[1] Poland has the longest tradition of modern empirical sociological research (surveys) of any country in Central Eastern Europe. Due to the numerous sociological institutes which conduct empirical research as well as the large number of surveys which have been implemented, it is quite impossible to enumerate and describe them in comprehensive terms. This report therefore provides an outline of the most important developments.
[2] During the so-called Stalinist period, sociology was considered to be a ”bourgeois science” and was banned. However, the first substantive, non-ideological and basically positive (in terms of conclusions) review of Paul Lazarsfeld’s ”The Language of Social Research” was published by one of Poland’s leading Marxist philosophers as far back as 1956, i.e. right after the so called ”Polish October” or ”political thaw”. Sociology as a legitimate discipline in social sciences was soon re-established in academic institutions and sociological research reappeared. Even before establishing the first professional research organizations, several sociologists conducted research on, for example, family models, rural transformation, the intelligentsia and workers, attitudes to inequality and students’ attitudes.
[3] Public Opinion Research Center (OBOP), the first professional survey agency in Central Eastern Europe, was established by Polish Radio (later the Polish Radio and TV Broadcasting Corporation) as early as 1958. Although it was an institution financed by public radio and TV, OBOP cooperated very closely with sociologists from the Faculty of Philosophy and Sociology of the University of Warsaw. Academic sociologists conducted numerous surveys and OBOP researchers participated in academic life at the University. The surveys naturally concerned current social and political problems (although the latter were not sufficiently investigated due to censorship and political restrictions), but many of them were of an academic nature.
[4] It must be stressed that the establishment of OBOP in 1958 coincided with the UNESCO Warsaw seminar on public opinion, in which such prominent researchers as Paul Lazarsfeld, Stein Rokkan and Jean Stoetzel took part. Somewhat later, Herbert Menzel, a close collaborator of Lazarsfeld, became a regular OBOP consultant and spent a longer period of time in Warsaw. Training of Polish researchers in modern methods and their theoretical basis was made possible, for example, by a very large grant from the Ford Foundation. This grant also enabled many sociologists to attend American universities.
[5] There were some ups and downs in OBOP’s activities, the latter caused by political restrictions. However, different sociological research was always conducted by the Center, which is again very active nowadays.
[6] Empirical sociology became quite strong very early on at the University of Łódź and in some institutes of the Polish Academy of Sciences. The first Polish representative public opinion survey and the first panel survey were conducted by the University of Łódź in 1958, not by OBOP.
[7] One very characteristic feature of early Polish sociological research was that their authors placed a great deal of emphasis on methodological considerations at various generalization levels, i.e. not only on the methods as such (”cook books”) but also on much more general methodological problems dealt with by logic and the philosophy of science (using American terminology) or meta-methodology (as it was called in Poland). This was obviously the result of close contact with Lazarsfeld’s academic circles and ideas.
[8] Survey research developed so rapid that a serious - though somewhat too ideologically driven - discussion took place in 1962 concerning so-called ”ankietomania” (”questionnaire mania”). However, this did not seriously hamper further research.
[9] One of the rapidly developing fields of empirical research concerned social stratification, inequalities, social mobility, occupational prestige and fair earnings. Contrary to all other Central Eastern European countries, where a Marxist class approach was the only official credo in social sciences, modern research on stratification and mobility was adopted in Poland by both non-Marxist and Marxist sociologists. This was also a period of thriving rural sociology.
[10] Some time later, Polish sociologists took part in their first international comparative projects, e.g. concerning attitudes to local (regional) policies and problems or psychological aspects of work and social status. They also played an active role in establishing a new field of social research, namely ”social ecology”, which was based primarily (though far from exclusively) on the analyses of official statistical data supplemented at times by survey data.
[11] In the early 1970s the second highly professional survey network was organized in Poland by the Institute of Philosophy and Sociology of the Polish Academy of Sciences as a separate branch of the Institute. In the mid-1970s another survey network was organized by the Polish Sociological Association. Its aims are to provide services to academic researchers and earn some money for the Association. With some modifications, these two networks have existed and conducted sociological research up until now, although the first one network mostly subcontracts out surveys.
