Lecture 14 — Sampling and Sampling Distributions

第14讲——抽样与抽样分布


Slide 1 — Sampling and Sampling Distributions

第1页——抽样与抽样分布

Knowledge Points (知识点)

  1. Sampling(抽样)
  2. Sampling distribution(抽样分布)
  3. Key measures: , (关键统计量)

Knowledge Point 1 — Sampling(抽样)

  • Explanation(解释)
    Sampling is the process of selecting a subset of a population to estimate population characteristics.
    抽样是从总体中选择部分个体,以推断总体特征的统计过程。

  • Example(例子)
    A researcher surveys 100 WKU students instead of all students to estimate average weekly spending.
    研究者通过调查100名温州肯恩大学学生,而非全体学生,以估计平均每周支出。

  • Extension(拓展)
    Sampling reduces cost and time, and when properly designed, yields accurate estimates of population parameters.
    抽样可以节约成本和时间,只要设计合理,也能提供对总体参数的准确估计。

  • Summary(总结)
    Sampling allows inference about the whole population without a complete census.
    抽样能在不进行全面调查的情况下推断总体特征。


Knowledge Point 2 — Sampling Distribution(抽样分布)

  • Explanation(解释)
    A sampling distribution shows how a sample statistic varies when repeated samples are drawn from the same population.
    抽样分布描述了在重复抽样时样本统计量的变动规律。

  • Example(例子)
    Drawing multiple samples of 50 students each and calculating their mean GPA produces a distribution of sample means — the sampling distribution of .
    若多次抽取50名学生并计算平均绩点,这些均值的分布即为样本均值 的抽样分布。

  • Extension(拓展)
    The concept of sampling distribution is the foundation for understanding probability-based inference such as confidence intervals and hypothesis testing.
    抽样分布概念是理解基于概率的统计推断(如置信区间与假设检验)的基础。

  • Summary(总结)
    Sampling distribution connects sample variability with population characteristics through probability.
    抽样分布通过概率理论将样本变异与总体特征联系起来。


Knowledge Point 3 — Key Measures: and (关键统计量)

  • Explanation(解释)
    The sample mean estimates population mean ; the sample proportion estimates population proportion .
    样本均值 用于估计总体均值 ,样本比例 用于估计总体比例

  • Example(例子)
    If 60 of 100 surveyed students use mobile payments, then , an estimate for .
    若100名学生中有60人使用移动支付,则 ,为总体比例 的估计。

  • Extension(拓展)
    These sample statistics serve as point estimators that summarize information about the population.
    这些样本统计量作为点估计量,用于概括总体信息。

  • Summary(总结)
    and are central statistics for inference about means and proportions.
    是均值与比例推断中的核心统计量。


Slide 2 — Selecting a Sample (Finite Population)

第2页——抽样选择(有限总体)

Knowledge Points (知识点)

  1. Finite population(有限总体)
  2. Equal chance selection(等概率抽样)
  3. Randomness and representativeness(随机性与代表性)

Knowledge Point 1 — Finite Population(有限总体)

  • Explanation(解释)
    A finite population has a fixed and countable number of elements, such as 900 applicants or 500 customers.
    有限总体由固定数量的成员构成,如900名申请者或500位顾客。

  • Example(例子)
    St. Andrew’s College has applications, and a sample of is randomly selected for review.
    圣安德鲁学院共有 份申请,随机选取 名申请者进行评估。

  • Extension(拓展)
    Sampling from a finite population allows for exact probability assignment and control over bias.
    从有限总体抽样能精确控制概率与偏差,结果更可重复。

  • Summary(总结)
    Finite populations provide a known frame for random selection, ensuring every member has a measurable chance.
    有限总体提供明确抽样框,使每个成员都有可测的选中概率。


Knowledge Point 2 — Equal Chance Selection(等概率抽样)

  • Explanation(解释)
    Every element in the population has an equal probability of being chosen.
    总体中的每个成员被选中的概率相同。

  • Example(例子)
    Assign numbers 1–900 to applicants and use a random number generator to select 30 IDs.
    将900位申请者编号为1至900,并使用随机数生成器抽取30人。

