i RIGHTS OF RESEARCHERS AND GOVERNMENTS TO NATIONAL RECORDS WHO OWNS CONTRACT AND GRANT DATA AND WHO CAN USE IT? BY Ekkehard Mochmann Zent ral archi V fur Empirische Sozi al forschung University of Cologne This article is reprinted with permission from the editors of EUROPEAN POLITICAL :DATA NEWSLETTER. I. Freedom of research, but no access to data? The legal context for social science data access in Germany is significantly different from the American situation. While the federal and state data protec- tion laws have been enacted in the late seventies (like in most European countries), a general data access regulation as the equally important component of information legislation is still lacking. In this situation Article 5 of the German Federal Republic's Constitution is referenced as the most authoritative written norm. It guarantees freedom of arts and science, research and teaching in general, but does not say anything specific about data access. The interpretation of this article by courts and experts, however, acknowledges in principle the right to information access. This position is contrasted by the actual behaviour of the German Adminis- tration. An orientation to keep information under its control is prevailing (1). In practice it is the researcher who has to justify the information request and has to prove that he cannot achieve his results by other means. There is no regulation demanding from the Administration to justify its refusal. On the con- trary the researcher has to convince the data holding administration of the impor- :tance and legitimacy of his research intention. In short: There is no equivalent to the Freedom of Information Act. Before privacy legislation was enacted, the legal situation in Western Germany vas reasonably well characterized by the statement "That the owner of data-- including personal data--was more or less regarded as proprietor who could dispose Df them as long as he did not violate the rights of other persons" (2). In terms jf data accessibility (not protection) privacy legislation changed the situation :o the worse: It was frequently misused as an argument to prevent access--even in :ases where privacy legislation did not apply. ! Access to data from the Federal Statistical Agency The mandate of the Federal Statistical Agency is (among others) to provide juantitative Information about social and economic development of the society, 'he same can be said for the state and commune agencies. More than sixty legal norms regulate the procedure for more than two-hundred separate statistical counts, which have to be conducted. Whenever a special legal norm is enacted, it specifies all details for the survey (sample, variables) including access and disposition rights. These normally rest with the office, which initiated the data collection. There are cases, in v\;hich statistical offices of the communes collected the data for the state office, but were not entitled to analyze the data themselves, either before or after transfer to the state office. in other cases public access was explicitely guaranteed. Whether access actually can be achieved is frequently a question of the fees which have to be paid for usage (e.g., up to DM 6000 for the copy of a tape from a \% sample of the Micro census). Apart from these specific regulations the general procedure for the Federal Agency was defined in the Federal Statistics Act in 1953 which was revised in I98O. This new version responded to problems arising from recent privacy legislation. According to the new regulation the statistics can be used by scientific institutions and other interested bodies, only if the data is anonymized. The data flow is seriously hampered by the fact that appropriate anonymi zat ion procedures have to be developed and implemented (3). In fact, given the problems of anonymi zat i on , even public use files are not available. On the other hand the Federal Statistic Agency offers the services of STATIS-BUND. This is a network-oriented service for access to data and appropriate statistical procedures. You can analyze the data in the bank, you can bring in additional data which you had collected yourself, but you cannot transfer the data to your own installation (k) . 3. Access to administrative (process-produced) data The transfer of anonymous data is permissible, transfer of identifiable is restricted. Again data can be transferred only if the researcher can convince the Agency and data protection commissioner that the research interests are considerably higher than contradicting interests and that the research goal cannot be achieved by other means. Basically the data access is under complete control of the Admi n i St rat i on . h. Access to contract data Government agencies have a high demand of data, but hardly any personnel resources for data collection or data analysis. These activities are usually contracted out. Given this situation they are not interested to lay their hands on the data itself. Their needs are satisfied when they receive the research report and tables. There are hardly any resources for data analysis in the offices of the Aermi ni st rat i on . As a consequence no attention is being paid to clarify the rights regarding the data in the contract. Only recently attempts to alert the responsible administrators to the fact that publicly financed data collections are an important resource for secondary analysis are gaining increasing attention. Let me characterize the situation regarding access to cross-section surveys by two contrasting experiences. Since the first story given an example of very questionable performance in a critical political situation, it may be particularly important for discussion. Nevertheless I will protect the identity of this office, since I cannot give a fair account of the detailed arguments here. In 1979 one of the more prominent offices of the Federal Administration asked a commercial research institute to contract a study on the right-wing radical potential. This research team had contracted the field work for the cross-national survey to one of the leading opinion research institutes. The results which were finally reported in the media were substantially contra- dicting the findings of a prominent German sociologist, who had contributed to most important research findings