[12] By the end of the 1960s a relatively small though very active sociological research section had been established in the Central Statistical Office. In addition to various small-scale research activities, this section conducted in 1972 the socio-occupational mobility survey on the representative sample of 0.5 percent of the Polish working population. This is still the largest mobility survey both in Poland and the whole world, and its data are still used internationally. A great deal of sociologically relevant data were available from the ongoing statistical survey of household budgets (earnings, expenditure and consumption). Two large time budget surveys were conducted in connection with this survey. Sociologists and statisticians from the Central Statistical Office took part in international cooperation aimed at creating a system of socio-demographic statistics coordinated by the UN Statistical Division, the UN Conference of European Statisticians and the CMEA (”Comecon”) Statistical Commission.
[13] Sociological research and analysis have also been conducted since the mid-1970s by the Institute of Basic Problems of Marxism, which was established by the Central Committee of the PUWP (Communist Party). The Institute established a modern, computerized bank for its data. However, due to the particular political connotations and the Institute’s links to the party authorities, this bank was never widely used and ceased to exist after the institute was dissolved.
[14] The birth of ”Solidarity” in 1980 and the subsequent temporary democratization of social and political life, which was curbed in 1981 by martial law, resulted in the intensification of social research on socio-political and economic attitudes. The first wave of the survey ”Poles ‘80” was conducted, although its results were not disseminated until a few years later due to martial law restrictions. This was a very comprehensive survey of public attitudes towards a wide variety of social, political and economic issues.
[15] One of the peculiarities of Polish martial law imposed in 1981 was not only the relatively small number of people who were killed compared with similar events in other countries, especially Latin America, but the fact that the regime wanted to ascertain public opinion and established a government Public Opinion Research Center (Centrum Badania Opinii Spolecznej – CBOS). On the other hand, several – at that time – illegal public opinion surveys were conducted during martial law by researchers connected to the underground opposition. Given the circumstances, the representativeness of these surveys was dubious and the results were biased towards the opposition. However, the results of ”official” surveys, which were conducted by CBOS interviewers who presented themselves as representatives of a government research institution and were often recruited from military personnel or their families, were also biased due to the fear of repression or simply due to the prevailing distrust and dislike of the government.
[16] CBOS was founded in the autumn of 1982 and the first research was carried out in the following year. It was then a unit of government. The parliamentary act decreed its goals as follows: ”To provide information about social opinions concerning the economic and social decisions made and implemented by the government.” In order to distinguish it from the existing OBOP, the term ”social opinion” (opinia spoleczna) was used instead of ”public opinion” (opinia publiczna) in its name, although ”public opinion” was used in the English translation of both names. There was a very unconvincing attempt to explain the use of the term ”social opinion” from an ideological point of view. Up until 1989 CBOS conducted 134 surveys and published 441 reports, many of them confidential, and made a number of smaller contributions to periodical publications.
[17] was established by the military government during the period of martial law in Poland. Its motives are probably best summed up by a renowned Polish sociologist, who said that ”after the bitter lesson they had learned in 1980/81, the rulers decided to listen to public opinion, even if they did not follow what they heard.” The first director of CBOS was a professional social researcher from the military.
[18] The circumstances of the establishment of CBOS were reflected in the topics of the first research which it conducted. The first project, which was conducted in the spring of 1983, was about ”wages as perceived by workers in industrial enterprises in Warsaw.” A number of different research methods were used during the initial period of the activity of CBOS. Both surveys and content analysis of letters ”from the citizens” to the government were carried out. The contents of the press were also analyzed and a ‘telephone hotline’ was activated. At this early stage so-called ‘workshops with opinion-makers’ were organized. These workshops were mainly attended by scientists, journalists and, at times, opposition activists.
[19] With the passage of time, however, most non-survey methods of research were abandoned. Before 1989 CBOS was already specializing in survey research methods, mostly conducted on the basis of samples representative of the adult population. Quota sampling was used initially, and then CBOS started to use random address samples. In the 1980s it was quite common to conduct surveys on randomly selected social strata, e.g. ‘working class in large industrial enterprises’, due to the political sensitivity of that group, but also on ‘managerial cadres’, youth, the intelligentsia, etc.
