消除回答误差的随机化模型

同其他统计调查方法一样,抽样调查结果也会出现误差。抽样调查的误差可分为抽样误差和非抽样误差两类,其中抽样误差是由抽样的随机性引起的,是不可消除的,但我们可以对它进行估计,并控制在要求的范围内。在样本量固定的情况下,我们通过采用适当的抽样方法和估计量来最大限度地减少抽样误差,并将其控制在一定的限度内。非抽样误差是发生在回答、登记、汇总、计算等过程中的误差,我们可以通过加强统计法规建设,提高统计人员的素质来消除或减少登记、汇总、计算等过程中的误差,而回答误差由于受多种因素的影响,非常难以控制和消除。在调查项目…     

抽样方法中调查误差的控制

本文探讨如何对抽样调查中的调查误差加以硬性控制,限调查误差于最小范围内。如何提高调查人员的素质以减少调查误差不是本文的主题。调查误差与抽样误差有着辩证关系。通过抽样调查得到的样本的综合指标与总体相应的综合指标必定存在误差。其一是抽样误差,其二是调查误差。当调查误差一定时,抽样误差的大小决定全部误差的大小;当抽样误差一定时,调查误差的大小决定全部误差的大小。如要求全部误差控制在一定范围之内,实践证明,抽样误差与调查误差便存在着一个此长彼消的辩证关系。即:当样本量加大时,抽样误差减少,但因工作难度加大,人员增加,调查误差的控制难度亦就越大;样本量减少,抽样误差扩大,但调查误差易控制,可提高调查精度。两种情况使总体误差基本稳定在一个范围内。     

应重视抽样调查的调查误差

目前抽样调查的优势及取代逐级上报统计方法的必然性已经成为共识 ,人们将解决统计数据失真的全部希望寄托于抽样调查 ,甚至出现“一抽就准”的幻想。因此我们必须对抽样调查的误差有一个清醒的认识 ,打破“一抽就准”的幻想。抽样调查误差来源于两个方面 ,一个是单纯由于不同的随机样本得出不同的估计量而产生的抽样误差 ,它是抽样调查所固有的 ,可控制的误差 ;另一个是执行随机原则或调查方案发生偏差而形成的调查误差 ,这是一种人为因素造成的、模糊的、无法计算和修正的误差。现实中抽样调查数据的准确性可以说主要取决于调查误差。而我…     

关于降低城镇失业率抽样调查非抽样误差的尝试和设想

抽样调查误差来自许多不同的方面,概括起来划分成两大类:一类是抽样误差,一类是非抽样误差。所谓非抽样误差是指在调查活动过程中产生的各种各样与样本抽取无关的误差。本文主要讨论的是城镇失业率抽样调查过程中的非抽样误差问题。     

浅析统计调查中的非抽样误差

在统计理论中把总误差分为两类,一类是抽样误差,另一类是非抽样误差。现阶段的统计实践工作中非抽样误差在总误差中占很大比重,通过减少非抽样误差,从而降低总误差有着非常现实的重要意义。一、非抽样误差的产生非抽样误差是指除抽样误差以外的,由于各种原因而引起的误差,在抽样调查与全面调查中非抽样误差都可能存在。从调查开始时的历史数据的收集和调查方案的设计阶段,到调查的实施和数据收集阶段,以     

论容许误差和把握程度

抽样调查的一个重要特点,就在于可以在调查开始前,计算和控制抽样误差。怎么样计算和控制抽样误差?调查结果的容许误差应当有多大?把握程度如何?只有把这些问题回答清楚,才能真正懂得抽样调查的道理,熟练地掌握和运用抽样调查方法。     

小微企业抽样调查的样本轮换

目前,小微企业抽样调查数据受到各级政府和社会各界的高度关注。针对小微企业单位新增、消亡变动频繁的特点,研究了总体单位及样本量变动的一般条件下的样本轮换理论,对样本轮换率和估计量进行了探讨,扩大了研究结果的适用范围,得到了简单随机抽样、分层抽样中样本轮换的有关结论,对估计量的抽样误差进行了有效控制,并进行了相关实证研究,最后提出了构造适合小微企业连续性抽样调查的样本轮换设计模式和方法。     

