The microcensus provides official representative statistics of the population and the labour market in Germany. The labour force survey of the European Union (EU labour force survey) forms an integral part of the microcensus. Since 1957 – in the new Länder (including Berlin-East) since 1991 – the microcensus has supplied statistical information in a detailed subject-related and regional breakdown on the population structure, the economic and social situation of the population, families, consensual unions and households, on employment, job search, education/training and continuing education/training, the housing situation and health.
The purpose of the microcensus is to provide benchmark and structural data at regular and short intervals on the above-mentioned topics and their development, thereby bridging the data gap between two population censuses. The design of the microcensus as a multi-topic survey abolishes the need for a considerable number of single-topic surveys which would otherwise be required. This has a cost-saving effect for official statistics and reduces the burden imposed on the respondents. With regard to a number of small-scale surveys of empirical social and opinion research as well as of official statistics, the microcensus is a tool for extrapolation, adjustment and control. The integrated labour force survey, which supplies comparative statistical data on the level, structure and development of employment and unemployment in the EU member states, is an important instrument for the EU Commission. The majority of the variables covered by the labour force survey are also microcensus variables.
The results of the microcensus are used in government reports and in the annual report of the German Council of Economic Experts; they provide the basis for current employment research, for the Federal Government's annual Pension Insurance Report and much more besides. The results of the labour force survey are used, for example, in the allocation of the European Union's regional and social funds.
How is the microcensus conducted?
The microcensus is organised as a decentralised statistics, which means that the organisational and technical preparation is done at the Federal Statistical Office, while executing the survey and processing the data are tasks of the statistical offices of the Länder. It is based on a legal provision with a limited period of validity, the Microcensus Law. In 2004, the former "Law on the Execution of a Sample Survey of the Population and the Labour Market and of the Housing Situation of Households" of 17 January 1996 (Microcensus Law 1996; Federal Law Gazette I, p. 34) was replaced by the same-named Microcensus Law of 24 June 2004 (Microcensus Law 1995; Federal Law Gazette I, p. 1350). The new legal basis for the microcensus has led to content-related and methodological changes in the microcensus design for the years 2005 to 2012.
Until 2004, the survey design of the microcensus was characterised by a fixed reference week, which means that the majority of questions related to the situation in one single reference week per year. Usually that was the last holiday-free week of April. Consequently, the microcensus results up to 2004 provided a snap-shot of the conditions in spring and were affected to differing degrees – depending on the variable concerned – by seasonal variations. Article 3 of the Microcensus Law 2005 provides for an infra-annual continuous survey. In such a survey, the entire scope of the survey is spread evenly across all calendar weeks of the year, with the last week prior to the data collection being the reference week (so-called moving reference week). Hence it is possible to provide the microcensus users not only with annual but also with quarterly average results, and thus to offer a significantly larger and more timely supply of data with an increased informative value.
As regards the list of questions included in the microcensus, Article 4 of the Microcensus Law 2005 lays down the topics to be covered in the individual years and their periodicities. In contrast to the law applicable up to 2004, no sub-samples will be examined any more because a sampling fraction of 1% of the population has generally been defined for all variables. This makes the microcensus the largest annual household survey in Europe. As in the years from 1996 to 2004, the question programme consists of a standard programme of variables that are surveyed every year, and a number of variables that are covered only at four-year intervals (4-yearly additional programmes, for example questions on health).
The content-related changes in the microcensus survey programme as from 2005 mainly relate to the inclusion of a new set of topics on "migration and integration". One part of the respective questions is covered every year, the other at four-year intervals. The section on "education" for the first time includes questions on the field of highest vocational qualification and on the kind of vocational qualification besides a higher education degree. To avoid increasing the burden on the respondents, some variables have been deleted from the microcensus question programme (for example year of marriage; current attendance of nursery school, crèche or day care centres for school children; usual and actual hours worked per week (in terms of days); need for long-term nursing care; change of establishment).