[20] In the early stages of activity there were two types of surveys, i.e. comprehensive and topical. The comprehensive surveys, which started in 1984, usually focused on the attitude towards authorities and evaluations of current events (social sentiments). At that time they were first conducted twice a year and then four times a year. The reports from the surveys were partially classified information, but this subsequently became more seldom.
[21] In 1989 at the start of the new political regime, CBOS was already an institute with considerable experience and research achievements. However, its professional position was not matched by corresponding legitimation. There was also a fairly widespread, but only partially justified conviction that research was instrumental to goals of the ruling elites. There appeared a need for CBOS to find a place in the new democratic system so that it did not serve the political authorities, but instead specific societal groups and public institutions. It needed to function in a situation where public opinion research is an ”instrument of democracy” while also being subject to market competition.
[22] The lifting of martial law and gradual liberalization of the state socialist system resulted in an increasing number of surveys conducted by various academic institutions. The second wave of the comprehensive survey ”Poles ‘88”, which was conducted by the Institute of Philosophy and Sociology of the Polish Academy of Sciences, is probably the best example of this trend.
[23] The change in the political and economic system increased both the interest of academic circles in empirical sociological research and the possibility of conducting such research. Sociologists from almost all sociological departments of the universities have conducted their own surveys on a great variety of topics. The research activity of the Institute of Philosophy and Sociology of the Polish Academy of Sciences has also intensified. Especially important were two subsequent waves of ”Poles ‘90” and ”Poles ‘95”, which were conducted by this Institute. This series of four fully comparable surveys, which dealt with a large and comprehensive selection of important problems during the upheaval and transformation periods (1980-1995), was created in this way.
[24] The establishment of two new academic research institutes was particularly important. The first institute was the Institute of Political Studies of the Polish Academy of Sciences while the second was the Institute for Social Studies of the University of Warsaw.
[25] The Institute of Political Studies has conducted many surveys. Special mention must be made of a new series of the Polish General Electoral Survey, which has been repeated on the occasion of every subsequent presidential and parliamentary election since 1990. The Institute also attempted to organize the first central archive of Polish sociological research surveys. This attempt was unsuccessful, primarily on account of the resistance from the community of researchers, who wrongly perceived the project as a threat to their independence and autonomy. The idea was finally abandoned.
[26] The Institute for Social Studies conducts the annual Polish General Social Survey (PGSS). The Institute participates in the International Social Survey Programme (ISSP), which constitutes an integral part of the Polish General Social Survey (PGSS). The data from all surveys conducted by the Institute are documented and archived according to international standards. This is therefore the only institute in Poland which has a well-established and professionally run data archive for its research. (See appendix).
[27] It must be emphasized that, since 1990, Polish sociologists have participated in all major international research projects such as the World Values Survey, International Social Survey Programme, New Europe Barometer, Consolidation of Democracy Project, International Survey of Economic Attitudes, etc.
[28] Due to their great number and the lack of any registration body, it is impossible to present all academic sociological surveys comprehensively and systematically without extensive research.
[29] Two major public opinion research centers, OBOP and CBOS, have taken quite different development routes.
[30] CBOS, formerly a government institution, has acquired the status of an independent foundation (NGO) with a guaranteed annual grant from the state budget (although without any guaranteed amount of grant money). From 1990 onwards the comprehensive surveys have been conducted on a monthly basis. The reports were now open to the public and were used increasingly by the mass media. At that time the goals of the Institute were redefined: on the one hand, CBOS was to ”inspire, organize and carry out public opinion research in the sphere of interest of the government”, but on the other hand ”carry out public opinion research on its own initiative.” The statutes make it mandatory for CBOS to ”make the survey data available to the general public and the representatives of the scientific community.” Quite often, investigated problems are not only of topical importance, but are also of a more general, academic nature. The intensity and variety of the research activities of CBOS are best illustrated by the fact that between 1995 and 2001 it published as many as 1315 thematic reports and several books. In addition to its statutory activity, CBOS conducts commercial surveys for both academic institutions and business organizations.
