The National Data Center proposals between macro modelling and micro targeting

Abstract As computer technology got more advanced during the 1960s, scientists used it to build economic models, run simulations and make predictions. But the gist of any method was a solid data base. Therefore, a committee of the Social Science Research Council developed a proposal for a Federal Data Center within the United States government. The Bureau of the Budget developed the idea further into a National Data Center proposal provoking a well-known debate on privacy. Less is known about how the proposals originally emerged. One goal was the access to microdata on individual units which were necessary for statistical operations such as correlation or matching of data sets. The article argues that the data center demonstrates the shift towards statistics based on micro units to build macro models. Scientists, however, faced obstacles such as rules concerning confidentiality and disclosure of data as well as a decentral structure of federal statistics, and their interests differed from the interests of politics.


PUBLIC INTEREST STATEMENT
Personal data is sometimes dubbed as the new oil.It has a commercial value for companies to target customers, it is valuable for politicians to create efficient programs, and it is essential for social scientists to study relations in society.Yet the gathering of large amounts of data may conflict with individual rights such as privacy.The question how to strike a balance between these competing interests can be traced back to the 1960s.Back then, at the brink of the computer age, a small group of scientists drafted a proposal to store data tapes by US federal agencies centrally and make them accessible.One motivation was to improve economic models using statistical methods like correlations.When government officials took up the proposal, a debate on privacy emerged in Congress.This debate is still relevant today, in the context of big data and artificial intelligence.

Introduction
In our information society there is a need for accurate data, data that improves statistical analysis and political decisions.Yet there is an underlying conflict of how to balance interests of access to and confidentiality of data.Historically, this conflict reaches back to the 1960s when scientists started using computers to create economic models, apply statistical methods and run simulations.Scientists at the time strived to make predictions, companies tried to influence elections and to forecast riots (Lepore, 2020b, p. 349f.).Altogether, the late 1960s and early 1970s were "the high point of future research" (Andersson, 2018, p. 3).During President Lyndon Johnson's War on Poverty, politicians tried to tackle problems of poverty by analyzing its causes and developing its cure with scientific methods.As a result, the knowledge about poverty became a technical and professional discipline: "Besides being social scientific, this knowledge is based principally on quantitative, national-level data" (O 'Connor, 2001, p. 4).These approaches had one thing in common: They relied on detailed information.The advancement in data processing, statistics and the advent of the computer in the 1950s and 1960s went hand in hand with a rising awareness of data protection.This debate on data protection is relevant today as governments are struggling how to regulate the handling of data in the age of the internet.While, for instance, the European Union implemented a relatively strong General Data Protection Regulation of 2018 (Taal, 2022, pp.4-7), the United States relies on relatively weak rules.Recently, in July 2023, the EU and the US agreed on a Data Privacy Framework (European Commission, 2023), but it remains to be seen whether the agreement will be upheld by the courts.
In the United States, regulation such as the Privacy Act are mainly a result of a debate that arose during the 1960s and 1970s.One origin of the privacy debate was a project of economists and social scientists to pool governmental data in one center (Cappello, 2019, p. 187-192; Frohman, 2021, p. 75 n89; Igo, 2018, pp.221-23).As Priscilla Regan has argued, this debate about a centralized data facility brought issues of "privacy, computers, and government information practice" on the political agenda (Regan, 1995, p. 73).Already at the time, these plans served as key witness that computers could have a negative effect on individual rights.For instance, legal scholar Alan F. Westin referred to the data center in his seminal study Privacy and Freedom (Westin, 1967, p. 317).Recent research on the National Data Center focuses on statistical and scientific aspects.Rebecca S. Kraus, historian at the U.S. Census Bureau, describes the work of governmental agencies and poses the question which conclusions may be drawn to improve statistics: "What can we learn from 1965 that can help us form sound statistical information policies in the 21st century?" (Kraus, 2013, p. 1).Also, Jill Lepore discusses the proposals from the perspective of the present, of "the data-driven twenty-first century" (Lepore, 2020a, p. 279ff.).Likewise, Dan Bouk refers to the National Data Center with regard to a "history of the political economy of data" (Bouk, 2018, p. 628).From a social science perspective, Christopher Loughnane and William Aspray analyze the proposals with regard to archives, and they are interested in the motives of social scientists to suggest a data center, asking: "why the need for centralization was felt so urgently within the social science community" (Loughnane & Aspray, 2018, p. 203).This article may contribute to answering that question.In summary, the latest research on the National Data Center proposals focusses on the implications for data processing and statistics and points out the interests of agencies and scientists.
The process of how the proposals for a central data facility evolved, however, has not yet been analyzed.The question is how the idea of a data center came about and why it failed.What were the scientific premises, how was the institutional setting characterized, which problems did the statisticians intend to solve, which problems did they face?In this article, I elaborate three hypotheses on the emergence and failure of the project.The first hypothesis assumes that it failed due to privacy concerns.In this regard, Arthur R. Miller has discussed the proposals in Assault on Privacy (Miller, 1971, pp.54-66).The second hypothesis focuses on the decentral structure of the federal system that could not easily be turned into a central structure.This aspect of centralization versus decentralization has been discussed by Rein Turn et al., for instance (Turn  et al., 1976, p. 1).The third hypothesis deals with competing interests of the scientific community and the bureaucracy.
The article analyzes, on the one hand, how scientists reached the conclusion to centralize governmental data and by which scientific goals they were driven.On the other hand, the article takes a look at how governmental officials dealt with scientific proposals and points out the networks between the two communities.Methodologically, the article draws on Bruno Latour following scientists in society: "our entry into science and technology will be through the back door of science in the making" (Latour, 2003, p. 4).History needs to put technology into its social context; it needs to consider the statistical methods, the data resources and possible implementations at the time.From the perspective of a history of technology, statistics evolved from macroeconomic analysis to correlations in-between individual data sets with linked attributes.In this article, I argue that the idea of a data center was driven by new possibilities of analyzing interdependencies within sets of microdata.
