Labour Force Survey 4th quarter 2013 - 3rd quarter 2015, panel

Labour Force Survey 4th quarter 2013 - 3rd quarter 2015, panel (8 quarters)

As of the 1st quarter of 1972, SSB has conducted official quarterly labour force surveys (AKU). These surveys aim to give the labour force authorities (and other people interested) knowledge of the occupational structure of the population and how it develops over time. The surveys are meant to give a foundation and statistical material for occupational prognoses and labour research.

In 1996, AKU was significantly revised: The questionnaire, the file description and the standard for coding of industry and occupation. The data collection also changed to CATI - Computer Assisted Telephone Interviewing. A new classification of industry was put into use (NOS C 182, based on the EU standard NACE, Rev.1). This standard was updated in 2002 and 2007. Also, the new Norwegian standard classification of occupations (STYRK) based on ISCO 88 was used from 1996 and onwards. The variable indicating socio-economic status was omitted, as a similar variable was not developed in the new occupational classification.

Every second quarter AKU is extended with a number of questions to collect data accordant to Eurostats / EUs spesifications.

From 1996, the rotation schedule changed, each family now attends eight consecutive quarters. As of 1996 the selection shedule was changed as well with stratification at the county level.

Summary of AKU panels with detailed explanations for panel files:
Data collection for the new rotation plan as the basis for this file is shown below. Participants in the 4th quarter of 1997 is the first batch of 8 possible completed interviews.

Pool |2009 |2010|2011|2012|2013|2014|2015|2016|
Quarter |1234|1234 |1234|1234 |1234|1234|1234|1234|
---------------------------------------------------------------------------------------------
|____|____|____|____|____|____|____|____| 2013,1
|____|____|____|____|____|____|____|____| 2013,2
|____|____|____|____|____|____|____|____| 2013,3
|____|____|____|XXXX|XXXX|____|____|____| 2013,4
---------------------------------------------------------------------------------------------
|____|____|____|_XXX|XXXX|X___|____|____| 2014,1
|____|____|____|__XX|XXXX|XX__|____|____| 2014,2
|____|____|____|___X|XXXX|XXX_|____|____| 2014,3
|____|____|____|____|XXXX|XXXX|____|____| 2014,4
---------------------------------------------------------------------------------------------
|____|____|____|____|_XXX|XXXX|X___|____| 2015,1
|____|____|____|____|__XX|XXXX|XX__|____| 2015,2
|____|____|____|____|___X|XXXX|XXX_|____| 2015,3
|____|____|____|____|____|____|____|____| 2015,4

Panel files are created by linking eight interviews from an ordinary interview round of the Labour Force Survey. In the table above means that one could connect all the batches (lines) marked X.
Participation 8 times used as a criterion for selection.

Annual files and quarterly files consist of respondents who are participating for 1st time, 2nd time, and so on. The weights allow such cross-sections (columns) can be balanced with national figures. If 4 quarters is merged into an annual volume, the quarterly weights be divided by four to attain correct weighted annual numbers.

Each panel (panel files: horisontal in the table) will have a weight compiled for each quarter. Although the weights do not differ much from each other, they will be compiled independently, and each weight should be used only for the data which the quarterly weight is made for. Weighted numbers is valid for one pool, which means ca 12,5 % (1/8) of a total quarterly selection. Therefore, one must multiply by 8 to approximate national quarterly figures

By "Approximate national quarterly figures" it means that differences from the quarterly national weighted figures can appear, because of the selections and whitdrawal in the connection of the panels. Because of the whitdrawal inflated numbers will be lower than the national total figures. The main point of the weights is not to is not to give national totals, but rather to ensure the best possible representation.

2013年第四季度至2015年第三季度劳动力调查面板数据（共8个季度）

自1972年第一季度起，挪威统计局（SSB）便开展官方季度劳动力调查（「AKU」）。该调查旨在为劳动力管理部门及其他相关方提供人口职业结构及其随时间演变的相关信息，为职业预测与劳动力研究提供基础数据与统计素材。

1996年，「AKU」进行了重大修订：更新了问卷、文件描述以及行业与职业编码标准。数据采集方式调整为计算机辅助电话访谈（CATI, Computer Assisted Telephone Interviewing）。同时启用了基于欧盟标准《经济活动统计分类第1修订版》（NACE Rev.1）的行业分类标准（NOS C 182），该标准分别于2002年与2007年完成更新。此外，1996年起正式采用基于《国际标准职业分类1988》（ISCO 88）的新版挪威职业分类标准（STYRK）。由于新版职业分类体系未开发对应变量，原社会经济地位指示变量被移除。

每两个季度，「AKU」会新增一批调研问题，以符合欧盟统计署（Eurostats）的规范要求采集数据。

1996年起，调查轮换机制作出调整：每个受访家庭将连续参与8个季度的调查。与此同时，抽样方案同步调整为以县级为单位开展分层抽样。

以下为基于新版轮换方案生成本数据集的相关说明。1997年第四季度的受访对象为首批完成8次完整访谈的样本组。

样本池 |2009 |2010|2011|2012|2013|2014|2015|2016|
季度 |1234 |1234|1234|1234 |1234|1234|1234|1234|
---------------------------------------------------------------------------------------------
|____|____|____|____|____|____|____|____| 2013年第一季度
|____|____|____|____|____|____|____|____| 2013年第二季度
|____|____|____|____|____|____|____|____| 2013年第三季度
|____|____|____|____|XXXX|XXXX|____|____| 2013年第四季度
---------------------------------------------------------------------------------------------
|____|____|____|____|_XXX|XXXX|X___|____| 2014年第一季度
|____|____|____|____|__XX|XXXX|XX__|____| 2014年第二季度
|____|____|____|____|___X|XXXX|XXX_|____| 2014年第三季度
|____|____|____|____|____|XXXX|XXXX|____| 2014年第四季度
---------------------------------------------------------------------------------------------
|____|____|____|____|____|_XXX|XXXX|X___|____| 2015年第一季度
|____|____|____|____|____|__XX|XXXX|XX__|____| 2015年第二季度
|____|____|____|____|____|___X|XXXX|XXX_|____| 2015年第三季度
|____|____|____|____|____|____|____|____| 2015年第四季度

本面板数据集通过将劳动力调查常规访谈轮次中的8次访谈记录进行关联整合而成。上表中标记为「X」的样本组均可完成关联整合。参与8次调查为样本入选的核心筛选标准。

年度文件与季度文件涵盖首次参与、二次参与等不同受访频次的受访者。权重设置可确保各横截面样本（列维度）与全国统计数据保持均衡。若将4个季度的数据合并为年度数据集，则需将季度权重除以4，以得到准确的加权年度统计值。

每个面板（即表中横向的面板文件）将为每个季度单独编制权重。尽管各季度权重差异较小，但均为独立生成，且仅可用于对应季度的数据分析。加权统计值仅适用于单个样本池，约占季度总抽样样本的12.5%（1/8）。因此，若需估算全国季度统计数据，需将加权值乘以8。

此处提及的「估算全国季度统计数据」，是指由于面板抽样与样本流失问题，估算结果可能与全国季度加权统计值存在偏差。因样本流失，加权统计值会低于全国总统计值。权重设置的核心目的并非生成全国总量数据，而是尽可能确保样本的代表性。