[31] At present the data from individual surveys are stored as separate files in ASCII format. The documentation of the data (including the questionnaires) is available for public use in the CBOS Information Center where the researchers can locate the relevant survey, check out the variables and obtain the data. However, data from each survey have to be retrieved individually. An integrated dataset which serves as a basis for regional analyses has already been established and work is in progress regarding to organization of the comprehensive data archive linking reports, questions and individual data.
[32] OBOP has been privatized. The majority of its shares are now owned by one of the large international market research companies while some shares remain in the hands of the Polish Public TV Broadcasting Corporation. This means that the main emphasis is on TV audience research. However, a typical public opinion survey is also conducted every month.
[33] There are at least three commercial market and social research organizations which systematically or very often conduct public opinion research. Some of this research is commissioned by the media (one of the leading Polish daily newspapers and one weekly journal order the surveys systematically, and others do it occasionally), while some research is conducted in order to present the results publicly and to make the firm better known and renowned. Public opinion research is therefore often conducted as a public relations exercise. (There are also other commercial market research companies that conduct sociological or public opinion research very occasionally, either on the orders of their clients or as a public relations exercise).
[34] There are two major sources of sociologically relevant statistical data from the Central Statistical Office (CSO). The first source is an ongoing survey on household earnings, expenditure and consumption while the second is a monthly representative survey of the workforce. Both contain from time to time some questions other than their main topics. In addition, the Central Statistical Office conducts ad hoc social surveys. They constitute, however, a kind of research activity of secondary importance to the Office.
[35] There are no universal regulations relating to any kind of research institutions. They must therefore be discussed separately.
[36] A general rule so far is that the principal researchers and their research teams have full control of the data. In principle, it is almost entirely up to the researchers as to whether or not they publicize the data. The researchers themselves are very divided on this issue. Some researchers publicize data, some do not, and some do it on a very selective basis.
[37] The situation differs between the institutions and between the categories of research. As a public institution, Public Opinion Research Center (CBOS) observes the rule that all individual datasets from its ”statutory” surveys are available for the public as soon as all thematic reports from a given survey are published (usually 1-2 months after the survey, but sometimes even earlier). In the publicly accessible part of the archive it stores all the data – in an electronically coded format – from surveys from the years 1983-2001. Any person or institution can order any data from any survey, unless the survey was purchased by one of our commercial clients.
[38] The general public and the mass media mostly use the CBOS data through the research reports from the monthly omnibus surveys as well special, topical surveys whose results are also sometimes published as reports. The data from the extended topical surveys are also used in analyses which are contained in larger scale publications: books and scientific articles. However, the data from the surveys ordered by outside customers constitute their own property and the CBOS has no right to distribute them. Some clients (especially if they are scientific institutions) allow the findings to be publicly accessible. Other privatized research agencies are the owners of the data from their own (non-commissioned) surveys. There are no rules concerning their availability. The data from ordered surveys are the property of the clients (as above).
[39] The CSO can not make public any individual data which would allow the identification of the respondent or other investigated unit (business, institution, etc.). However, fully anonymous data are usually given to academic institutions and other legitimate researchers on request.
[40] Most of the sociological research is now financed by the national funding agency called the Committee for Scientific Research (similar to NSF, etc). However, there are also many other smaller sources of funds, including international foundations.
[41] There is growing interest among sociologists and public opinion researchers to establish a central archive for social research data. However, this depends on two vital issues, namely the funding and the consent of the research community. Current economic problems in Poland make it unrealistic to expect the availability of money necessary to establish a full scale archiving institution in the near future. However, some money may be available for developing and modernizing existing archives, such as those in the CBOS or the University of Warsaw. The most realistic idea is to initially attempt to organize the archive solely for those researchers and institutions who would voluntarily agree to share their data. This would be the first step towards a more comprehensive project based on some generally binding regulations. A good example would certainly lead to more consent.