In order to investigate the making of science, to answer the question and to address the hypotheses, this article draws on primary sources of institutions that were involved in the project.Documents of the Committee on the Preservation and Use of Economic Data (CPUED) funded by the Social Science Research Council (SSRC) show how scientists developed the report that suggested a Federal Data Center. 1 In the following, the Bureau of the Budget (BOB) took up the proposal and commissioned an external study that, in 1965, recommended a National Data Center.Afterwards, the executive office set up another external task force that once again recommended a National Data Center as well as an internal task force that worked on a proposal for a National Statistical Data Center. 2In addition, the article analyzes how the White House treated the question of how to handle economic data. 3Most studies on the National Data Center proposals only refer to the published hearings of several congressional committees that dealt with the proposals.Lately, Kraus drew on primary sources of the BOB but did not give the exact location (Kraus, 2013, p. 35).These documents correspond to some of the material used in this article.
The article comprises of five chapters.The first chapter discusses the paradigm shift in statistics towards analysis based on microdata and correlation.Chapter two describes how economists and social scientists requested microdata from several agencies and wanted to improve storage of information.Soon a committee of scientists developed the idea to install a Federal Data Center to manage the storage and usage of data, as chapter three shows.Some government officials were attracted by the idea to centralize the federal statistics, as chapter four analyzes, but Congressional hearings resulted in a debate on privacy.Meanwhile, subject to chapter five, government officials set up an internal task force to develop the proposals further.Yet other government officials followed another approach of gradual reform.The proposals eventually failed, while the debate on the limits of privacy and efficient program planning continued.The episode sheds light on how to balance interests with regard to the analysis of personal data, a debate that is still crucial, since 1960s-style modelling and planning gave way to approaches of big data and artificial intelligence that often target individuals.

Paradigm Shift In Statistics
Neither the theory of statistics nor the utilization of statistical methods by government agencies was invented in the 1960s.The starting point of a "history of quantification" might be traced back to around 1830 (Porter, 1996, p. xi), and the "practice of measurement" took shape in 1840 (Hacking, 2009, p. 5).While in the nineteenth century the print of numbers gained importance (Hacking, 2009, p. 3), the mid-twentieth century witnessed another decisive innovation: computers processing digital numbers.Computers facilitated calculations and allowed statistics of large surveys based on individual entities, so-called microdata.The shift towards quantitative methods was accompanied by a quest for data, data on which statistical modelling or cost-benefit analysis relied.The latter had been applied by economists by the mid-1960s to diverse fields such as infrastructure projects or city planning and was regarded as a mean for a rational evaluation, even though the data basis may have sometimes been weak: "Perhaps the available data were less than ideal for some measures" (Porter, 1996, p. 188).Access to detailed data therefore became a main concern for social scientists and economics, and it also became part of business models.In the United Kingdom, for instance, social scientists planned to set up a data facility in 1963 and 1964 which was later called the Social Science Research Council Data Bank.Scientists wanted to collect surveys by research, industry as well as government facilities, but they were confronted with problems: "One serious obstacle to the SSRC Data Bank requisitioning all old government surveys was the assurance of confidentiality of information given to contributors" (Agar, 2003, p. 359).Access to information was restricted.
One US company creating simulations had to struggle with the technological shortcomings during the 1960s, with limited computer power, rudimentary methods and first and foremost a scarcity of data: "Data was scarce.Models were weak.Computers were slow" (Lepore, 2020a, p. 321).The company became subject of public debate on the impact of computers.Social critic Vance Packard later commented in The naked society on the episode asking: "whether by 1984 power might be held by the party with the best computers, or the one most adept at exploiting information stored in computers" (Packard, 1964, p. 202).This comment illustrates the discomfort of contemporary critics with regard to computer technology.Usually, economic models that were based on theories of John M. Keynes, for instance, relied on aggregated data sets (Desrosières,  2011, p. 209, 333).Besides Keynesianism, a new paradigm emerged during the 1960s: "the War on Poverty led the way to a 'microdata' (household and individual level) revolution" (O 'Connor, 2001,  p. 182).The scientific background for this revolution, however, was developed earlier than the programs of the War on Poverty.Compiling samples was one option to provide data on a small scale.For instance, Conrad Taeuber, Assistant Director of the Census, told the SSRC that the Bureau was planning to publish a nation-wide sample, while complying with rules of confidentiality and preventing the possible identification of specific persons: "The plan is to offer to make the tapes containing the information for each person within the 60,000 households available on a cost basis." 4The advent of the computer and the quest for microdata were driving forces behind a project later known as the National Data Center.

A Committee Requests Machine-Readable Microdata
The story began in the year 1960 when scientists wanted to improve the storage and usage of research data.On 25 and 26 March 1960, the Executive Committee of the American Economic Association (AEA) made a reservation to set up a committee regarding a possible data facility "until the dimensions of the project could be thoroughly explored," suggested that economist Richard Ruggles, assisted by qualified personnel, should elaborate these dimensions, and the AEA requested an exploratory study. 5A couple of days later, in April 1960, Ruggles, Ford Foundation, told Paul Webbink of the SSRC in a letter that he had discussed "the question of statistical data libraries" with colleagues. 6Attached to the letter Ruggles sent a "Proposal for an Exploratory Committee on Data Repositories in Economics," in which he considered, first, how scholars could learn about existing research data sets by the government, research institutions or foundations, data that would otherwise remain unused or needed to be replicated.Second, he demanded facilities for discarded data that was no longer utilized: "no group at the present time seems to be in a position to handle any significant amount of statistical data generated as the by-product of government or business operations." 7Third, research data of universities could be managed centrally. 8This proposal summarized the principal goals that scientists pursued in the following years, demanding not only data explicitly collected for research purposes but any kind of data produced by the government or industry.
