1988 Income Sample of Blue-Collar and White-Collar GDR Households (EKS)

Data on individuals

1988 EKS: content and survey method used

In addition to the census, household budget survey, and the time-use survey, the EKS was one of the most important population surveys carried out in the GDR.

The EKS was first drawn in 1959; sampling was carried out regularly every two or three years thereafter, with the last sample drawn in September of 1988. It was designed as a representative sample survey in which households were surveyed at their workplaces. That is, information was gathered about individuals and other members of that individual's household through interviews conducted at the workplace. Respondents were chosen by a process of restricted random selection (see below), with participation in the survey voluntary.

The EKS' purpose was to gather meaningful information from both blue- and white-collar households about income - its levels, composition, and changes. Such information was differentiated by various criteria, including household size, marital status, and education levels. Information was also gathered regarding the possession of durable consumer goods, the ownership and use of yards and weekend properties (GDR Statistical Office, 1990). About 30,000 households, consisting of around 80,000 persons, were surveyed by the 1988 EKS. According to the former Statistical Office of the GDR, this represented approximately 0.75% of the blue- and white-collar households at the time.

Three main characteristics were examined in the survey:

household income (monetary income derived from work, transfer payments, etc., for all those who lived in and contributed to the household). All such figures were given for the month of August. One-time income that did not fall in the month of August was also surveyed and  included proportionately in determining average net monthly household income.

household possession of durable consumer goods such as televisions, automobiles, refrigerators, freezers, and the like.

Survey Methods Used

The income sample was restricted to the target population of blue- and white-collar households. In analysing the EKS, however, it is worth noting that the terms "blue- and white-collar households" are not as representative as they might appear. Rather, the structure of the EKS reflects a number of GDR-specific features, including the systematic exclusion of certain population groups. As a result, in interpreting the findings, it should be taken into account that there is a difference between the target population and the sampled population (Krause and Schwarze 1990), as can be seen in the following description of some of the problems.

Thus, though the observed units of the EKS were intended to encompass blue- and white-collar households, including those of still employed pensioners, the selection of households to be surveyed was conducted solely at the workplace. Yet the likelihood that an individual household would be selected depended on the number of household members who were categorized as currently employed blue- or white-collar workers. The problem with this method is that in a properly representative household survey, households are randomly selected by address, such that each household has an equal chance of being selected regardless of the number of persons working in that household or their employment status. If households are instead surveyed at the workplace, as was true for the EKS, the likelihood that a specific household would be included in the sample increased the more blue- or white-collar workers there were in that household (Krause and Schwarze 1990). To what extent this selection method led to an over or under representation of specific household types can not be estimated, because baseline figures on household types are currently not available for comparison.

Another problem lay in the fact that the EKS employed a (restricted) stratified, two-stage random sample design. In the first stage, the companies were drawn, and in the second, the blue- and white-collar workers who were to be interviewed were selected. One should therefore speak of a "restricted" random sample, because the population to be sampled was limited from the outset. In the following section, this will be described in more detail (Krause and Schwarze 1990; Frick, Krause, and Schwarze 1991).

Stage 1: Selection of Companies

The population to be sampled in the first stage included all companies that were part of the GDR domestic economy, including agriculture and forestry, as well as enterprises and facilities of consumer cooperatives (Konsumgenossenschaften).

At the time the 1988 survey was taken, about 92% of all blue- and white-collar workers were included within the scope of this survey. Explicitly excluded from this selection, however, were socialist cooperatives such as political parties, interest groups, private firms dealing solely with socialist enterprises, private firms with state participation, and entirely private firms (Arbeitsrichtlinie für das Bezirkszählbüro [Working Guidelines for the Regional Census Bureau] 1988). Members of such enterprises could only be indirectly included in the survey if they lived in the same household with a person otherwise surveyed.

The target size to be included in the sample survey was set at about 30,000 households, which in turn served as the basis for selecting the businesses. This target size was determined proportionately to the economic sector and the number of blue- and white-collar workers in the enterprises, and was differentiated according to the 15 GDR regions. In addition, a list of businesses was created that was sorted by economic sectors. Within this list, enterprises were ranked in ascending order by mean gross income per blue- and white-collar worker. Based on these two lists, the respective regional census bureau selected companies that fell into the lower, middle, and upper income ranges, and then determined the number of interviews that were to be conducted, respectively, in each enterprise. The average gross wage per blue- and white-collar worker contained in this sample of selected enterprises was then compared with the average wages for that economic sector throughout the GDR. If the average gross wage for the nation fell outside the range contained in the list of selected enterprises, the selection of enterprises had to be re-examined, and where necessary, companies were substituted for one another in the selection (SZS 1988a). It is unknown to what degree these kinds of corrections were actually undertaken.