[42] CBOS has created an integrated dataset which permits analysis of several important attitudes on a regional basis. This database is constructed by combining individual files from the omnibus surveys from one year (12 different samples, mid-2000 to mid-2001; then for a consecutive year) into one file. Throughout the year, several indicators were repeated every month:
[43] The variables from different surveys are assigned one name, thus creating a file with around 12,000 cases. This means that each “voivodship” contains enough respondents to draw conclusions on regional differences.
[44] The most significant archiving project undertaken by CBOSn of a database that will contain the data from the omnibus surveys from the years 1983-2002 as a single data universe. CBOS is constructing a relational database which will, in the final stage, comprise data from all omnibus surveys in one dataset with tools enabling the researcher to instantly find the data on the same question through documentation based on the DDI (Data Documentation Initiative). The DDI is ”an effort to establish an international criterion and methodology for the content, presentation, transport and preservation of ‘metadata’ (...) With the achievements of the DDI, codebooks can now be created in a uniform, highly structured format which is easily searchable, lends itself well to simultaneous use of multiple datasets and will significantly improve the content and usability of metadata.” 1 (Note1: from: http://www.icpsr.umich.edu/DDI)
[45] This type of database is particularly suitable for datasets composed of multiple files with long time series. Usually, academic surveys pertain to one particular issue and time-series data can be created after several separate data files are browsed, which requires extensive efforts on the part of the researcher. Comparability is achieved through recoding. However the CBOS database is different from strictly academic surveys:
[46] The target version of the dataset will be serviced by client programs for the database, which will make it possible to search the whole dataset using standardized codebook elements (DDI). At present standardized documentation (according to the DDI standard) is applied to the datasets and software, thus enabling users to locate the data on identical questions through which the data file is created or adapted. In the next step the data will be integrated with the relational database and with the client search software enabling the users to find the relevant data. The search engine will enable the users to locate data with the help of:
[47] Through the search engine it will be possible to locate both time series of the same questions and blocks of questions in a survey. It will be possible to generate a SPSS data file with the help of the above-mentioned search options and to calculate frequencies online as well.
[48] It is expected that the data from current years (i.e. 2001 and 2002) will be available by the beginning of mid-2002. Data from previous years will gradually be accessible and will eventually go all the way back to 1983. Due to the shortage of funds, it is impossible to tell when the complete archive will be available.
[49] The Polish General Social Survey (PGSS) is the first, and as yet the only, academic social survey research program in Poland, for which the data and full methodological and technical documentation are public and non-commercially disseminated. It is a program in which the main emphasis is placed on the literal repetition of questions and items and item sequences. As of 1992, close to 400 indicators have been strictly replicated in each subsequent survey. The fundamental aim of replication is to allow for systematic trend studies connected with the social, political and economic transformation process in Poland. The Polish General Social Survey (PGSS) allows for systematic study of the effects of systemic transformation in Poland in cross-national comparative perspective. This is possible due to co-operation within the International Social Survey Programme (ISSP) as well as due to the literal replication of about 50 indicators from the American General Social Survey (GSS)General Social Survey (GSS)/University of Chicago:National Organization for Research at the University of Chicago (NORC), and from the German General Social Survey (ALLBUS/ZUMA, Mannheim). Many substantive Polish General Social Survey (PGSS) indicators are also comparable to indicators in the British Social Attitudes Survey (BSA/SCPR, London). The entire (printed up to 1995 and since the 1997 available only in electronic form) PGSS methodological documentation (codebooks), along with a set of control statements to read the dataset into standard computer statistical packages, are also provided in English. Thanks to this, Polish General Social Survey (PGSS) data are also available to the international community of social scientists and are disseminated by the international social data archives (ICPSR), Central Archive for Empirical Social Research (ZA), Cologne, Roper Center, Economic and Social Research Council (ESRC), UK, Columbia University:Center for International Earth Science Information Network (CIESIN), TÁRKI).