  • Extension(拓展)
    Equal chance sampling ensures fairness and prevents selection bias.
    等概率抽样可保证抽样公平,避免选择偏差。

  • Summary(总结)
    Equal probability is the foundation of valid random sampling.
    等概率原则是随机抽样的基本条件。


Knowledge Point 3 — Randomness and Representativeness(随机性与代表性)

  • Explanation(解释)
    A representative sample accurately reflects population characteristics; randomness helps achieve that.
    具有代表性的样本能准确反映总体特征,而随机性能帮助实现这一目标。

  • Example(例子)
    Selecting 30 students randomly from different majors avoids concentration bias.
    从不同专业随机选取30名学生可避免样本集中偏差。

  • Extension(拓展)
    Representativeness depends on both random selection and sufficient sample size.
    样本代表性取决于随机抽样与样本量的充足。

  • Summary(总结)
    A random and representative sample supports accurate statistical inference.
    随机且具代表性的样本能提供更准确的推断。


Slide 3 — Selecting a Sample (Infinite Population)

第3页——抽样选择(无限总体)

Knowledge Points (知识点)

  1. Infinite population(无限总体)
  2. Random sampling from infinite population(无限总体的随机抽样)
  3. Independence of observations(观测独立性)

Knowledge Point 1 — Infinite Population(无限总体)

  • Explanation(解释)
    An infinite population occurs when new elements continuously appear, e.g., production output or customer arrivals.
    无限总体指不断生成新数据的总体,如生产线输出或顾客到达。

  • Example(例子)
    A company monitoring daily website visits treats each new visit as part of an infinite population.
    公司监测每日网站访问量时,将每次访问视为无限总体的一部分。

  • Extension(拓展)
    Infinite populations are modeled probabilistically since their total number is unobservable.
    无限总体无法穷尽,只能通过概率模型进行刻画。

  • Summary(总结)
    Infinite populations describe processes rather than finite groups.
    无限总体用于描述持续过程而非固定群体。


Knowledge Point 2 — Random Sampling from Infinite Population(无限总体随机抽样)

  • Explanation(解释)
    Sampling from an infinite population requires independent selection of each observation.
    从无限总体抽样时,每个观测值必须独立选择。

  • Example(例子)
    Selecting every 100th customer transaction in a continuous system maintains randomness.
    在连续交易系统中,每隔100笔抽样一次能保持随机性。

  • Extension(拓展)
    Independence ensures unbiased estimation and valid inference from streaming data.
    保证样本独立能确保无偏估计与有效推断。

  • Summary(总结)
    Independent random sampling is essential when dealing with infinite populations.
    独立随机抽样是无限总体分析的必要条件。


Knowledge Point 3 — Independence of Observations(观测独立性)

  • Explanation(解释)
    Independence means the selection of one element does not affect another’s chance.
    独立性指一个样本被选中不会影响另一个样本的选中概率。

  • Example(例子)
    When drawing data from automated sensors, each reading is independent of previous ones.
    从自动传感器收集的数据中,每次测量结果相互独立。

  • Extension(拓展)
    Violating independence leads to biased estimates and underestimated variability.
    若独立性被破坏,将导致估计偏差与方差低估。

  • Summary(总结)
    Independence preserves randomness and validity in sampling results.
    样本独立性保证随机性与结果的有效性。


Slide 4 — Point Estimation

第4页——点估计

Knowledge Points (知识点)

  1. Point estimation(点估计)
  2. Sample statistic vs population parameter(样本统计量与总体参数)
  3. Unbiasedness and efficiency(无偏性与有效性)

Knowledge Point 1 — Point Estimation(点估计)

  • Explanation(解释)
    Point estimation uses sample statistics to estimate population parameters.
    点估计是利用样本统计量来估计总体参数的过程。

  • Example(例子)
    Use to estimate , or to estimate .
    以样本均值 估计总体均值 ,以样本比例 估计总体比例

  • Extension(拓展)
    Good estimators should be unbiased (), consistent, and efficient.
    优秀的估计量应具备无偏性(一致性)与高效率。

