第6章抽样推断88821373.pptxVIP

第6章 抽样推断;6.1 抽样推断的基本概念 ;抽样推断具有如下几个特点;抽样推断的作用;6.1.1 总体和样本 ;2．样本（sample）又称子样，它是从全及总体中随机抽取出来，作为代表这一总体的那部分单位组成的集合体。 ;6.1.2 参数和统计量 ;（1）总体平均数 Population mean ;（2）总体成数 Proportion;【实例6．1】红光灯泡厂生产的10000只灯泡中，有9550只是合格品，有450只是不合格产品，则总体成数即灯泡的合格率和不合格率分别为： ;（3）总体方差和总体标准差 Variance standard deviation;2．统计量(statistic) 。根据样本各单位标志值计算的反映样本特征的指标称为统计量，也称作样本指标sample statistic 。它是用来估计总体参数的。;（1）样本平均数 Sample mean;样本是非标志的方差S2和标准差S的计算公式分别为 ;6.1.3 样本容量和样本个数 ;6.1.4 重复抽样和不重复抽样 ;6.2 抽样误差;抽样调查中误差有以下两个来源 ;Dubious Sampling: The Literary Digest Case;First, there was selection bias. The Digest mailed questionnaires to 10 million people whose names had been taken from various lists such as its own subscribers, telephone directories, and automobile-registration rolls. During the Great Depression, higher-income people typically voted Republican, and these people were well represented in the Digest’s sample. On the other hand, lower-income people, who heavily favored the Democrats, were underrepresented because a lower percentage could afford magazine subscriptions, telephones, and automobiles.;Second, there was nonresponse bias. Only 2.4 million of the l0 million questionnaires were mailed back. Although this made the survey the largest sample ever taken, more educated people are more likely to respond to mail questionnaires than less educated ones. The former, again, tended to favor Republicans; the latter, the Democrats. Thus, a much larger percentage of the nonrespondents than of the respondents were for Roosevelt. This bias reinforced the selection bias.;Note: The Digest never survived the debacle and folded shortly thereafter. At the same time, George Gallup was setting up his survey organization, and he correctly forecast the Roosevelt Victory from a mere sample of 50,000 people. Yet in 1948, using another dubious procedure (a form of judgment sampling, called quota sampling), Gallups organization (along with Crossleys and Ropers) incorrectly predicted the victory of Thomas Dewey over Harr