In order to address the aspects of access to data and its storage, the SSRC organized a conference.At the "Conference on Preservation and Use of Economic Data" taking place on 19 December 1960 at the SSRC in New York, participants from universities, foundations, research institutions as well as from federal agencies discussed the availability of research data, preferably microdata: "The symptom that is probably most responsible for the calling of a conference on data preservation is the scarcity, for research purposes, of good data on individual units." 9The discussion focused on the federal government because participants hoped for standardized data sets collected on a national scale. 10A common technology to store machine-readable data was magnetic tape, other data could first be stored on microfilm.Beyond the question of storage, researchers wanted to access data immediately. 11Furthermore, participants discussed the problem that data held by the government was often marked as confidential.To solve this problem, either government officials needed to handle data, or scientists needed to conduct research within the respective agencies.Another possible solution was managing information in "data centers" with computing resources: "Also the problem of disclosure could be met by placing all processing of primary data under the supervision of authorized personnel." 12One operation to make data more valuable was to link the individual units of separate data sets, a costly "matching operation." 13Participants discussed to install "several data depositories" 14 and generally agreed that "action is needed." 15During the year 1960, scientists established ties with each other and laid the groundwork for a study on socioeconomic data.The starting point was a focus on the federal government and the idea to install certain facilities.Yet the problem that data could not easily be disclosed to any research facility was present from day one.
In 1961, a committee was established, and it soon faced difficulties.Government officials not involved with the project favored a cautious approach of starting a narrow trial.In response to the conference report, Ross Eckler, Acting Director of the Bureau of the Census (BOC), advised to start the project on a small scale and to subsequently evaluate problems and discuss technologies, suggesting a "piece-meal approach." 16In March 1961, Webbink presented the conference report to the SSRC Committee on Problems and Policy (P&P), pointing out to a "technological revolution" since the early 1950s when the advent of computers and magnetic tape allowed for the production of vast amounts of data that could be preserved for research purposes. 17Eventually, the SSRC approved an "Exploratory Committee on the Preservation and Use of Economic Data" led by Richard Ruggles of Yale University.Further members were Ernest Enquist of the Internal Revenue Service (IRS), Lebergott of the BOB, Guy Orcutt of the University of Wisconsin as well as Joseph Pechman of the National Committee on Government Finance, 18 which belonged to the Brookings Institutions.Scientists from the academia, from research institutions and from federal agencies cooperated to negotiate the interests of science and politics concerning research data.The committee soon forgot about its exploratory nature and the term "Exploratory" was dropped altogether.
The committee focused exclusively on data of federal agencies and, at first, decided to request data of agencies case by case, but it had to address the question of disclosure: "There is considerable need for a general set of rules which would meet the research requirements of scholars and at the same time protect the agencies involved." 19The committee members listed a variety of statistical data sets to be obtained from federal agencies and suggested to install a "Committee on Access and Disclosure" led by the SSRC, located at the Brookings Institution in Washington, DC, a committee that was supposed to fulfil the function of liaison and coordination with agencies. 20Solutions to access data for research purposes existed only on a small scale.For example, the Census Bureau published a sample about 60.000 households, and the Brookings Institutions cooperated with the IRS to create "tax model" tapes from samples of roughly 100.000 tax returns. 21The committee members noticed a growing demand especially "for general purpose sets of data which provide information at the micro level," data sets on individual entities, households or companies. 22But access to such microdata was restricted.Difficulties obtaining data remained, as the CPUED report of September 1961 said: "The committee recognizes that problems of disclosure may arise in certain areas of government statistics but it feels that various devices can be utilized to meet this problem." 23There were basically two approaches concerning disclosure: Either identifying features had to be removed from data in "special tabulations" or, if that was not possible, calculations planned by researchers had to be run by government officials. 24In 1961, the exploratory committee dealt with difficulties accessing research data in machinereadable form, preferably non-aggregated data.Yet there was no general solution in sight.Specific projects required specific approaches.
During the year 1962, members of the committee began lobbying in Washington, DC, and realized that it was a difficult task to acquire data on magnetic tape from agencies.The committee chose an "ad hoc" approach focusing on specific samples and surveys of the BOC, the IRS respectively the Bureau of Labor Statistics (BLS). 25In particular, the BLS attracted the committee's attention since it planned to outsource computing activities and to create a "data center", though the plans were still in an early stage.First and foremost, such a center had to deal with disclosure rules: "Since disclosure apparently is by far the largest stumbling block, even larger than money, some kind of rules for data release (to avoid disclosure) might be appropriate." 26The committee did not ask how to comply with disclosure rules but thought about new rules to bypass these provisions.As Arnold Chase, BLS, explained further, a "data center" had to be installed outside the BLS for financial reasons, possibly at a university where two or three government employees would supervise the operations with regard to disclosure rules. 27Altogether, the committee reached only minor successes.For example, George Jaszi, Office of Business Economics (OBE), offered to put existing OBE publications as the "Survey of Current Business" or "Business Cycle Indicators" on magnetic tape for research purposes, and Raymond Bowman, Assistant Director, Office of Statistical Standards (OSS) at the BOB, told the committee that duplicate governmental tapes could be provided to researchers, except for restricted microdata. 28At this point in time, the committee might have come to a conclusion and put together the results about the options and hurdles to acquire research data in an exploratory report.But the committee kept going looking for ever bigger and more general solutions, following an ambitious approach of gathering data on a large scale.