非抽样误差及其控制方法

所谓非抽样误差,是与抽样误差——代表性误差相对应来说的,泛指在抽样调查的整个过程中,由于调查工作方面的原因而产生的各种误差。通常把非抽样误差也叫做调查误差或工作误差。     

基于农产量抽样调查误差的原因分析

现在新的农产量抽样调查只是农村多目标抽样调查中的一个组成部分,相比于过去的农产量抽样调查,虽然"专一性"弱了,但是抽样的环节却减少了,由过去的五级(省抽县--县抽乡--乡抽村--村抽户--户抽地块)减为现在的二级(省抽村--村抽地块),因此,抽样效率大大提高了.但是新方案与旧方案一样,仅对抽样误差做了要求,即抽样误差系数一般控制在±2%,把握度一般为95%.而这总可以通过抽取足够的调查单位来做到.两者均未对非抽样误差做出明确的要求和规定,也缺乏有效的控制方法.这主要是因为非抽样误差来源多种多样,产生于调查工作的各个环节,且不同的调查内容会产生不同的非抽样误差,不具有通性且较难察觉,而且缺乏成熟实用的理论工具.因此,很难硬性要求和规定,这使得调查结果的非抽样误差究竟有多大,很难讲清楚.但是,非抽样误差如果不进行有效的控制,即使抽样误差已经控制到很小的程度,整个调查结果仍有可能反映不了真实的情况.因而,有效地分析非抽样误差的来源,无论实践上还是理论上都具有极其重要的意义.     

抽样调查在第三产业统计中的应用

第三产业是一个复杂的总 体,对它进行抽样调查, 较之其它专业调查而言,是一项很艰难的工作。本文就南京市1993年第三产业抽样调查的实践,谈一谈怎样控制抽样误差和非抽样误差的问题。     

使用普查数据模拟MPPS抽样方法的研究

MPPS抽样即多变量与规模成比例的概率抽样,是20世纪90年代才提出来的一种抽样设计。近年来,中国有关部门与美国农业部国家农业署合作,进行了MPPS抽样设计的试点,来解决多目标调查问题。但是MPPS抽样在中国的应用非常有限。对MPPS抽样进行简单的回顾,介绍了它的基本估计,并对其应用进行了数据模拟研究。模拟中采用了系统抽样和泊松抽样的方法,根据实际调查数据得到了明确的结果。还对泊松抽样的一种变形永久随机数抽样的方法进行了模拟研究,并对它的一种误用情况进行了模拟比较,得到了具有说服力的结果。     

PPS抽样设计的比估计及其模拟分析

本文在介绍简单随机抽样比估计的基础上,对PPS抽样设计的比估计进行了系统的研究,并使用了第一次农业普查数据对方法进行了模拟。从模拟结果可以看到,比估计方法可以解决一些调查精度和多目标调查的问题。     

非抽样误差函数的构建——基于不完备抽样框下无回答误差的讨论

抽样调查中,使用不完备的抽样框及存在无回答单元是产生非抽样误差的两类主要原因。运用概率统计方法,分析抽样调查中使用三类不完备抽样框且存在无回答单元时对估计误差的影响,尝试构建基于不完备抽样框及无回答共存时的非抽样误差函数,为抽样调查中非抽样误差的定量分析与控制提供依据。     

Variance estimation with randomized response

Presenting a general procedure of eliciting a randomized response (RR) from selected persons in order to estimate the total of a sensitive variable related to a finite survey population, we consider two estimators along with variance estimators treating the case of sampling with probabilities proportional to (known) size measures (PPS) with replacement (WR), drawing analogies with multi-stage sampling and note their relative efficacies.     

Order Sampling Design with Prescribed Inclusion Probabilities

Order sampling with fixed distribution shape is a class of sampling schemes with inclusion probabilities approximately proportional to given size measures. In a recent article, methods were provided to compute the exact first and second order inclusion probabilities numerically when the distribution shape is of the Pareto type. In the same article, procedures were also provided for this case to adjust the parameters to get predetermined inclusion probabilities. In this paper we prove the existence and uniqueness of a solution for the latter problem, in general for any order sampling of fixed distribution shape.     

