6102.0.55.001 - Labour Statistics: Concepts, Sources and Methods, Apr 2007
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 19/04/2007
|Page tools: Print Page Print All|
CHAPTER 24. SURVEY OF EMPLOYMENT AND UNEMPLOYMENT PATTERNS
24.1 The Survey of Employment and Unemployment Patterns was a longitudinal survey covering the reference period September 1994 to September 1997. Data for the survey were collected in three waves - with each wave covering a 12 month reference period.
24.2 Longitudinal surveys involve collecting data from the same respondents at intervals over an extended period of time, with the information collected on each occasion linked together to provide insights into transitions over time.
24.3 In May 1994 the Government released a White Paper on Employment and Growth, titled 'Working Nation'. The overriding objective of Working Nation was to provide a comprehensive program to boost jobs growth, increase skill formation in the workforce, and ensure the long-term unemployed were not left behind during the economic recovery.
24.4 The central feature of the Working Nation initiatives was the 'Job Compact'. The Job Compact was targeted at the long-term unemployed, and provided individual case management and access to a range of labour market programs leading to a firm offer of a job placement for 6 to 12 months. The assessment of these initiatives was the driving force for the creation of the survey.
24.5 The objectives of the survey were to provide information on the dynamics of the labour market and to assist in the assessment of the Working Nation initiatives.
24.6 A range of data was made available from the survey in the form of publications, unit record data, special tabulations and occasional papers. Broad estimates were published in Australians' Employment and Unemployment Patterns, 1994 to 1997 (cat. no. 6286.0). An expanded Confidentialised Unit Record File (CURF) is available via the Remote Access Data Laboratory (RADL). For more information on this CURF, see Technical Manual: Australians' Employment and Unemployment Patterns: Expanded Confidentialised Unit Record File, Australia, 1994 to 1997 (cat. no. 6286.0.55.002). Information on CURFs and the RADL can be found on the ABS web site, under Services - Microdata - CURFs and TableBuilder.
TYPES OF VARIABLES
24.7 As part of the survey a wide range of labour market related data was collected. Variables were grouped into five categories - fixed, dynamic, episodal, occurrence, and summary.
Episodes of labour market activity
24.8 Episodes of labour market activity encompassed every day of the reference period and comprised periods of working, looking for work, or absence from the labour market (i.e. neither working nor looking). If a respondent had two or more jobs at the same time, each was treated as a separate episode. A change of employer constituted a new episode. An episode of working would overlap with an episode of looking for work if a respondent was working and looking for work at the same time. However, neither episodes of work nor episodes of looking for work could overlap with episodes of absence from the labour market.
Episodes of DEETYA labour market support
24.9 Episodes of DEETYA labour market support comprised periods of registration with the then Commonwealth Employment Service (CES), case management and Labour Market Program participation. Data for these episodes were provided by DEETYA and required the respondents' consent.
Episodes of income support
24.10 Episodes of income support comprised periods during which the respondent received income support. Data for these episodes were provided by the then Department of Social Security (DSS) and required the respondents' consent.
LEVELS OF OUTPUT
24.11 The focus of the survey was on the selected respondent, and so detailed information at the income unit, family and household level is not available. However, because a limited amount of information was collected from the respondent about their spouse and other usual residents, some broad household and family level information is available.
24.12 Estimates from the survey include:
Sex, age, marital status, birthplace, year of arrival in Australia, Aboriginal or Torres Strait Islander origin, educational attainment, language use, disabilities, housing tenure, household/family structure, and geographic region.
Labour force status at the time of each interview, trade union membership, and employment history.
Episodes of working
Start and finish date of each episode, status in employment, hours worked, permanent/casual, industry, occupation, earnings, sector, job preferences, and method of job attainment.
Episodes of looking for work
Start and finish date of each episode, whether looking for full-time or part-time work, active steps taken to find work, difficulties in finding work, mobility, and reservation wage.
Episodes of absence from the labour market
Start and finish date of each episode, availability to start work, and main activity.
Offers of employment, whether offer was accepted, and reasons for not accepting offer.
Types of training course, time spent on course, field of course, and course outcome.
Income amount, and sources of annual income.
Labour Market Support from DEETYA
Start and finish date of CES registration, reason ceased CES registration;
type of Labour Market Programs, start and finish date of Labour Market Program, post-program outcomes; start and finish date of case management, case management outcome; and reading, writing and speaking proficiency.
Income support from DSS
Start and finish date of income support, type of income support, and amount of income support.
THE TARGET POPULATION
24.13 The target population consisted of those people considered to be most likely to be currently eligible for labour market assistance or likely to become eligible for assistance in the near future. It was determined after consultation with government officials, labour market analysts and other users and was designed to enable the survey objectives to be met in the most efficient manner.
24.14 In broad terms, the target population comprised persons resident in private dwellings who were aged 15 to 59. Within this broad population, there were three subgroups:
SCOPE AND COVERAGE
24.15 The scope of the survey was all persons aged 15-59 years except:
24.16 The survey was conducted in urban and rural areas in all States and Territories and only included persons living in private dwellings.
24.17 For the Jobseeker and Population reference group subgroups, coverage rules were applied to ensure each person in scope was associated with only one dwelling and hence had only one chance of selection. Coverage rules were not needed for the Labour Market Program subgroup as specific people had already been identified as members of this subgroup.