[50] The originality of the Polish General Social Survey (PGSS) mission is based, above all, on the dissemination of data allowing other researchers to systematically study the social consequences of systemic change in Poland. In the past, the Polish academic milieu conducted numerous and inspiring survey studies on social, political and economic attitudes and behavior. The data gathered from these studies, however, have never been made publicly accessible nor used by "secondary" researchers. In effect, the exchange of information in a broader scientific environment took place only at the stage of the final scientific product (publications, printed communiqués with research findings and conference presentations).
[51] This situation, in which there was only limited exchange of data and information in the broader scientific community could be explained by, amongst other things, the normative legacy of science from the communist period. Intense institutionalization and centralization of scientific system and its technological infrastructure, favored the definition of intellectual scientific property as the property of the organization, or of the people gathering and producing scientific data, but not as the property of the scientific community as a whole, regardless of its structural and functional divisions.
[52] On the other hand, as in the majority of countries in the world, the dominant style of conducting social science research in Poland is to concentrate responsibility for all phases of research project in the hands of a few researchers (or centers and teams), from the formulation of the problem, through data gathering, processing and analysis, to the publication presentation of the research findings. This philosophy of "global responsibility" for a project is dysfunctional, even when the project researchers are interested in the secondary analysis of "their" data by researchers from other centers. It usually turns out that the simplified, working documentation on the research and the standards of data preparation which are prepared for the internal needs of the given center, (dataset integrity, data file structure, variable construction, codes and codebook documentation) are either insufficient or unintelligible, even to experienced researchers from other centers. When studies encompass large samples, and the information structure is more complicated (cross-section or longitudinal surveys for example), use of data from other centers often becomes technically impossible.
[53] The transformations in Poland have produced a huge increase in the number of commercial and marketing social research centers, the principle of whose work is, of course, protection of data and findings from public access. From the perspective of the development of „open” academic social studies, the development of commercial survey centers has become dysfunctional, as these centers, by offering very attractive wages, are able to employ the most talented graduates of the empirical social sciences, or researchers from universities who wish to improve their material status through additional work in the private sector – in one of the social science departments at University of Warsaw, for example, more than half of the doctoral graduates are employed in a second job in commercial social research centers.
[54] The systemic changes in Poland during the 1990s have, of course, had favorable effects in the sphere of democratization of the institutional scientific system (an increase in the autonomy of universities; independence of the financing of scientific studies from political power structures; and independence of a scientific career from ideological criteria). However the creation of an open system for the exchange of information in the social sciences in the context of the accumulated data gathered (often duplicated!) by autonomous research centers, still remains a problem. The main reason for this is to be found in the radical limitations on public spending for science and scientific research in this period of deep economic crisis that developed at the outset of the political transformation in Poland and other post-communist countries. Limited spending on social research has resulted in radical cuts, within scientific centers, to those funds which are devoted to the implementation of concrete projects that could be designated for archiving and public dissemination. This can be more readily appreciated if we take the experience of the Polish General Social Survey (PGSS) as an example. The preparation of PGSS data for deposit in public archives (for example, cleaning and checking the integrity of the cumulative datasets; preparation and publication of bilingual codebooks, machine-readable documentation, bilingual datasets and tools helpful in viewing results or finding desirable variables) consumes more than one-third of the total expenditures on annual PGSS research, and every year absorbs four months of work from three researchers.
[55] It should be added that the limitation on public funds for science has become a serious hindrance to the development of an independent social data archive in Poland. In this situation, the costs and efforts of data preparation and distribution have to become the responsibility of the centre and researchers carrying out the project.
[56] It was with precisely this conviction that we initiated preparation in 1991 of the Polish General Social Survey project, and founded the Institute of Social Studies at the University of Warsaw.
[57] Institute for Social Studies (ISS) was established in January 1991 as an independent research unit of the University of Warsaw on the basis of a Memorandum of Agreement between the University of Warsaw and the University of Michigan. Under this agreement, the Institute of Social Studies was affiliated with the internationally known Institute for Social Research at the University of Michigan, Ann Arbor. The ISS is an interdisciplinary research-oriented institution, bringing together sociologists, psychologists, economists and political scientists. Most senior members of the institute staff hold permanent appointments in the University of Warsaw social science departments.