Designing optimal double-sampling plan based on process capability index

Process capability analysis is applied to monitor the process quality. Process capability can be quantified by process capability index. These indices have wide application in quality control methods and acceptance sampling plans. In this paper, we introduce a double-sampling plan based on process capability index. In this type of scheme, under a decision rule and with the specified rejection and acceptance numbers, the second sample is selected and the decision of rejection or acceptance is made based on the information obtained from two samples. The purpose of this scheme is to reduce the average sample number in order to reduce the time and cost of sampling. A comparison study has been conducted in order to evaluate the performance of proposed method in comparison with classical single sampling plans.     

Developing and designing of an efficient variables sampling system based on the process capability index

Process capability indices (PCIs) are extensively used in the manufacturing industries in order to confirm whether the manufactured products meet their specifications or not. PCIs can be used to judge the process precision, process accuracy, and the process performance. So developing of sampling plans based on PCIs is inevitable and those plans will be very much useful for maintaining and improving the product quality in the manufacturing industries. In view of this, we propose a variables sampling system based on the process capability index C_pmk, which takes into account of process yield and process loss, when the quality characteristic under study will have double specification limits. The proposed sampling system will be effective in compliance testing. The advantages of this system over the existing sampling plans are also discussed. In order to determine the optimal parameters, tables are also constructed by formulating the problem as a nonlinear programming in which the average sample number is minimized by satisfying the producer and consumer risks.     

Optimal designing of an SkSP-R double sampling plan

In this article, we propose a method of planning and determining the optimum parameters of a SkSP-R skip-lot sampling plan by using the attribute double sampling plan as the reference plan. The SkSP-R plan is a new type of skip-lot sampling plan which has a provision for re-inspecting the submitted lots. The optimal plan parameters of the suggested sampling plan are estimated with the target that the average sample number be minimized and satisfying both the specified producer's as well as the consumer's risks simultaneously. In order to obtain the optimum parameters, tables are also built for different combinations of the acceptable quality level and the limiting quality level in conjunctions with different producer's and consumer's risks. An illustrative example is provided for the implementation of the suggested plan. The advantages of the suggested plan over the existing conventional sampling plans and other existing skip-lot sampling plans are also described.     

A mixed double sampling plan based on Cpk

Abstract

Acceptance sampling plans are quality tools for the manufacturer and the customer. The ultimate result of reduction of nonconforming items will increase the profit of the manufacturer and enhance the satisfaction of the consumer. In this article, a mixed double sampling plan is proposed in which the attribute double sampling inspection is used in the first stage and a variables sampling plan based on the process capability index C_pk is used in the second stage. The optimal parameters are determined so that the producer’s and the consumer’s risks are to be satisfied with minimum average sample number. The optimal parameters of the proposed plan are estimated using different plan settings using two points on the operating characteristic curve approach. In designing the proposed mixed double sampling plan, we consider the symmetric and asymmetric nonconforming cases under variables inspection. The efficiency of the proposed plan is discussed and compared with the existing sampling plans. Tables are constructed for easy selection of the optimal plan parameters and an industrial example is also included for implementation of the proposed plan.     

Randomized response techniques: An application to the Botswana AIDS impact survey

The performance of the sampling strategy used in the Botswana Aids Impact Survey II (BAISII) has been studied in detail under a randomized response technique. We have shown that alternative strategies based on the Rao–Harley–Cochran (RHC) sampling scheme for the selection of first stage units perform much better than other strategies. In particular, the combination RHC for the selection of first stage units (fsu's) and systematic sampling for the selection of second stage units (ssu's) perform the best when the sample size is small where as the RHC and SRSWOR perform the best when the sample size is large. In view of the present findings it is recommended that the BAISII survey should be studied in more detail incorporating more indicators and increased sample sizes. This is because the BAISII survey design is extensively in use for large scales surveys in Southern African countries.