The Idea of a Federal Data Center Emerges
The idea to establish a nationwide data center was born during the year 1963.Based on the statements of the BLS, the committee shifted the focus of its investigation and instead of requesting data case-by-case, it started searching for locations for possible data centers, where data would be provided and research could be conducted.Preferably, the main center was located in the area Washington, DC. 29 Agencies followed different projects for automated data processing.According to IRS officials, specific data on the taxation model were supposed to be filed and stored permanently on IBM or UNIVAC tapes.This kind of data could be made available for academic research, and the bureau accepted "direct payments for special tabulations of data," and according to the committee minutes, Ernest Enquist, IRS, regarded disclosure "to be only a small stumbling block." 30The IRS archived data in several "federal records centers" supervised by the National Archive and the General Service Administration. 31In addition, Richard Holton, Assistant Secretary, DOC, suggested to install research facilities within the Census, but he mentioned "problems of disclosure" concerning the idea to set up data facilities at universities outside of the federal agencies. 32Despite these problems, the current or planned data centers of federal agencies soon inspired the committee to planning for a data center, too.The creation of a specific facility was not necessarily the most promising choice to obtain data.Other research institutions were able to work with government data on a project basis for specific purposes.For instance, one researcher from Harvard University analyzed census data on large companies, he was therefore authorized by the Census, worked at a dependence and was supplied with an assistant. 33Yet the CPUED searched for a more general solution.In spring 1963, Edwin Kuh of the Massachusetts Institute of Technology joined the committee. 34Still, the CPUED did not reach a major success to convince federal agencies to release information other than as samples or in aggregated form.
The CPUED members were not satisfied with the situation of requesting data sets on a case-bycase basis because agencies did not have the resources to support economic research, and researchers carried the burden to file agreements with each agency separately. 35Therefore, the committee wanted to found a "Data Center", an institution that would also deal with disclosure since it became clear: "the problem of disclosure is a very real one." 36Data sets could be edited and identifying information removed, or government officials could perform the calculations: "the agency concerned is then able to check the results to insure that there is no violation of confidentiality." 37Agencies were supposed to transfer magnetic tape to the "Data Center" which would provide data to researchers in compliance with disclosure rules.Furthermore, the "Data Center" would function as an archive to store historic data. 38It was supposed to store the information of some hundred thousand of punch cards on several magnetic tapes and to oversee regional centers: "The central Data Center in Washington thus would provide the basic mechanism for supervision and control of data made available for research purposes." 39This sounded somewhat like wishful thinking.The question remained how the central facility would obtain data in the first place since problems of access remained.
Instead of a solution there was a growing perplexity.In its annual report 1962 to 1963, the CPUED considered that either a government official should be present in a possible statistical center or researchers should take an oath becoming officers themselves temporarily, suggestions that should prevent the identification of persons within economic data sets and that should tackle the disclosure problem. 40Besides, it remained unclear which data sets to preserve: "There is clearly no present agreement among economists, or among other social scientists, as to what data should be collected and kept available for extensive research." 41As a result, the committee wanted to store as much data as possible.The CPUED pursued the plan to install a central facility to provide data for research not because it was a suitable way to organize available data sets but because of administrative barriers.However, promising new statistical approaches may have been, the access to microdata was restricted by statutory law.The committee kept working on its proposal ignoring the obstacles.Rarely committee members like Guy Orcutt complained about an "apparent minimization of the problems involved in establishing a data center." 42Problems were well-known but no solution seemed practical: "The committee agreed that prying data loose from the generating agencies was a big step because of the disclosure problem." 43The CPUED identified the National Archives as a suitable institution to store and provide data tapes.Rules of confidentiality also applied to the National Archives, and as Wayne Grover informed the CPUED, researchers needed to address requests directly to the respective agencies concerning the "disclosure problem." 44But those agencies were not allowed to release confidential data either.The proposal of external data facilities seemed flawed from the very beginning.
Government officials frequently referred to problems of disclosure and rules of confidentiality.In this regard, agencies gave similar responses to the CPUED.Within the Social Security Administration, Ida Merriam, Director of the Division of Research and Statistics, who was "extremely sensitive to the disclosure problem," told the CPUED about a counseling board that decided about requests for data.William Levin, financial statistician of the Federal Trade Commission (FTC) showed interest in the project but refused to transfer any data to the archive. 45Actually, data centers could not solve problems to access data, but regulations concerning disclosure and confidentiality prevented agencies from transferring data to any center in the first place.An important step towards a center was creating an inventory of existing information within the federal government.Richard Ruggles told Raymond Bowman, OSS, that this inventory of punch cards and magnetic tapes was decisive in order to plan for the storage of data and the installation of a "federal data center." 46This letter contained the first mentioning of the term.Ruggles emphasized the need for microdata: "For a grat many analytic purposes, it is necessary to compute ratios and/or study behavior and characteristics at the individual respondent level prior to aggregation." 47Yet aggregated data was also valuable and supposed to be integrated in a center.- 48Altogether the requests by the CPUED for microdata remained unsuccessful.Any statistician dealing with microdata of federal agencies had to address two major obstacles, legal provisions of confidentiality and rules of disclosure restricting the access to data.It became a fixed idea to install a central data facility in order to overcome these obstacles.
The idea of a Federal Data Center developed further in 1964 when the BOB director suggested to set up a center within the government itself.In January, Paul Krueger, Assistant Chief OSS, sent a draft inventory that was compiled without evaluating the importance of data and included an almost complete list of existing tapes. 49According to Ross Eckler of the BOC, the bureau considered installing a separate organization for outside requests, but statutes of confidentiality still applied. 50In February 1964, Ruggles met with representatives of the National Archives and the BOB and considered the lists as thorough but not very detailed concerning the economic variables of the data.As the CPUED minutes of this meeting said: "There was unanimous agreement that a Federal Data "Computor" [sic] should be set up as soon as possible." 51Richard Holton of the Department of Commerce (DOC) told the CPUED about a board within the Census that was supposed to decide about requests of researchers, and he mentioned a possible "Commerce data library." 52The project of the CPUED seemed stalled.According to a preliminary committee report, problems persisted how "micro data" in machine-readable form could be provided for research purposes, because agencies had strong reservations to publish information that could reveal the identity of persons. 53From the agencies' point of view, an external data facility was hardly possible because of existing rules, rules that such a facility was originally supposed to solve.Setting up a non-governmental center for microdata seemed more and more unrealistic.