  • Summary(总结)
    Point estimation provides a single best guess of population parameters.
    点估计为总体参数提供最佳单点估计。


Knowledge Point 2 — Sample Statistic vs Population Parameter(样本统计量与总体参数)

Sample Statistic(样本统计量)Population Parameter(总体参数)
  • Explanation(解释)
    Sample statistics are used as estimators for corresponding population parameters.
    样本统计量是相应总体参数的估计量。

  • Example(例子)
    If the sample variance is , it estimates population variance .
    若样本方差为 ,则它用于估计总体方差

  • Extension(拓展)
    These relationships are key to connecting data to theoretical parameters.
    这些对应关系是从样本到总体推断的基础。

  • Summary(总结)
    Each sample statistic provides information about a specific population characteristic.
    每个样本统计量反映总体的某一特征。


Knowledge Point 3 — Unbiasedness and Efficiency(无偏性与有效性)

  • Explanation(解释)
    An estimator is unbiased if its expected value equals the true parameter.
    若估计量的期望值等于总体真实值,则称其为无偏。

  • Example(例子)
    Since , the sample mean is an unbiased estimator of the population mean.
    ,故样本均值是总体均值的无偏估计。

  • Extension(拓展)
    Efficiency compares the variances of unbiased estimators; smaller variance means higher efficiency.
    有效性比较无偏估计量的方差,方差越小效率越高。

  • Summary(总结)
    Unbiased and efficient estimators yield reliable and precise statistical inferences.
    无偏且有效的估计量能提供更精确的推断。


Slide 5 — Comments on Sampling

第5页——关于抽样的说明

Knowledge Points (知识点)

  1. Population vs Sample(总体与样本)
  2. Representativeness(样本代表性)
  3. Sampling bias(抽样偏差)

Knowledge Point 1 — Population vs Sample(总体与样本)

  • Explanation(解释)
    The population is the complete set of elements we want to study, while the sample is a smaller group taken from it.
    总体是研究目标的全部成员,而样本是从总体中抽取的一部分。

  • Example(例子)
    A company wants to know employee satisfaction across all 2,000 workers, but surveys only 200 randomly chosen employees.
    某公司想了解2000名员工的满意度,仅随机调查200人作为样本。

  • Extension(拓展)
    The closer the sample characteristics are to the population, the more valid the statistical inference.
    样本特征越接近总体特征,统计推断结果越可靠。

  • Summary(总结)
    Sampling provides manageable data while maintaining a link to the population’s properties.
    抽样使数据更可管理,同时保持总体特征的代表性。


Knowledge Point 2 — Representativeness(样本代表性)

  • Explanation(解释)
    Representativeness means the sample accurately mirrors the diversity and proportions of the population.
    代表性意味着样本准确反映总体的多样性与比例。

  • Example(例子)
    If 60% of WKU students are female, a representative sample should also have roughly 60% females.
    若温肯大学学生中女性占60%,则具有代表性的样本中女性比例也应接近60%。

  • Extension(拓展)
    Random sampling, stratified sampling, and sufficient sample size increase representativeness.
    随机抽样、分层抽样及足够的样本量能提高样本代表性。

  • Summary(总结)
    Representative samples ensure that sample statistics reflect true population parameters.
    代表性样本能保证样本统计量准确反映总体参数。


Knowledge Point 3 — Sampling Bias(抽样偏差)

  • Explanation(解释)
    Sampling bias occurs when the method of selection systematically favors certain outcomes.
    抽样偏差指抽样方法系统性地偏向某一结果。

  • Example(例子)
    Conducting an online survey about internet use may exclude individuals without online access.
    通过网络问卷调查上网习惯会排除没有网络的人群,从而产生偏差。

  • Extension(拓展)
    Bias can be reduced through randomization and strict adherence to sampling procedures.
    通过随机化与规范抽样程序可减少偏差。

  • Summary(总结)
    Avoiding sampling bias is crucial for reliable and generalizable conclusions.
    避免抽样偏差是获得可靠、可推广结论的关键。


Slide 6 — Sampling Distribution of

第6页——样本均值的抽样分布

Knowledge Points (知识点)

