2.228是每尾α/2 = 0.025的临界值。
4. Finding One-Tailed Probabilities | 求单尾概率
When you are asked to find the p‑value for a one‑sided test (e.g., H₁: μ > μ₀), you calculate the test statistic t = (x̄ − μ₀) / (s/√n racketed. For example, with ν = 15, if t = 2.0, the table might show 1.753 for α=0.05 (one‑tail) and 2.131 for α=0.025. Hence, 2.0 lies between these, so p‑value lies between 0.025 and 0.05. Modern GDCs can give precise p‑values.
当要求计算单侧检验（如H₁: μ > μ₀）的p值时，你需计算检验统计量t = (x̄ − μ₀) / (s/√n)，并在表上根据相应ν找出其相较于临界值的位置。确切概率可以给出范围。例如，ν = 15时，若t = 2.0，表上可能显示α=0.05（单尾）的临界值为1.753，α=0.025为2.131。因此2.0位于两者之间，故p值介于0.025与0.05之间。现代图形计算器可以给出精确p值。
Always indicate whether you are using a one‑tailed or two‑tailed approach. In IB exams, you must write the p‑value and compare it with the significance level, then make a conclusion.
始终指明你使用的是单尾还是双尾方法。在IB考试中，你必须写出p值并将其与显著性水平比较，然后得出结论。
5. Finding Two-Tailed Probabilities | 求双尾概率
For a two‑tailed test (H₁: μ ≠ μ₀), the p‑value is P(|T| ≥ |t|). Using the symmetry of the t-distribution, this equals 2 × P(T ≥ |t|). Hence, you take the absolute value of the test statistic, find the one‑ 6 for α=0.025 and 2.896 for α=0.01. Thus P(T ≥ 2.5) is between 0.01 and 0.025, and the two‑tailed p‑value is between 0.02 and 0.05.
对于双尾检验（H₁: μ ≠ μ₀），p值为P(|T| ≥ |t|)。利用t分布的对称性，这等同于2 × P(T ≥ |t|)。因此，取检验统计量的绝对值，从表中找出单尾概率（或范围），然后加倍。例如，ν = 8时，t = −2.5。绝对值为2.5。表中可能给出α=0.025对应2.306，α=0.01对应2.896。因此P(T ≥ 2.5)介于0.01与0.025之间，双尾p值介于0.02与0.05之间。
Remember that doubling a probability is only valid for symmetric, two‑tailed tests. Never double for a one‑tailed test.
请记住，概率加倍仅对对称的双尾检验有效。单尾检验绝不可加倍。
6. Calculating p-Values for t-Tests | 为t检验计算p值
can be carried out linearly on the reciprocal scale or by using software. The IB accepts either stating the range (e.g., 0.02 < p < 0.05) or using a GDC to obtain p = 0.029 for a specific t.
在假设检验中，p值是在原假设下，得到与观察值一样极端或更极端的检验统计量的概率。使用t分布表时，通常只能确定p值的范围，除非使用内插法或技术。可以在倒数尺度上进行线性内插，或借助软件。IB接受陈述范围（如0.02 < p < 0.05）或使用图形计算器得出具体p值（例如p = 0.029）。
When using a GDC (TI‑84, TI‑Nspire, Casio), the command tcdf(lower, upper, df) computes the probability between two t-values. For a right‑tail probability, use tcdf(t, 9999, df). For a two‑tailed, multiply the result by 2 if the distribution is symmetric and you have taken the absolute t.
当使用图形计算器（TI‑84, TI‑Nspire, Casio）时，命令tcdf(lower, upper, df)会计算两个t值之间的概率。对于右侧尾部概率，使用tcdf(t, 9999, df)。对于双尾，如果分布对称且你取了绝对值t，则将结果乘以2。
7. Using Technology: GDC or Software | 使用技术：图形计算器或软件
Graphing calculators are indispensable in the IB examination. On the TI‑Nspire, you can access the t-distribution probability via Menu > Statistics > Distributions > t Cdf. Enter lower bound (the t-value for an upper‑tail or -9999 for cumulative lower), upper bound, and degrees of freedom. The calculator returns the exact probability. You can
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