in this field. Of course he wanted to reanalyze this new data set. We asked the financee for access to the data. The response was posi- tive, but conditional on the agreement of the contract institute. This agency was positive too, but conditional on the o.k. of the sub-contractor. This was very positive, but we did not receive the data. After several i terat ions--and even poli- tical i ntervent ions--we were informed by the Government Office that data transfer seemed not to be advisable in this given situation. Almost parallel to that we were informed by the contractor that the interested researchers certainly could inspect the data in the contractor's office; apart from that the results would be published on the book market in short. The book is available, we are still waiting for the data. One positive experience stems from negotiations with the State of Horthrhine- Westfalia and the Federal Post Minister's Office. They currently are conducting implementation and evaluation studies for the two-way communication system, which is called B i 1 dsch i rmtext in Germany (Prestel in England, Telidon in Canada and Antiope in France, just to name a few). The contracts with the research insti- tutes clearly define that all rights regarding the data rest with the financees, and they are interested in an intensive usage of these data sets. To prove this: We have produced the codebook and a character data set for two of the major studies al ready (5) • Likewise we received and distributed three big data sets from the Federal Labour Minister's office with results from recent studies of unemployment to quote just another positive experience. I could continue with an amazing example from a postord i nated agency (nachgeordnete Bundesbehbrde) , which holds numerous labour market data sets. The most positive declaration to start data transfer to the social science community via the Zent ra larchi V was unfortunately restricted by including a privacy clause, which explicitly stated that data can only be used for the purposes of this office. It might have been phrased slightly different to open access to the social science community in general. But I prefer to follow the positive terms described before. 5. Federal Archives act A law for the Federal Archives (Bundesarchi v) has been drafted. There was none in existence before. Reactions to privacy legislations promoted the idea of a Federal Archives Act. This draft explicitely offers access to research--that , with some exceptions, can wait 30 to 120 or even I50 years. I understand that these regulations are similar to those in many other countries. They can be interesting for historical research, they hardly will be useful for empirical social research. All materials produced or received by federal agencies, will be subject to this law unless considered not worth archiving. Nothing is said about contract data. All materials have to be offered to the Federal Archive as soon as they are no longer needed in the public administration. A decentralized principle is followed in so far as materials from local or state agencies can be sorted in their respec- tive archives. Nothing is said about contract data. This point, however, is tapped in the Swiss neighbour's "Guidelines for processing personal data in the federal admin i strata! on". In case a state (Kanton) or commune, a private person or organiza- tion is given a contract, data protection rules have to be specified by contract or order and have to be supervi sed-- i f possible. Nothing is said about access regula- tions, howeve r (6) . 6. Access to grant data This situation can be characterized very quickly. Besides research foundations. Government agencies give a significant amount of funds in form of research grants. As a rule the data collected under these grants is under complete control of the researcher who collected it (of course subject to current privacy and other legis- lation). Unfortunately the practice of data sharing in the research community is still less popular than arguing against data protection and restricted data access regul at ions . 7. Result I There is no clear answer to the question of ownership and access to contract data. Discussion of negative effects of privacy legislation and positive experi- ences with special access regulations as well as references to the international development gradually develop the feeling for the importance of this subject in our research community. The experts' discussion about the topics is reported in spe- cialized journals and monographs. Wh i le the State of Hessen was first to pass a privacy protection law we have to catch up with respect to data access regulations (7). Part of the game is to balance political, commercial and research interests. REFERENCES J. Scherer "Datenzuqang des Forschers zwischen I nformat ionsanspruch and Gehe imha 1 tungsgrundsatz", in: M. Kaase ; H.J. Krupp; M. Pflanz; E.K. Scheuch; S. Simitis (eds.): Datenzugang und Datenschutz, Konigstein 1980, p. 96. U. Dammann und R. Brennecke: "Country Report Federal Republic of Germany", in: E. Mochmann, P.J. Muller (eds.): Data Protection and Social Science Research, Frankfurt 1979, p. 129- B. Pohl: "Datenschutz in de r amt 1 i chen Statistik", in: Datenschutz und Datens i cherung , Volume 2, 1981, p. 69- (k) Before access is permitted you have to sign a contract specifying access conditions and fees. Descriptive information about STATISBUND, its holdings and program libraries are available from "Stat i st i ches Bundesamt, Wiesbaden" (5) i nf ratest-Nul lerhebung im Rahmen der wi ssenschaft 1 i chen Beglei t forschung zu Bi 1 dschi rmtext and Bi 1 dschi rmtext-Voruntersuchung (Langenbucher , Scheuch, Treinen, Date collection by Infratest Munchen) (6) R.J. Schweizer "Die Richtlinien des schwe i zeri schen Buncesrates uber den Datenschutz in der Bundesverwal tung", in: Datenschutz und Datens i cherung. Vol. ^, 1981, pp. 239-2'i2, and "Richtlinien fur die Bearbeitung von Pe rsonal daten in der Bundesverwa 1 tung" , paragraph 3^*. in: Datenschutz und Datensi cherung. Vol. ^4, I98I, p. 2't3- The German Science Foundation has implemented a committee on Privacy protection and Data access problems. see also. M. Kasse; H.J. Krupp; M. Pflanz, E. K. Scheuch; S. Simitis (eds.): "Datenzugang, Datenschutz", Konigstein I98O.