  1. Sampling distribution of (样本均值的分布)
  2. Expected value (样本均值的期望)
  3. Unbiased estimator(无偏估计量)

Knowledge Point 1 — Sampling Distribution of (样本均值的分布)

  • Explanation(解释)
    The sampling distribution of shows the probability distribution of all possible sample means from samples of size .
    样本均值 的抽样分布描述了所有样本量为 的样本均值的概率分布。

  • Example(例子)
    Suppose we take 100 different samples of 50 students each, compute each mean GPA, and plot their distribution—it forms the sampling distribution of .
    假设抽取100个样本(每个样本包含50名学生),计算各样本平均绩点,其分布即为 的抽样分布。

  • Extension(拓展)
    The sampling distribution allows the application of probability to describe the variability of statistics.
    抽样分布使我们能够用概率刻画统计量的变动性。

  • Summary(总结)
    Sampling distributions bridge descriptive statistics and inferential analysis.
    抽样分布连接了描述统计与推断统计。


Knowledge Point 2 — Expected Value (样本均值的期望)

  • Explanation(解释)
    The expected value of equals the true population mean :

    样本均值的期望等于总体均值 ,即

  • Example(例子)
    If the population mean height of students is cm, the average of sample means across repeated sampling will also be 170 cm.
    若总体平均身高 cm,多次抽样计算样本均值的平均值也为170 cm。

  • Extension(拓展)
    This property shows that the sample mean is centered around , ensuring unbiased estimation.
    这一特征表明样本均值 以总体均值为中心,是无偏的。

  • Summary(总结)
    The equality is the mathematical foundation of unbiased estimation.
    是无偏估计的重要数学基础。


Knowledge Point 3 — Unbiased Estimator(无偏估计量)

  • Explanation(解释)
    An estimator is unbiased if its expected value equals the true parameter being estimated.
    若估计量的期望值等于所估参数的真实值,则称其为无偏估计量。

  • Example(例子)
    Since , is an unbiased estimator of .
    ,所以 的无偏估计量。

  • Extension(拓展)
    Other unbiased estimators include sample proportion for population , and sample variance for .
    其他无偏估计量还包括:样本比例 估计总体比例 ,样本方差 估计总体方差

  • Summary(总结)
    Unbiasedness ensures estimators neither overestimate nor underestimate true values on average.
    无偏性保证估计量在长期平均下既不高估也不低估总体参数。


Slide 7 — Making Statistical Inference

第7页——统计推断的过程

Knowledge Points (知识点)

  1. Relationship between and (样本均值与总体均值的关系)
  2. Statistical inference steps(统计推断步骤)
  3. Estimation and decision making(估计与决策)

Knowledge Point 1 — Relationship between and (样本均值与总体均值)

  • Explanation(解释)
    The sample mean serves as a point estimator for the population mean .
    样本均值 是总体均值 的点估计量。

  • Example(例子)
    If is obtained from a random sample of 50 students, we infer .
    若从50名学生的样本中得到 ,则可推测总体均值

  • Extension(拓展)
    The reliability of as an estimator depends on sample size and sampling variability.
    作为估计量的可靠性取决于样本量 与抽样变异程度。

  • Summary(总结)
    connects observed data to the theoretical population mean .
    样本均值 将观测数据与总体参数 联系起来。


Knowledge Point 2 — Statistical Inference Steps(统计推断步骤)

  • Explanation(解释)
    Statistical inference converts sample information into knowledge about the population.
    统计推断是将样本信息转化为总体结论的过程。

  • Example(例子)
    Steps:
    1️⃣ Select a random sample
    2️⃣ Compute
    3️⃣ Use to estimate or test hypotheses about .
    推断过程包括:① 随机抽样;② 计算样本均值;③ 用 推断或检验

  • Extension(拓展)
    These steps form the foundation for hypothesis testing and confidence interval estimation.
    该过程为假设检验与置信区间估计奠定基础。

  • Summary(总结)
    Statistical inference generalizes sample findings to population conclusions.
    统计推断使样本结论能推广至总体。


Knowledge Point 3 — Estimation and Decision Making(估计与决策)