[58] The ISS is largely self-supporting through grants and contracts from domestic and international foundations, corporations and the Polish government. Most research is funded by the Polish National Committee of Scientific Research and the Ministry of Education and Science.
[59] The main objectives of the Institute are:
[60] The PGSS, which is produced and distributed by the ISS, is an example of how the latter mission of the Institute is being achieved.
[61] PGSS data come from personal interviews administered to national samples using a standard questionnaire. The sample covers adults, 18 years of age and older, living in non-institutional households. Each PGSS survey uses an independent, multistage, stratified probability sample about 1,600 cases (and since 1997 – about 2200 cases).
[62] PGSS data include a wide range of variables, touching on many areas of current interest to social scientists. Over 1100 different variables exist in the 1992-1999 PGSS cumulative data file. The content of the survey changes somewhat from year to year, but core questions, and the order in which they are asked, have been kept intact, thus enabling comparative analyses across time. PGSS core variables include socio-economic and demographic variables with an emphasis on stratification measures (occupation, labor force status, education, income) of respondents, their spouses and parents. Additionally, there are attitudinal variables concerning political opinions and ideology, national spending, religious beliefs, social inequality, job and occupational values, tolerance, educational values, attitudes toward other countries, traditional sex roles, family issues, abortion, and homosexuality. Other variables gauge subjective well-being, social class identification, satisfaction with different spheres of life, and confidence in different public institutions. Respondents are also queried about their voting behavior, social interactions, religiosity and health. Each year, additional topical modules of questions from the ISSP have been added: "Inequality II” (1992), "Environment" (1993), "Family and Changing Gender Roles II" and "Sexual Behavior" (1994), "National Identity" (1995), „Role of Government III”(1996), and „Work Orientations II” (1997). In 1999 there were two more international modules conducted which will be distributed with the data-set originating from 1999 PGSS edition, „Religion II” (1998) and „Inequality III” (1999).
[63] Given the limited PGSS funds a new biennial, split-sample design has been used since 1997. The sample consists of two parallel sub-samples of approximately 1,200 cases each.
[64] The criteria for the selection of questions and indicators for the PGSS were indicated by the overall aims of the project. The major goal of this project has been to provide the social science community with large-scale substantively important data of high quality. Who decides at is important for the science community? And who evaluates the quality of the data? At the beginning of the planning of the PGSS survey, we had a clear answer to these questions an independent advisory group to the PGSS principal investigator, which should function as an intermediary between the social science research community and the PGSS. In 1992, the ISS Scientific Council appointed the PGSS Advisory Council, consisting of 10 representatives of various disciplines in the social sciences and from different organizational sectors of Polish science (universities, academy of sciences). Among its tasks are counseling in the research conducted as part of the PGSS program, the evaluation of methodological standards, and supervision over the structure of successive PGSS editions. The Council also examines proposals from research teams, and individual researchers, to include new questions, indicators or self-contained problem supplements in the PGSS project.
[65] With the hindsight of nine years' experience of the PGSS, it could be argued that consensus between the principal investigator and the PGSS Advisory Council is most important. In this way, the PGSS serves both those for whom the PGSS is an important vehicle for studying social trends in Poland (and also in cross-country comparative perspective), as well as those for whom the PGSS represents a vehicle for studying the methodological issues presented by complex surveys, or for testing the possible limits of knowledge about society as marked out by social survey technology and tools.
[66] The field research is carried out by a network of professional interviewers. The survey operation unit of the ISS exercises control over the substance of the field research, which includes: supervision over selection of the sample, testing of the research instruments, developing documentation connected with the field work, training the field managers and interviewers, and control functions during the field phase of the research project.
[67] The response rate is one of the criteria for evaluating the standards of realization of the survey research. When we add together the ineligible addresses (out of sample, not dwelling units, vacant dwelling units), the non-response from refusals, non-contacts, and others, the total non-response for the PGSS surveys ranges from 17.7% (the PGSS 1992) to 28.7% (the PGSS 1999).