PANEL ESTABLISHMENT AND DATA COLLECTION
24.18 Screening interviews were conducted between 24 April and 7 July 1995, for simplicity referred to as 'May 1995', across approximately 69,000 households. The purpose of these interviews was to identify respondents falling into the Jobseeker subgroup, and to establish the Population reference group. The screening interviews were conducted on an 'Any Responsible Adult' (ARA) basis. Following the identification of a target group member, personal interviews were conducted with potential respondents. Of those identified, 95% were recruited.
24.19 The Labour Market Program subgroup was not asked screening questions, as specific people had already been identified as members of this subgroup.
24.20 There were three collection waves for the survey:
24.21 Data for each collection wave were collected over a six week period following the end of the reference period. Questions relating to episodal data required the respondent to recall for what period(s) they had been working, looking for work or absent from the labour market. To obtain occurrence data, respondents needed to recall details of training courses and job offers received in the past 12 months.
24.22 Computer Assisted Interviewing was used in the second and third waves.
USE OF DOCUMENTARY SOURCES (ADMINISTRATIVE DATA)
24.23 With respondents' consent, data collected directly from them during the interview were supplemented with data about CES registration, case management and Labour Market Program participation (from DEETYA) and data about income support (from DSS). This minimised the interview time for respondents and ensured that accurate information was available about their involvement with labour market assistance programs and about their receipt of income support.
24.24 A probability sample design was used. The Jobseeker and Population reference group samples were drawn from the Population Survey Master Sample and excluded special dwellings and the remote and sparsely settled stratum. The Labour Market Program participants sample was drawn from a list of such participants provided by DEETYA.
24.25 The design used for the Jobseeker subgroup sample was the same as that used in the Labour Force Survey (see Chapter 20 for more detail). However, three major adjustments were made to the sample selection methodology:
24.26 For the Jobseeker subgroup, all persons aged 15 to 59 years in selected dwellings were 'screened' to determine whether or not they were a 'Jobseeker'. Only Jobseekers were recruited to the Jobseeker subgroup.
LABOUR MARKET PROGRAM PARTICIPANTS
24.27 The sample methodology for this subgroup was that of a list-based probability sample. DEETYA provided the ABS with a list of eligible persons. In order to produce a more efficient sample, the list was stratified into homogenous (similar) groups (strata) and then a random sample was selected from each group. The stratification variables used were: State of usual residence, length of unemployment, and job placement/training status. In total 2,300 people were selected. This number allowed for various types of sample loss, such as persons refusing to allow DEETYA to pass their name to the ABS, incorrect addresses, and overlap with the Labour Force Survey sample.
24.28 Procedures were put in place to ensure that the probability sample chosen for this subgroup was in some sense geographically close to the Jobseeker subgroup and the Population reference group (see below), but not overlapping these subgroups, nor overlapping the Labour Force Survey or any other ABS household survey.
POPULATION REFERENCE GROUP
24.29 A random subsample of the dwellings selected for screening in the Jobseeker sample was selected in the Population reference group dwelling sample. From each dwelling in the Population reference group sample, a randomly selected in-scope person was selected in the Population reference group, if such a person existed (some dwellings selected in the Population reference group dwelling sample did not contain any persons in the population of interest).
24.30 It is worth noting that, using this sampling methodology, it was possible for a person to be selected both in the Jobseeker subgroup and the Population reference group.
PANEL SIZE AND MAINTENANCE
24.31 The ability to maintain contact with a relatively high proportion of the panel was critical to the usefulness of the survey data. A number of strategies were put in place to help the ABS stay in contact with respondents between interviews. These included: 'change of details' cards for respondents to advise a new address, etc.; a toll-free telephone number for the respondent to call; asking the respondent for contact details of up to three people who were likely to know the respondent's whereabouts; and regular mail contact throughout the survey. However, it was inevitable that some sample attrition would occur when people were unwilling or unable to cooperate, or when they could not be contacted.
24.32 The attrition rate is the percentage of previous wave respondents who did not respond in the current wave. Attrition between waves caused a permanent drop in the sample size as the survey did not replace non-respondents. Although the weighting procedure for each wave partly corrected for attrition in the sample, there are some small differences in estimates between publications for each wave. Analysis of the attrition shows that higher than average sample loss occurred for males, young people, and people who were renting accommodation. Table 24.1 shows the size and composition of the panel at waves 1, 2 and 3, and the overall attrition rate.
24.1 Composition and Size of the Panel (Persons)
24.33 Calibration estimation techniques were used. Estimation was undertaken separately for the three components of the panel. Longitudinal weights were derived at each wave, benchmarking back to the population size/composition at the time of panel establishment. The use of longitudinal weights had the effect of always producing population estimates that related to the initial point of recruitment.
24.34 Labour Force Survey estimates (employed, unemployed and not in the labour force), were used to supplement independent demographic benchmarks (State/Territory of usual residence, part of State of usual residence, age and sex).
24.35 Aside from these 'standard' weights, it was also necessary to provide separate DSS, DEETYA and combined DSS/DEETYA weights. Although these weights produced similar estimates to the standard weights, their use was limited to analysis that used data from administrative sources. This is because these weights apply to a smaller sample; for example, the combined DEETYA and DSS weight should only be used in analysis that involves both DEETYA and DSS data.
RELIABILITY OF THE ESTIMATES
24.36 Estimates from the survey were subject to both sampling and non-sampling error (see Chapter 17 for more detail). The relative standard errors of survey estimates were published in each survey publication.
DATA COMPARABILITY OVER TIME
24.37 This was a one-off survey.
24.38 For further details contact the Labour Market Statistics Section, on Canberra (02) 6252 7206 or email <firstname.lastname@example.org>.