  • Explanation(解释)
    Once estimates , managers can use this estimate for strategic or policy decisions.
    用于估计 后,管理者可据此进行战略或政策决策。

  • Example(例子)
    A firm estimates average customer spending \bar{x} = \85\bar{x} = $85$,据此调整定价与库存策略。

  • Extension(拓展)
    Inferential results must be interpreted cautiously, considering sampling error and confidence level.
    推断结果应结合抽样误差与置信水平谨慎解释。

  • Summary(总结)
    Estimation links data analysis to actionable business and policy insights.
    统计估计将数据分析转化为可操作的商业与政策见解。


Slide 8 — Sampling Distribution of (Standard Error)

第8页——样本均值的抽样分布(标准误差)

Knowledge Points (知识点)

  1. Definition of standard error(标准误差定义)
  2. Finite population correction factor(有限总体修正系数)
  3. Infinite population approximation(无限总体近似)

Knowledge Point 1 — Definition of Standard Error(标准误差定义)

  • Explanation(解释)
    The standard deviation of the sampling distribution of , denoted , is called the standard error of the mean.
    样本均值抽样分布的标准差称为均值的标准误差,记作

  • Example(例子)
    For a population with and sample size ,

  • Extension(拓展)
    The standard error measures how much sample means vary around the population mean.
    标准误差表示样本均值围绕总体均值的波动程度,反映估计精度。

  • Summary(总结)
    Smaller means higher precision in estimating .
    越小,估计总体均值 的精度越高。


Knowledge Point 2 — Finite Population Correction Factor(有限总体修正系数)

  • Explanation(解释)
    For a finite population of size , the standard error is adjusted using the finite population correction factor (FPC):

    其中 即为有限总体修正系数。

  • Example(例子)
    If , , and , then

  • Extension(拓展)
    When , the FPC must be applied to avoid overestimating variability.
    当抽样比例 时,必须使用修正系数以防高估变异性。

  • Summary(总结)
    FPC corrects the reduction of variability caused by sampling without replacement.
    修正系数用于校正无放回抽样导致的样本依赖问题。


Knowledge Point 3 — Infinite Population Approximation(无限总体近似)

  • Explanation(解释)
    For large or infinite populations,

    applies, since the effect of FPC becomes negligible.

  • Example(例子)
    When , we can treat the population as infinite.
    ,则可将总体视为无限总体。

  • Extension(拓展)
    This assumption simplifies calculations while maintaining high accuracy for large .
    对于较大总体,此假设能简化计算且精度高。

  • Summary(总结)
    Infinite population assumption is reasonable when sample fraction is small.
    当抽样比例极小,视总体为无限是合理近似。


Slide 9 — Normal Distribution of (Central Limit Theorem)

第9页——样本均值的正态分布(中心极限定理)

Knowledge Points (知识点)

  1. Sampling distribution of (样本均值的分布)
  2. Rule of sample size(样本量规则)
  3. Central Limit Theorem(中心极限定理)

Knowledge Point 1 — Sampling Distribution of (样本均值的分布)

  • Explanation(解释)
    If the population follows a normal distribution, then is normally distributed for any .
    若总体服从正态分布,则样本均值 在任意样本量下也服从正态分布。

  • Example(例子)
    If student IQ scores follow , then sample mean also follows a normal distribution with the same mean.
    若学生智商服从 ,则样本均值 亦服从均值为100的正态分布。

  • Extension(拓展)
    This property simplifies inferential analysis for populations known to be normal.
    若总体正态,则简化推断计算,可直接应用标准正态法则。

  • Summary(总结)
    retains normality when the population itself is normally distributed.
    若总体正态,样本均值也保持正态。


Knowledge Point 2 — Rule of Sample Size(样本量规则)

  • Explanation(解释)
    For non-normal populations, the sampling distribution of approaches normal when sample size is large ().
    若总体非正态,当样本量较大()时,样本均值分布近似正态。

  • Example(例子)
    In skewed income data, sample means of or approximate a normal curve.
    对偏态收入数据,样本量50或100时,其样本均值分布近似正态。

