📚 Finding t-scores given probabilities | 给定概率求t分数
In many statistical problems, especially those involving small samples where the population standard deviation is unknown, we need to find the t-score that corresponds to a given probability (or tail area). This process is the reverse of calculating a p-value: instead of starting with a test statistic and finding the probability, we start with a significance level or confidence level and determine the critical t-value.
在许多统计问题中，特别是涉及小样本且总体标准差未知时，我们需要找到与给定概率（或尾部面积）对应的t分数。这个过程是计算p值的逆过程：我们从显著性水平或置信水平出发，确定临界t值，而不是从检验统计量出发求概率。
1. Introduction to t-scores and probabilities | t分数与概率简介
The t-distribution arises when we estimate the population mean using the sample mean and the sample standard deviation s. Because s varies from sample to sample, the sampling distribution of the standardized sample mean follows a t-distribution with n−1 degrees of freedom (df). Given a probability such as 0.05 in the right tail, we need to find the t-score that cuts off that area.
当我们用样本均值和样本标准差s估计总体均值时，就产生了t分布。由于s因样本而异，标准化样本均值的抽样分布服从自由度为n−1的t分布。给定一个概率（例如右尾0.05），我们需要找到截断该面积的t分数。
The notation tₐ,df is often used, where a is the tail probability and df is the degrees of freedom. For instance, t₀.₀₅,₁₀ represents the t-score with 10 degrees of freedom that has 5% of the area to its right.
常用记号tₐ,df，其中a是尾部概率，df是自由度。例如，t₀.₀₅,₁₀表示自由度为10、右侧尾部面积为5%的t分数。
2. Understanding the t-distribution | 理解t分布
The t-distribution is symmetric and bell-shaped like the standard normal distribution but has heavier tails. As the degrees of freedom increase, the t-distribution approaches the standard normal distribution. For infinite df, the t-distribution is exactly N(0,1). This property helps us understand why t-scores change with sample size: smaller df lead to larger critical values for the same tail probability.
t分布与标准正态分布一样对称且呈钟形，但尾部更厚。随着自由度增加，t分布趋近于标准正态分布。当自由度为无穷大时，t分布就是标准正态N(0,1)。这一性质有助于理解为什么t分数随样本量变化：对于相同的尾部概率，自由度越小，临界值越大。
3. Degrees of freedom (df) | 自由度
n − 1. For two independent samples, the df can be estimated using the Welch-Satterthwaite formula, but in IB examinations we typically work with the simpler one-sample case. Always determine the correct df before looking up a t-score – an incorrect df will yield a wrong critical value.
对于单样本t过程，df = n − 1。对于两个独立样本，自由度可使用Welch-Satterthwaite公式估计，但在IB考试中，我们通常处理较简单的单样本情形。在查找t分数之前，一定要确定正确的自由度——错误的df会导致错误的临界值。
y probability statement about the t-statistic, such as P(T > c) = 0.01, requires t₀.₀₁,₁₉.
示例：若抽取容量为20的样本，df = 19。任何关于t统计量的概率陈述，如P(T > c) = 0.01，都需要t₀.₀₁,₁₉。
4. One-tailed vs two-tailed probabilities | 单尾与双尾概率
sh between one-tailed and two-tailed situations. A confidence interval is two-tailed: a 95% obability 0.05 in the right tail). The t-score you find depends on the tail area you need.
在处理概率时，区分单尾和双尾情况至关重要。置信区间是双尾的：95%置信水平意味着α = 0.05，平均分配到两个尾部，即每尾α/2 = 0.025。单尾假设检验将整个α放在一个尾部（例如，α = 0.05的右尾检验意味着右尾概率为0.05）。所需的t分数取决于你所用的尾部面积。
For a two-tailed test with significance level α, the critical t-scores are ±tα/2, df. For a one-tailed test, the critical value is tα, df (right-tailed) or −tα, df (left-tailed).
对于显著性水平为α的双尾检验，临界t分数为±tα/2, df。对于单尾检验，临界值为tα, df（右尾）或−tα, df（左尾）。
5. Using t-tables in reverse | 反向使用t分布表
10 and one-tail probability 0.05, the table might give t = 1.812. If your table shows two‑tail probabilities, you must match the appropriate column: for a two‑ e t-score of 2.228 for df=10 (because 0.05 in two tails corresponds to 0.025 in one tail).
传统的t分布表列出了一系列尾部概率和自由度对应的临界值。要查找t分数，先找到你所需自由度的行，然后横向读出匹配所需尾部概率的列。例如，df=10且单尾概率为0.05时，表上可能给出t=1.812。如果你的表格显示的是双尾概率，则需匹配正确的列：对于双尾概率0.05，df=10时同一表格的该列将给出t=2.228（因为双尾0.05对应于单尾0.025）。
Always double-check whether the table provides one-tail or two-tail probabilities. IB formula booklets often include both types; read the column headings carefully.
务必再次确认表格提供的是单尾还是双尾概率。IB公式手册常包含两种类型；仔细阅读列标题。
6. Using technology (GDC) | 使用图形计算器
Graphical display calculators (GDC) such as TI-Nspire, TI-84, or Casio models have an inverse
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