The microcensus is a random sample in which all households have the same probability of selection. Within the territory of the Federal Republic of Germany, areas (sampling districts) are selected in which all households and persons are interviewed (one-stage cluster sample). Every year, a quarter of the households (or sampling districts) included in the sample are exchanged. This means that each household remains in the sample for four years (partial rotation procedure).
A variety of survey instruments are used in the microcensus. The majority of data are collected by interviewers from the statistical offices of the Länder, who conduct face-to-face interviews with all persons in a household. However, the household members may also complete the questionnaire in writing. There is an obligation to provide information for most of the questions, while only few variables are legally exempted from that obligation. Variables which are exclusively covered for purposes of the EU labour force survey are always subject to voluntary response. Proxy interviews are permissible in the microcensus, which means that one household member may supply information on other household members. Proxy data are available on about 25% to 30% of all persons aged 15 years or over.
Due to the obligation to provide information, the proportion of known non-response by households (unit non-response) is very small at about 5%. The proportion of non-response for individual questions or variables (item non-response) is generally well below 10%. It may however be significantly higher in individual cases, especially for sensitive variables with voluntary response (for example 24% on an annual average in 2005 for the question relating to body weight). Compared with the microcensuses conducted up to 2004, item non-response has decreased. This is probably due to the fact that the change-over to the continuous survey has been accompanied across Germany by an increasing use of laptops, which in turn resulted in a further standardisation of the interviews.
When are the microcensus results published?
Quarterly results are available about three months after the end of the reference quarter and annual results about three months after the end of the survey year. The monthly and quarterly movements of the years 2005 and 2006 were characterised by unusual effects related among other things to the introduction of the continuous microcensus. This is why the Federal Statistical Office so far has published only annual results (annual averages) for the years from 2005 to 2008. The Federal Statistical Office plans to release also quarterly results in the future. The federal results of the microcensus are published as online products in the publication service as well as in various Subject-Matter Series and Cross-Section Publications of the Federal Statistical Office. Comments on the results and methodological papers are published, among others, in the journal "Wirtschaft und Statistik" (Economy and Statistics, published monthly by the Federal Statistical Office). Results concerning topical issues are presented at a press conference and in press releases. The release calendar and the press releases can be found on the Federal Statistical Office's website. Apart from federal results, a great variety of data for the Länder are available, which are published by the respective statistical offices of the Länder.
How accurate is the microcensus?
Sample surveys like the microcensus are affected by two kinds of errors: random sampling errors and systematic errors. Random sampling errors are deviations arising from the fact that not all units of the population are included in the survey. The standard error, which is calculated using the individual data of the sample, serves as an estimate for the random sampling error. The simple relative standard error exceeds 15% for extrapolated annual results of less than 5,000 and – if based on provisional figures – extrapolated quarterly results of less than 20,000, which relate to fewer than 50 persons in the sample. Such results are of little informative value and should therefore not be used for comparisons. This is why extrapolated cell frequencies of less than 5,000 at an annual level or less than 20,000 at a quarterly level are not recorded in microcensus publications but are replaced by a slash ("/").