  • Extension(拓展)
    For highly skewed data or extreme outliers, a larger sample () is required.
    若总体严重偏态或存在极端值,则应采用 的样本量。

  • Summary(总结)
    Larger samples yield more symmetric and normal-like sampling distributions.
    样本量越大,样本均值分布越接近正态。


Knowledge Point 3 — Central Limit Theorem(中心极限定理)

  • Explanation(解释)
    The Central Limit Theorem (CLT) states that as increases, the sampling distribution of approaches a normal distribution regardless of the population’s shape.
    中心极限定理指出:当样本量 增大时,无论总体分布形态如何,样本均值分布都会趋近于正态。

  • Example(例子)
    Even if population sales data are skewed, sample means of are nearly normal.
    即使销售数据偏态,当样本量为40时,样本均值分布也接近正态。

  • Extension(拓展)
    CLT justifies using normal models in estimation and hypothesis testing.
    中心极限定理为估计与假设检验中使用正态模型提供理论依据。

  • Summary(总结)
    CLT is the foundation for inferential statistics based on large samples.
    中心极限定理是大样本统计推断的理论基础。


Slide 10 — Example: St Andrew’s College (SAT Distribution)

第10页——案例:圣安德鲁学院(SAT分数抽样分布)

Knowledge Points (知识点)

  1. Sampling distribution of (样本均值分布)
  2. Standard error computation(标准误差计算)
  3. Probability estimation(概率估计)

Knowledge Point 1 — Sampling Distribution of (样本均值分布)

  • Explanation(解释)
    For SAT scores with , , and , the mean of the sampling distribution is .
    当 SAT 分数总体均值 、标准差 、样本量 时,样本均值分布的期望为

  • Example(例子)

    均值标准误差为 14.6。

  • Extension(拓展)
    The narrower the sampling distribution, the higher the precision of .
    抽样分布越窄,样本均值估计越精确。

  • Summary(总结)
    follows a normal distribution centered at , with .
    样本均值服从均值为1090、标准误差为14.6的正态分布。


Knowledge Point 2 — Probability Estimation(概率估计)

  • Explanation(解释)
    We want for a sample of 30 students.
    求样本均值与总体均值差不超过10的概率。

  • Example(例子)


    概率约为 0.5066,即约 50.66%。

  • Extension(拓展)
    About half of all 30-student samples will have mean SAT scores within ±10 of 1090.
    约50%的30人样本均值会落在总体均值1090的±10范围内。

  • Summary(总结)
    The example illustrates using the -score and normal probability to assess sampling accuracy.
    此案例展示了如何用标准正态分布求样本均值精度概率。


Knowledge Point 3 — Statistical Interpretation(统计解释)

  • Explanation(解释)
    A higher sample size reduces , increasing the probability that is close to .
    样本量越大,标准误差越小,样本均值越可能接近总体均值。

  • Example(例子)
    If doubles to 60, then , and accuracy improves.
    若样本量增至60,标准误差变为10.3,精确度提升。

  • Extension(拓展)
    This relationship shows why large samples lead to more stable statistical inference.
    该关系解释了为何大样本能带来更稳定的统计推断。

  • Summary(总结)
    Increasing enhances confidence that approximates .
    样本量越大,对 接近 的置信程度越高。


Slide 12 — Example: Effect of Sample Size on Sampling Distribution

第12页——样本量对抽样分布的影响

Knowledge Points (知识点)

  1. Relationship between sample size and standard error(样本量与标准误差的关系)
  2. Probability comparison ( vs. )(不同样本量下的概率比较)
  3. Interpretation of variability(变异性的解释)

Knowledge Point 1 — Relationship Between Sample Size and Standard Error(样本量与标准误差的关系)

  • Explanation(解释)
    As the sample size increases, the standard error decreases, making the sampling distribution narrower.
    随着样本量 的增加,标准误差 会减小,从而使样本均值的分布更集中、更窄。

  • Example(例子)
    When , ; when , .
    样本量为30时,标准误差为14.6;样本量为100时,标准误差为8。

  • Extension(拓展)
    This shows that larger samples provide more consistent estimates of , reducing random fluctuations.
    这说明样本量越大,对总体均值 的估计越稳定,随机波动越小。