Soon, another problem besides disclosure became apparent, namely the problem of centralization.The federal government had various departments, agencies and offices dealing with statistics, two of which had a coordinating function, the DOC and the BOB.While officers of the BOB felt attracted by the idea of a data center, the Census within the DOC shared a different view.At a meeting at Brookings Institution Morris Hansen, BOC, objected a corresponding report to the board of the SSRC: "Census opposes placing a Federal Data Center in Archives," 54 and he said that the Census as a "general statistical agency" rejected installing a rivaling institution, while agencies needed to keep a certain control over data.Hansen pointed out that the Census already published the "Statistical Abstract" covering various agencies. 55Yet Kermit Gordon, director of the BOB, suggested Ruggles to draft a bill implementing a separate institution within the federal government, a "central statistical service," getting beyond the structure of the Archives, the Census and the OSS. 56In a letter of November 1964, Ruggles told Gordon that the committee acted on the suggestion to propagate a special institution to support agencies as well as researchers: "the committee decided that it should take a forthright stand on the need for a Federal Data Center which would have access to all of the machine readable data produced in the Federal government." 57This was the birth hour of the FDC idea.Since the agencies would not release any data to an external institution, scientists wanted to set up a center within the government, an idea that was encouraged by officials of the BOB.
Finally, the CPUED, once established to prepare an exploratory study, submitted a report in early 1965 that proposed a large-scale project to the government.This incongruity was noted by the SSRC when Paul Webbink sent his comments on the draft report to Ruggles, suggesting not to use the term "recommendation" but to draw a conclusion instead: "Somehow I am squeamish about the ultimate impact of a 'recommendation' of a non-governmental committee to an important government Bureau." 58Finally, the SSRC approved the final report.In the final report, the authors criticized the decentralized statistical services and proposed to the BOB to install a central data facility: "Such a Federal Data Center should have the authority to obtain computer tapes and other machine readable data produced by all Federal agencies", a facility that was supposed to make data available to researchers, without however violating rules of disclosure. 59It is fair to say that the report went beyond an explanatory mandate, requesting data on magnetic tapes for an FDC from every federal agency.
In response to the report, Ewan Clague, Commissioner of the BLS, pointed out problems of keeping data up to date: "Newly developed data, corrections to existing data, and revisions to benchmarks which have become available at a later date, all create problems and the additional workload will soon become large since they are all channeled into a central data bank." 60urthermore, he questioned that organizations outside of the government could take the responsibility for confidentiality pledge even if data was transferred, but said that the final report of the CPUED did not sufficiently consider the aspect of confidentiality: "I doubt that such information can be made available to any data bank external to the original agency receiving the data in confidence." 61Kermit Gordon, BOB, told Webbink that he supported the project: "In earlier conversation with Dick Ruggles, I have expressed my personal interest in the idea of a Federal Data Center." 62It remains uncertain whether or not Gordon made the suggestion to create a Federal Data Center.The concept of centralization was more a reflection of the wishes of the scientific community than of needs of federal departments and agencies.

The Federal Government Takes Up the Proposal
Since the Progressive Era, an increasing centralization fostered by government experts was subject to debates (Porter, 1996, p. 151).Federal statistics were no exception of this trend.The decentral structure of federal statistics had grown historically, and it met the specific needs of various departments and agencies for data: "most recommendations have favored the existing decentralized organization of the U.S. Federal Statistical System at least in broad concept" (Duncan &  Shelton, 1978, p. 160).A report of the Central Statistical Board (CBS) (later OSS) of 1934, said that decentral units were able to address special needs for statistics.To the contrary, in 1949, authors of a task-force report to Congresscomplained about overlapping activities and recommended that the collection and analysis of statistics should be concentrated in the Census (Duncan & Shelton,  1978, pp.145-69).The history of the federal statistics demonstrates how a double structure between BOB and DOC had emerged.The subordinated OSS respectively the BOC were interested to concentrate statistical functions without, however, giving in control over information.
During the year 1965, government officials developed the proposal of a Federal Data Center further.Soon after receiving the report, the administration took up the idea and pursued the plans for a data center, and the BOB commissioned Edgar S. Dunn, Resources for the Future, Inc., to prepare another study on how to integrate federal statistics. 63At this point in time, Dunn and the company were already familiar with the project.For example, the company was listed as a participant at a conference of the SSRC on data administration as early as December 1960. 64n January 1964, Dunn, Deputy Assistant, Secretary for Economic Affairs, DOC, suggested that a "user service facility" could deal with disclosure. 65Later in that year, Dunn took part at a meeting of the CPUED as a member of the American Statistical Association (ASA) in October 1964. 66ltogether, Dunn seemed to be involved in the project already, and as a result, his report did not promise an entirely new perspective.
Proponents of a data facility, of course, were hoping that centralized statistics would help the federal government with planning and executing its programs.For instance, Raymond Bowman of the OSS believed that "the information services being proposed have significant payoffs for the public decision-making process." 67There were a number of questions such as "the disclosure problem" that could be addressed by editing data with computers. 68Bowman considered "some degree of centralization as desirable." 69Within the administration, the lead elaborating a possible data center was taken by the OSS, an office that belonged to the BOB controlled by the Executive Office.Therefore, reports to the BOB stressed the advantages of a data center for research as well as policy making.In his report of 1 November 1965, Dunn elaborated proposals for a "National Data Center" and suggested to establish a "National Data Service Center" to make data available (U.S. House of Representatives, 1966, p. 267).Soon the project was transferred to higher levels of the administration (Kraus, 2013, p. 8ff.).Government officials tried to adopt the scientific proposals to the administrative needs.Centralized statistics were regarded as a tool for policy making.Officials of the National Bureau of Standards elaborated the design of a data center in a report of November 1965 aiming at a "better understanding of interdependencies within our pluralistic society, leading to better informed choices among alternative policies and programs, and more effective program implementation" (Glaser et al., 1967, p. 20).In this progressive view policy could be improved with statistical analysis by understanding how social phenomena were interrelated.One only had to tie all the data knobs together.