Systematic errors are non-random deviations which may arise from errors occurring at any stage of statistical production (for example deficiencies in questionnaire design and interviewer training, wrong information provided by respondents and interviewers, data capture errors). Comparisons with other sources of employment data, such as the employment accounts compiled as part of national accounts or the ILO (International Labour Organization) telephone survey conducted by the Federal Statistical Office, have shown that minor employment – especially minor part-time activities – and the search for such jobs have so far been subject to undercoverage in the microcensus. Various hypotheses may be used to explain the resulting deviations and they suggest that the deviations are caused by content-related errors in the microcensus. Wrong information from the respondents or mistakes of the interviewers may, for instance, be due to the fact that the very broad ILO--International Labour Organization concept, which is not based on the relevant national legal provisions, does not correspond to the public's general understanding of the main activity status. The most important difference between the ILO activity status concept and the main activity status concept is that small-scale paid jobs, which are often done by pupils, students, housewives and pensioners, are included in employment. Respondents who keep the main activity status in mind will usually not classify such jobs as employment. The same applies to the search for such jobs which is classified as unemployment under the ILO concept – but not under the main activity status concept. The proxy interviews, which are permitted in the microcensus, may be another source of errors. Proxy interviews account for some 25% to 30% of the information collected in the microcensus about persons aged 15 years or over. The age group of 15 to 19 year olds, which typically comprises the majority of pupils, reaches the highest percentage of proxy information at 75%. This is a likely explanation of why it is difficult to determine the activity status according to ILO standards especially among pupils: It must be assumed that many parents who provide proxy information for the microcensus are not aware of their son's or daughter's activities at the margins of the labour market. Further problems with the application of the ILO concept may arise from the fact that many persons in minor employment find themselves – subjectively or objectively – within the grey area surrounding illicit work. Any fears that the respondents may have about the use of their data may lead them to conceal such activities.
The undercoverage of minor employment does not only result in an underestimation of the overall level of employment and hence of the employment/population ratio. What is also affected is the structure of employment, for example the breakdown of persons in employment according to demographic variables such as age and sex, by part-time and full-time employment or the average number of hours worked. The undercoverage affecting the search for minor employment leads to an underestimation of unemployment. On the other hand, minor activities that are performed by unemployed persons, but not recorded in statistics, result in an overestimated level of unemployment. Both factors are likely to influence the structure of unemployment variables (for example length of job search, kind of job wanted). Detailed information on current inconsistencies between the microcensus, the employment accounts of national accounts and the ILO telephone survey conducted by the Federal Statistical Office can be found on the Office's website (only in German).
Explanations of terms
Self-employed and employees Self-employed persons and employees include self-employed without employees, self-employed with employees, public officials, judges, salaried employees, wage earners as well as homeworkers. Apprentices and persons doing alternative civilian service were deliberately excluded from the analysis of working conditions because they are subject to special provisions regarding protection.
Type of employment Within the microcensus, the type of employment (full-time, part-time, minor employment) is covered through self-assessment of the respondents. It must be taken into account here that persons with a usual number of hours worked of more than 31 hours were generally counted as full-time employees in the microcensus up to 2008, using relevant data editing. When analysing the actual number of weekly hours worked, only the first or main job was taken account of. Consequently, all respondents for whom a minor job is the only job, or main job, were classified as "in minor employment".
Weekly working hours When analysing the weekly working hours as collected by the 2007 microcensus, only the hours actually worked in the reference week (week preceding the survey) in the first or main job were taken into account. Also, only those self-employed and employees were included in the evaluation whose hours worked in the reference week corresponded to the hours usually worked, or who worked more in the reference week than they did usually, or who worked less in the reference week than they did usually because of (special) vacation or compensatory time off (also flexible working hours). The number of hours actually worked in the reference week is based on the respondents’ self-assessment and is indicated in full hours. It should be noted that covering the hours usually or actually worked in the microcensus obviously involves some rounding effects.
Ad hoc module 2007 In 2007, an ad hoc module included questions on “work-related health problems (excluding accidents at work)” and “strain factors at work” in addition to the survey programme of the microcensus and the European labour force survey. The questions of the ad hoc module were covered only in a 10% subsample of the entire microcensus population. Answering those questions was voluntary. The analysis included only those respondents who answered the entire group of questions on work-related health problems or on strain factors at work. The response rates for the former group of questions was 81.7% for nurses, 78.8% for employees in health service occupations and 81.5% for employees in the overall economy. The response rates for the latter group of questions were 78.3% for nurses, 73.4% for employees in health service occupations and 78.2% for employees in the overall economy.
Topics Archive
STATmagazin