  • Summary(总结)
    Increasing sample size leads to higher precision in estimating population parameters.
    增加样本量能显著提高总体参数估计的精确度。


Knowledge Point 2 — Probability Comparison: vs. (样本量下的概率比较)

  • Explanation(解释)
    We compare the probability that is within ±10 of for two sample sizes.
    比较在样本量不同(30与100)时,样本均值距总体均值 ±10 的概率。

  • Example(例子)
    For :

    For :

    时,概率为 0.5066;当 时,概率提升至 0.7887。

  • Extension(拓展)
    As increases, becomes more concentrated around , raising the probability of accurate estimation.
    随着样本量增加,样本均值更集中于总体均值周围,估计的准确概率上升。

  • Summary(总结)
    Larger sample size improves precision and confidence in estimation.
    样本量越大,估计越精确,对结果的置信度越高。


Knowledge Point 3 — Interpretation of Variability(变异性的解释)

  • Explanation(解释)
    The blue curve (n = 30) shows greater spread, while the yellow curve (n = 100) is narrower, meaning less variability.
    蓝色曲线(n=30)分布更宽,黄色曲线(n=100)更窄,表示样本均值的变异性降低。

  • Example(例子)
    With , values are more dispersed; with , values cluster near .
    当标准误差为14.6时,样本均值分布较散;当标准误差降至8时,样本均值更集中在1090附近。

  • Extension(拓展)
    Reduced variability increases reliability of conclusions drawn from the sample.
    较小的变异性提高了基于样本得出的结论的可靠性。

  • Summary(总结)
    Smaller means less uncertainty and higher confidence in results.
    标准误差越小,不确定性越低,对结果的置信度越高。


Slide 13 — Relationship Between Sample Size and Sampling Distribution

第13页——样本量与抽样分布的关系总结

Knowledge Points (知识点)

  1. Sample size effect(样本量效应)
  2. Variability of (样本均值的变异性)
  3. Practical implication(实际应用意义)

Knowledge Point 1 — Sample Size Effect(样本量效应)

  • Explanation(解释)
    Selecting a larger sample (e.g., 100 applicants) yields a smaller standard error than a smaller one (e.g., 30 applicants).
    较大的样本(如100人)相比于较小样本(如30人)会产生更小的标准误差。

  • Example(例子)

    数值表明样本量越大,标准误差显著下降。

  • Extension(拓展)
    With reduced standard error, becomes a more stable and accurate estimator of .
    标准误差下降意味着样本均值对总体均值的估计更稳定、更准确。

  • Summary(总结)
    Larger → Smaller → More reliable inference.
    样本量越大 → 标准误差越小 → 推断越可靠。


Knowledge Point 2 — Variability of (样本均值的变异性)

  • Explanation(解释)
    Larger samples reduce variability in and make it more likely to approximate .
    较大的样本量能降低样本均值的变异,使其更接近总体均值。

  • Example(例子)
    For , has less spread around than for .
    样本量为100时, 围绕1090的分布更集中。

  • Extension(拓展)
    Reduced variability enhances the accuracy of population estimates and strengthens confidence intervals.
    变异性降低提高了总体估计的准确度,并使置信区间更窄。

  • Summary(总结)
    Less variability implies greater consistency and stronger predictive power.
    变异性越小,估计越一致,预测力越强。


Knowledge Point 3 — Practical Implication(实际应用意义)

  • Explanation(解释)
    Researchers must balance accuracy and cost when determining sample size.
    研究者在确定样本量时需平衡精度与成本。

  • Example(例子)
    A business survey may prefer for reliable results, though it costs more than .
    商业调查中,尽管样本量100成本更高,但能提供更可靠的结果。

  • Extension(拓展)
    In practice, choose large enough to minimize error within resource constraints.
    实务中应在资源允许范围内尽量增大样本量,以降低误差。

  • Summary(总结)
    Larger samples yield better inference, but efficiency and feasibility must be considered.
    样本越大推断越准,但需兼顾效率与可行性。