Pesquisa Nacional por Amostra de Domicílios 2002 - Brazil

Abstract
---------------------------

The National Household Survey - PNAD investigates annually and permanently, general characteristics of the population, education, labor, income and housing, and others with varying regularity, according to the information needs for the country. Such characteristics include migration, fertility , marriage, health, food security, among other topics. The survey of these statistics is an important instrument for the formulation, validation and evaluation of policies to socio-economic development and the improvement of living conditions in Brazil.

Geographic coverage
---------------------------

National

Analysis unit
---------------------------

- Households
- Individuals

Kind of data
---------------------------

Sample survey data [ssd]

Sampling procedure
---------------------------

The survey is conducted by a random sample of households. The information is provided by person resident or non-resident, considered capable of providing information for the whole neighborhood and the home. The interviewer is instructed not to accept a person under 14 years of age as an informant. The sampling plan uses cluster sampling, self-weighted in three stages (respectively municipalities, census tracts and households) with geographical stratification of the units of the first stage set for each state. The large municipalities in terms of population and those belonging to the metropolitan areas were each treated as a stratum and therefore included in the sample with certainty, being called autorrepresentativos. The other municipalities within the same geographic microregion were grouped into strata of approximately equal size, and designated non autorrepresentativos. Strata in these municipalities were selected systematically with probability proportional to size (ppt).
Sectors are the unit of selection in the second stage and also are selected systematically and ppt, in which case the size is measured by the number of households. The sectors were stratified according to the situation of urban and rural states of the northern region, except for Tocantins, to allow comparison of indicators from PNADs after 2004 with those performed before insertion of the rural area of the northern states. In other regions this stratification is only implicit, ie, there is an ordering for the situation of the sector before the systematic selection. Municipalities and selected sectors are kept in the sample until they are available new Census data, when they are selected new units for the sample.
Each year, in each sector selected for the sample is prepared (or updated) in the field a listing of households, producing an updated register for selection. An important characteristic of this listing operation refers to the Register of New Buildings, which is prepared to contain the buildings account for large changes in the sizes of sectors. The inventory of new construction is done in the municipalities of the sample, both in the sectors selected for the sample as those not selected. An area of new construction is excluded from the area of the original sector and is dealt with separately at the time of selection of households in this case is performed according to the sample fraction of the area. Households, which are units of the third selection stage, are formed by private households and the housing units in collective households occupied during the listing operation. The initial number of households per sector in the sample was set at 16.
The sampling fraction indicates the proportion of the population constituting the sample. Currently fractions ranging from 1/50 (rural area of Roraima) to 1/800 (Sao Paulo). How the selection of households in each selected sector for the sample is done systematically to ensure self-weighting sample, the selection range of households remains fixed from year to year. This procedure entails an annual increase in the number of households in the sample, it depends on the number of households upgraded the sector by listing operation. In PNAD 2008, approximately 151,000 households were selected. The final size of the sample of PNAD 2009 was approximately 851 municipalities, 7818 153837 sectors and households. In 2007 PNAD introduced the use of electronic collector ( Personal Digital Assistant - PDA) for carrying out data collection, making it possible to improve the research operating system . Also during PNAD 2007 the DIA system was used, which is an imputation system that automatically detects qualitative data errors. Developed by the National Institute of Statistics - INE of Spain, the software aims to facilitate debugging censuses and large statistical research. In this first year of use of the application, all steps of criticism usually applied to data from the National Household Survey core questionnaire were performed, followed by a process of simultaneous validation of the data collected. In 2008 PNAD used only the Canadian Census Edit and Imputation System - CANCEIS already including the procedures usually applied to critical data from the questionnaires. Starting from PNAD 2011 sample selection of Rondônia, Acre, Amazonas, Roraima, Pará and Amapá followed the same methodology in other units of the Federation.

Mode of data collection
---------------------------

Face-to-face [f2f]

{'Abstract': '本国家户调查（PNAD）每年定期对人口、教育、劳动、收入和住房等一般特征进行深入研究，以满足国家信息需求。此类特征包括移民、生育、婚姻、健康、食品安全等多个方面。对这些统计数据的调查是制定、验证和评估社会经济政策以及改善巴西生活条件的重要工具。', 'Geographic_coverage': '全国范围', 'Analysis_unit': {'- Households': '家庭', '- Individuals': '个人'}, 'Kind_of_data': '抽样调查数据 [ssd]', 'Sampling_procedure': '调查通过随机抽取家庭样本进行。信息由居住或非居住在调查范围内的个人提供，该个人被认为能够为整个社区和家庭提供信息。调查员被指示不接受14岁以下的个人作为信息提供者。抽样计划采用三阶段分层抽样（分别针对市镇、普查区和家庭），第一阶段的地形单位分层针对每个州设定。在人口规模较大的市镇和属于大都市区的市镇中，每个市镇被视为一个层，因此它们被确定性地纳入样本，称为自代表层。同一地理微区域内其他市镇被分为约等大小的层，并指定为非自代表层。这些市镇的层以与规模成比例的概率（ppt）系统地选择。在第二阶段，选择单位是部门，并且也以ppt系统地选择，此时规模由家庭数量衡量。部门根据北部地区城市和农村州的情况进行分层，但托坎廷斯州除外，以便将2004年后的PNAD指标与插入北部州农村地区之前的指标进行比较。在其他地区，这种分层仅是隐含的，即，在选择之前对部门的情况进行排序。市镇和选定的部门在样本中保留，直到有新的普查数据可用，届时将选择新的样本单位。每年，在选定的样本部门中，都会在现场准备（或更新）家庭清单，从而产生用于选择的更新登记册。这项清单操作的一个重要特征是新建建筑登记册，该登记册旨在包含对部门规模重大变化的建筑账户。新建筑的清单在样本市镇的市镇内进行，包括选定的样本部门和未选定的部门。新建筑区域从原始部门区域中排除，并在选择家庭时单独处理，在这种情况下，选择是根据该区域样本比例进行的。家庭是第三阶段选择阶段的单位，由私人家庭和清单操作期间被占用的集体住房单位组成。样本中每个部门的初始家庭数量设定为16。抽样比例表示构成样本的人口的比率。目前，比例从1/50（罗赖马农村地区）到1/800（圣保罗）。为了保证样本的自加权，每个选定的样本部门中的家庭选择是系统地进行的，选择范围的家庭数量每年保持固定。此程序导致样本中家庭数量的年度增加，这取决于通过清单操作升级的部门家庭数量。在PNAD 2008中，约选择了15,100个家庭。PNAD 2009的最终样本规模约为851个市镇、7818个153,837个部门和家庭。在2007年PNAD中，引入了电子收集器（个人数字助理 - PDA）的使用，用于进行数据收集，从而有可能改进研究操作系统。在2007年PNAD期间，还使用了DIA系统，这是一个自动检测定性数据错误的归因系统。该软件由西班牙国家统计局（INE）开发，旨在简化人口普查和大型统计研究。在应用的第一年，对国家户调核心问卷数据通常应用的批评步骤的所有步骤都进行了，随后是对收集到的数据的同步验证。从PNAD 2011年开始，罗纳尼亚、阿克里、亚马孙、罗赖马、帕拉和阿马帕的样本选择遵循联邦其他单位的相同方法。', 'Mode_of_data_collection': '面对面 [f2f]'}