In November 1965, Charles Schultze, director of the BOB, suggested to President Johnson to initiate a task force.The task force should elaborate a "Federal Data Center" and address problems of archiving data, referencing information not yet published, accessing data, and utilizing records of agencies. 70According to Schultze, computers would become essential to process federal statistics.Therefore, Schultze advertised the creation of a data facility: "A modern Federal Data Center would also increase the effectiveness of planning for the Great Society programs." 71ccording to this position, a central facility was needed to provide the data necessary to address social and economic problems.One month later, in December 1965, the White House established a task force, led by Carl Kaysen, Harvard University, to deal with storage of and access to governmental statistics.The White House had changed the list of members before, a representative of IBM was substituted with members from Berkeley and Los Angeles.As a handwritten comment by Johnson said: "These are all eastern.Tell him to provide us with a statement.We'll issue it down here," 72 noticing that the suggested members were all from the East coast.Apparently, members were recruited from a small circle of experts.One of the members was Ruggles who first proposed a FDC.Kaysen and Ruggles knew each other from before, as, for instance, Ruggles considered contacting Kaysen in June 1965 to discuss the project. 73It seemed likely that the task force would reach similar conclusions, in terms of centralization versus decentralization, as the initial report of March 1965, since the lead author of the former report became a member of the new task force and knew its chairman well.
The Kaysen task force started working on a report how to reorganize the federal statistics, but some government officials criticized the idea to turn the decentral setting into a central structure."I do not share Kaysen's view that complete or almost complete centralization of the collection of statistics is the ideal solution," wrote Raymond Bowman in March 1966. 74The question was how to improve and interrelate statistics in order to make machine-readable available "for both macro and micro analysis." 75Bowman suggested either a "Central Statistical Agency" as an independent body under the auspices of the Executive Office or a "Federal Data Center" set up within an existing agency.Yet he criticized the proposal and said: "Centralization per se will not suffice." 76he question was how to deal with the various statistical services that existed within the government.It seemed unlikely that departments and agencies were willing to give up competences and transfer them to a different institution.Kaysen submitted the task-force report to the BOB on 21 October 1966, suggesting to create within the Executive Office the position of a director for federal statistics, who supervised a National Data Center, and suggesting to implement standards for disclosure (U.S. Senate, 1967a, pp.33-36).Noteworthy, the report provided an important function to the Executive Office which was the very sponsor of the task force.Former assistant director of the Census, Stuart Rice, now with the Survey & Research Cooperation, disagreed with the Kaysen report that demanded centralization because federal statistics had become more complex, while past studies had concluded the opposite, pointing out the limits of centralization. 77This goal of centralization raised suspicion of members of Congress.
Meanwhile, Congress started investigating the proposals of consultant Dunn and criticized an alleged invasion of privacy (Kraus, 2013, pp.10-13).As early as May 1966, US News & World Report published a story on data banks and a possible "Government dossier bank" (Editor, 1966, p. 58).The article quoted Senator Edward Long, Democrat from Missouri, who held hearings on privacy at the time.He took a critical stance on the data center that "would be very undesirable for the American people, and would certainly result in great violation of their privacy" (U.S. Senate, 1967b,  p. 2398).The opinion of the press was mainly negative (Kraus, 2013, pp.13-17).The authors of reports on a Federal or National Data Center tried to defend the idea against critics.In hearings led by Congressman Cornelius Gallagher in July 1966, Ruggles emphasized that an analysis of individual data sets could have many benefits without, however, disclosing the identity of any person: "First, and foremost, it is essential to protect the individual from an invasion of his privacy and the misuse of information which may damage or embarrass him," said Ruggles (U.S. House of Representatives, 1966, p. 91f.).Several politicians made inquiries regarding the matter.For instance, Paul Krueger, Assistant Chief OSS, told Senator Walter Mondale, Democrat form Minnesota, that a "Federal statistical data center" was supposed to serve for statistical purposes, not to compile any dossiers. 78This criticism should not have taken experts and officials by surprise, since they faced ongoing problems concerning confidentiality and disclosure rules from the start, but the normative turn caught them on the wrong foot.In response to negative reviews of earlier data center proposals, the authors of the Kaysen report addressed the privacy problem in a separate chapter.These problems needed to be taken seriously, according to Kaysen, but they should not block advantages of technology for efficiency. 79There was an ostensible tradeoff between statistics and privacy.How to reap the benefits of technology without an outright centralization and integration of personal files was rarely discussed.