Stage 2: Selection of Individuals (Households)

A further restriction in the sample population occurred at the individual level. The following categories of persons were not included in the selection: employees of enterprises who were not GDR citizens, trainees and those who worked at home, employees who lived in households of members of the armed services (with the exception of those who were doing their required military service), or in households of full-time political party and other interest group employees. This last category was also known as the X-domain - persons who belonged to it were never included in sample populations, other than in the case of the national census.

In general, the selection of interviewees was determined by the target numbers recommended by the SZS. The selection basis was provided either through a wage or salary card index, or by a list of employees and a counting rule determined for each enterprise. Intervals were determined by the quotient of the total number of blue- and white-collar workers in a company and the specific number of responses mandated for that enterprise, such that, for example, every 17^th blue- or white-collar worker in a specific firm would be surveyed. Yet the selection of those to be surveyed was not just limited to specific groups of employees. It was also limited by the fact that once those who were to be surveyed were selected, a check was run to see whether the average gross wages of those selected was in accordance with the average gross wages of blue- and white-collar workers of the entire business, though a deviation of plus or minus 5% was acceptable.

Conclusion

As should be clear from the foregoing description, the EKS therefore provides only a select depiction of all the blue- and white-collar households in the GDR. As Krause and Schwarze (1990) note, the process by which this selection was undertaken can nevertheless be described and the effects of the limitations in the sample design therefore factored in. More problematic are the internal "controls" in the selection of the enterprises, as well as the "controls" on levels of income of those who are to be surveyed; there is no information about either, but both may affect the survey results. Krause and Schwarze assume that these corrections have the particular effect of excluding outliers, with the result that an analysis of the EKS that focused on the mean values is still possible, even though an extrapolation of the EKS survey results is not advisable.

Previous Publications:

The results of the 1988 EKS have been only sparsely reported in the 1990 GDR Statistical Yearbook (in the section on the characteristics of social statistics), and in a small brochure reporting the survey results entitled "Haushaltseinkommen und Ausstattung der Haushalte von Arbeitern und Angestellten 1980-1988" [Household Income and Household Durable Goods of Blue- and White-Collar Workers, 1980-1988].

References:

• Krause, P./Schwarze, J. (1990): Die Einkommensstichprobe in Arbeiter- und Angestelltenhaushalten der DDR vom August 1988 - Erhebungskonzeption und Datenbankzugriff. DIW-Diskussionspapier Nr.11, Berlin.

• Frick, J./Krause, P./Schwarze, J. (1991): Die Einkommensstichprobe in Arbeiter- und Angestelltenhaushalten der DDR vom August 1988. In: Das Sozio-Ökonomische Panel (Hg.): Lebenslagen im Wandel, Campus. S.42-56.

• Staatliche Zentralverwaltung für Statistik (1988a): Arbeitsrichtlinie für das Bezirkszählbüro zur Vorbereitung und Durchführung der Einkommensstichprobe 1988 in Arbeiter- und Angestelltenhaushalten. Drucksache E 4. Ministerrat der DDR, Berlin.

• Staatliche Zentralverwaltung für Statistik (1988b): Arbeitsrichtlinie für das Betriebsverantwortliche/Beauftragte der SZS zur Vorbereitung und Durchführung der Einkommensstichprobe 1988 in Arbeiter- und Angestelltenhaushalten. Drucksache E 4. Ministerrat der DDR, Berlin.

• Hartig, K.H. (1990): Übersicht über die Datenbasis der amtlichen Statistik der ehemaligen DDR zu ausgewählten Kennziffernkomplexen aus der Sicht der sozialwissenschaftlichen Analyse und Forschung. Unveröffentlichtes Manuskript.

• Statistisches Amt der DDR (1990): Übersicht über die regelmäßigen repräsentativen Bevölkerungsbefragungen des Statistischen Amtes der DDR. Berlin.