[68] In general, the PGSS samples closely resemble the distributions reported in the census. Some differences may occur due to the nature of the sample design (the probability of selection of the respondent is inversely proportional to the number of adults in the household), and due to the unequal response rates in different social strata (eg in rural and urban areas for example). To adjust for such differences in the PGSS file, users can apply a weight variable, which simultaneously calculates results corrected to population parameters.
[69] From the perspective of our experience, the preparation of data for public access is not only a very time-consuming and costly activity, but is also the most stressful stage of the survey research process. We realized that the PGSS will be judged, not only in terms of whether the indicators are "interesting" and the picture of Polish society "curious", but also – and above all – in terms of whether the collection of data contains horrible errors. Public and anonymous users of an archived dataset may find errors everywhere in particular variables or in the codebook, or may be unable to replicate the previously published research results based on the same dataset. Knowledge about the possible errors in survey data is common among both researchers and data users. Several tools have been developed to reduce the lumber of possible errors in the PGSS dataset to minimum.
[70] Many examples can be found which illustrate the thesis that the value of concepts and scientific research findings is intimately associated not only with the logic and rationality of the underlying argumentation, but also with the aesthetic clarity of the empirical proofs presented. The PGSS team shares this conviction. Moreover, it recognizes that the degree of interest in the PGSS findings also depends on the ease with which researchers, and especially young researchers, are able to search for indicators and for data to test their own hypotheses in the PGSS/ISSP dataset, which has accumulated a considerable amount of, often eclectic, material.
[71] Since 1992, the PGSS has been organized as a cumulative file in which each year is a sub file. The raw data from the latest year are merged with the existing cumulative file, and an updated cumulative file is created. At the point of the annual update, the project also adds in any new variables (variables from new ISSP modules for instance) to the old surveys and makes any corrections to past years. The merged PGSS data file is stored in ASCII form, along with a set of control statements to read it into a standard computer program (SPSS). PGSS data are also archived in an Export/Transport portable system file. In our experience, ASCII (row) PGSS data files are the most attractive form for Polish and international users because they are software-independent and they have a better chance of being read in the future, regardless of what happens to individual statistical packages. The SPSS control and comment statements are distributed in the Polish and English languages. This results from the ''internationalization" of the PGSS program. To facilitate the search for indicators that interest the user, the set of control statements contains extended labels for each of the variables and all value labels for almost every variable.
[72] The basic documentation for the PGSS is the Cumulative Codebook which is updated with the creation of every new cumulative dataset. It contains the exact wording of all questions and response categories, the physical (column) locations of all variables in the rectangular data file, the number and mnemonic names of all variables, and the frequency distributions for all items in each year of the PGSS/ISSP survey. Details on how each variable was constructed are provided for those variables created by the project staff. In addition, the Cumulative Codebook has 18 appendices containing the project description; information on sampling and weighting; recodes; distributions for multicolumn variables; codes and descriptions for the International Standard Classification of Occupation (ILO-1988); the codes and descriptions for the Polish and the International Industrial Classification. There are also three indexes of the PGSS/ISSP variables with their associated codebook pages: one using Q-numbers; the second using PGSS mnemonics; and the third which lists those variables it are also used in the GSS/University of Chicago:National Organization for Research at the University of Chicago (NORC) and ALLBUS/Center for Survey Research and Methodology (ZUMA), Mannheim surveys. The Cumulative Codebook also includes copies of the original data collection forms: the PGSS questionnaire, and the questionnaire for the ISSP topical modules. This is helpful for researchers, who often want to know the context in which a particular question was asked, and to see the survey instrument as a whole.
[73] Up to 1995 the PGSS Cumulative Codebook (in Polish and English) was distributed in both hard copy form (about 500 pages) and in computer-readable form (ASCII text). Since 1997 it is available in electronic form (MS Word format) only. Both forms of documentation are available in Polish and English.
[74] Due to the accumulation of data from subsequent editions of PGSS, data files have become quite large. The PGSS 1992-1999 dataset includes 11192 respondents and 1117 variables. To facilitate the use of this large data file, in 1997 a number of new solutions were introduced concerning archiving, distribution and searching.
26 April 2002