The Government Sets Up an Internal Task Force
The proposal for centralizing the federal statistical services originated in the scientific community.As a result, the purpose of a center and its benefits for the functioning of government remained vague.The National Data Center proposals have been portraited as a corner stone for the data processing of the Great Society (Lepore, 2020a, p. 279).But were they?Despite a growing interest in statistics and computer technology during the Johnson administration, there was no real need for such a center.The Johnson administration followed the idea of a Great Society, standing in the tradition of the Progressive Era and the New Deal.The "philosophy of liberalism" during the 1960s tried to detect and mitigate the causes of social problems (Milkis, 2005, p. 3).Poverty was one of the problems politics wanted to cure: "Like the New Deal, the Great Society tamed capitalism and secured a modicum of political peace" (Zeitz, 2018, p. 316).Even though the competences of the federal government grew, the reforms of the 1960s may not be regarded as "creeping socialism" as conservative critics did (Woods, 2016, p. 391).Over time, technology gained importance for the functioning of the government.As early as the 19 th century, Census applied the punch card machine of Herman Hollerith for the "mechanization of information" (Austrian, 1982, p. 4), and in the early 1950s, Census purchased a UNIVAC computer bringing data processing to a new level (Campbell-Kelly & Aspray, 2000, p. 121).In the late 1960s, computers had been established as an important tool within the government, as a BOB report pointed out. 80At the same time, agencies used more and more statistical information.From 1964 to 1966, the budget for federal statistics rose about 12 percent each year, as Paul Krueger calculated, reflecting a growing interest in "Aside from generally accelerating demands for more and better information as a basis for decision making, increasing emphasis on social programs gets us into areas in which data collection is expensive." 81A growing interest in statistics, however, did not imply that a central management was necessary.
The administration was undecided where to place a data center within the dual responsibility regarding federal statistics.This problem remained when the BOB started working on an internal proposal for a data center.In October 1966, Joseph Califano, Special Assistant to the President, entrusted Director Schultze with an in-house task force to elaborate how collection and analysis of economic data were organized and how statistics could be improved. 82In the following, Schultze contacted officials of commerce, census, finance and tax, agriculture, labor as well as the Council of Economic Advisers. 83The feedback of the agencies to the Kaysen proposal gave a mixed picture.Winn Finner, Acting Director, Agricultural Economics, made reservations about the project: "we are concerned about Congressional reaction to the Task Force proposal," 84 and he recommended, among others: "draft safeguards and ethics to be observed." 85In response to the Kaysen report, William Shaw, Assistant Secretary for Economic Affairs, sent a working paper that demanded a selection of specific data sets to be transferred to a center and said: "There should be a clear demonstration of how macrofiles of microinformation are in fact needed and used by the customers of the Federal statistical establishment." 86It insisted on "an explicit plan for the statutory and administrative implementation of provisions for assuring confidentiality and rights of privacy for records of the Center" before proceeding with the project. 87William Smith, Deputy Commissioner IRS, warned not to proceed with the project too fast: "any quick strong-arm approach could prove disastrous." 88Accordingly, the IRS objected placing a National Data Center under the aegis of the Executive Office, and insisted that tax returns were used for their primary purpose, not as a source for statistical analysis. 89There was little agreement about the organization and the function of a possible center.
It was a tricky issue where to locate a National Data Center and who to put in charge because DOC as well as BOB had competences in the field of federal statistics.Harold Seidman, Assistant Director, BOB, thought that "the suitable organizational structure and location becomes a key issue," and suggested to preferably install a possible National Data Center within the DOC or to create an independent agency. 90Concerning the administrative structure, a preliminary report gave three options: first, referring to the Kaysen report, to set up the position of a "Director of the Federal Statistical System" overseeing a National Data Center as well as the Census, second, the first option excluding the transfer of the Census, and third, to implement a "Data Center" within the Census under the DOC, while the BOB should control policy and programs. 91Especially the third option illustrates the organizational dilemma.It was a compromise to locate a "Data Center" at one agency but leaving important oversight functions at another.The compromise to give responsibility about federal statistics to both institutions, DOC and BOB, made its way.According to a proposal of December 1966, the DOC should host a "Federal Statistical Data Center" in close cooperation with the Census, while BOB should preside an interagency consulting group.The proposal should be presented to selected members of Congress and, if they were in favor, become part of the 1968 budget, otherwise the BOB should continue working on a proposal. 92Legislation should be drafted "to insure confidentiality and to safeguard against invasion of privacy through misuse of data by the Center." 93BOB Director Schultze told Califano that the heads of departments agreed to develop the proposal further, according to which a "National Statistical Data Center" should be installed close to the Census, while the Executive Office could still control policies and programs through the BOB. 94This institutional arrangement seemed less functional but more consensual than giving the priority to one single institution.John Connor, Secretary of Commerce, approved the BOB proposal of a "Federal Statistical Data Center" led by the DOC in order to reach "greater efficiency and effectiveness in administering Federal statistical programs," and he said that a new law was not required for legal reasons but sensible for political reasons. 95t the end of the year 1966, a government task force had developed its own proposals mainly focusing on centralization instead of a reform of the decentral system.
During the year 1967, agency officials continued to plan for centralization, but the Executive Office started working on a more gradual reform of federal statistics.It was still not clear whether a centralized system was in compliance with existing rules or whether it was necessary to draft a specific law.According to Arthur Focke, Office of General Counsel, there was no legal basis to set up a data center, because laws such as the Federal Reports Act or the Budget and Accounting Procedures Act rejected "the idea of centralization" and did not permit the use of data for other purposes than originally intended. 96Within the federal government the question of how to ensure confidentiality remained.As soon as a decision about a data center was made, the administration would introduce a law to Congress to "authorize its organization, to govern its operations, to insure against invasion of privacy, and with request for funds to finance it," according to a statement of the OSS, while a "Federal Statistical Center" should serve only statistical purposes and not contain any sensitive records. 97Despite these reassurances, privacy remained a major obstacle.Officials wanted to facilitate the statistical analysis of microdata for operations such as matching of different data sets or linking data over a certain time period.For instance, Carey Modlin, Council of Economic Advisers, sent the BOB a list of data sets to be included in a "Federal Data Center" and distinguished between "micro" or "macro" data depending on whether individual records were needed to be accessed for compiling "special tabulations." 98Likewise, Vito Natrella, Director of the Statistics Division, IRS, transferred a list of statistical data sets that could be used for "interrelation and collation with others." 99These statistical operations analyzed how certain characteristics of units were connected, a procedure that could not be performed with aggregated but with integrated data.
Yet the administration followed a different approach to improve statistical services by focusing on existing facilities.In 1967, when the press published mainly negative news stories on the data bank proposals, Califano demanded pragmatic solutions regarding the statistical system.For instance, Califano asked for a report on how to improve the work of the BOC. 100 One day later, Califano initiated a "Task Force on Quality of Economic Data for 1968," presided by Arthur Okun, Counsel of Economic Advisers. 101Furthermore, Califano contacted Schulte of the BOB, requesting a report on how the "right of privacy" related to "national data bank" proposals, 102 as well as a study how to improve "the collection, quality, and management of manpower data." 103Apparently, Califano was looking for alternatives to a central statistical facility, addressing single statistical services and areas.Yet Schultze rejected the idea of combining BLS and BOC without reorganizing the respective departments altogether: "We view the creation of a data center as a high priority part of any system's improvement effort." 104In November, the Task Force on the Quality of Economic Data presented its analysis, suggesting a "crash program" of five million dollars for the budget of the fiscal year 1969. 105he report itself contained various recommendations but did not mention any data center proposals.According to the authors, it was difficult to get funding for statistical programs in general, and Congress had rejected, among others, a Census sample for 1968. 106On request, Schultze sent a progress report of 6 November 1967, a draft bill as well as a proposal for another panel to deal with a "Statistical Data Center." 107Due to the privacy debate, the BOB decided to let an expert panel review the "Data Center" plans before proposing it officially and requesting Congressional approval. 108Despite the efforts of the White House to work on separate reforms of the decentral structure, the BOB continued to focus on centralization pursuing the plan of a single Federal Statistical Data Center.
Congressional hearings and the privacy debate continued in 1967.For instance, Dunn suggested to administrate a statistical system in such a way that prevented its utilization as a source for intelligence, but he acknowledged that there was no perfect solution either (Dunn, 1967, p. 23ff.).Critics as Lawrence Speiser of the American Civil Liberties Union explained in Senate hearings on Computer Privacy on 15 March 1967 that the proposals for a National Data Center had inherent threats for "civil liberties and privacy" concerning possible "personal dossiers" (U.S. Senate, 1967a,  p. 137).The public debate had shifted from administrative provisions to ensure confidentiality and comply with rules of disclosure to a normative question regarding individual rights and the role of the government.In addition, a congressional Subcommittee on Economic Statistics dealt with the question where to locate a data center and who was responsible within the administration and the Executive Office (U.S. Congress, 1967a, pp.1-4).In the meantime, the administration kept working on technical solutions.
Proponents of a data center believed in the benefits for political decisions and programs.The proposals inspired authors to speculate about a centralized society in the near future.For instance, Herman Kahn and Anthony Wiener of the Hudson Institute believed that society was lead from a single center like a pilot conducts a plane from a cockpit (Kahn, 1967, p. 351ff.).Author Robert MacBride projected for the future that industry and the finance sector would pool socioeconomic data in a single center and the economy would be controlled in real-time (MacBride, 1967, p. 123,  131).Despite these images, the reality looked quite different.Certainly, departments and agencies needed valid statistical data, but they ran their own offices and services for this purpose.Agencies did not demand a data center, but the initiative came from the OSS and in part the DOC.Here government officials were following the ideal of a Great Society and a War on Poverty, seeking to elaborate the phenomenon of poverty with information from a National Data Center (Downie,  1967, p. A2).From a political point of view, the project was already under severe pressure, since Congress investigated the matter pointing out threats to individual privacy.With regard to the privacy debate, Charles Zwick, Assistant Director of the BOB, suggested to postpone the project: "Given the emotion associated with the privacy issue, I doubt that we should make an attempt to win congressional approval during an election year." 109The protection of privacy moved to the center of the debate.For instance, authors of the report of the Joint Economic Committee asked: "Would further integration and particularly a national data center present a threat to personal privacy?" (U.S. Congress, 1967b, p. 7ff.)Centralization of statistics was either regarded as a threat to privacy or as a promising way to improve policy making.
The year 1968 marked the end of the project.Responding to the criticism concerning an alleged dossier center, Bowman said in February that he had appointed Krueger to work on a proposal for a "Federal Statistical Data Center", a center that would not contain any identifying information such as names or the Social Security Number in order to prevent a dossier function, waving the option to link or match data sets as a time series or a merger of different sources. 110Bowman argued that a "National Statistical Data Center" could strengthen rather than weaken privacy (Bowman, 1968, p. 812).There were still voices in favor of a National Data Center.In a comment on the Kaysen report the Harvard Law Review bought in the argument of "effective program planning" and the need for the government to address an "increasingly complex society" by setting up a National Data Center (Note, 1968, p. 400, 417).The paradigm of the benefits resulting from such a data center was rarely questioned.Jack Sawyer and Howard Schechter of Northwestern University argued: "a national data center necessarily has the clear potential to be used not only for administrative planning and basic research, but also as a means for evaluating the citizenry" (Sawyer & Schechter, 1968, p. 814).But operations connecting data allowed not only for analysis of relationships between individual characteristics but also for a closer look on a certain individual.
It seemed likely that a special law was necessary for a reform of the decentral structure of federal statistics.Finally, the BOB drafted a bill for the next Congress with the goal to establish a "Federal Statistical Service Center," 111 and the bureau wanted to cooperate with different agencies, consult an external panel and plan the budget. 112This "Federal Statistical Service Center" was supposed to be located within the DOC in order to archive data of different agencies and to provide machine-readable data for research, as the BOB proposed. 113The authors of a BOB report referred to two statistical operations relying on "Matching or Linkage" of data, either a "longitudinal matching" creating a time series of consecutive records, or a "cross file matching" creating a survey out of different sources that were connected for individual units. 114But transforming the decentral statistics into a central structure was too complicated to be realized soon.Time was running out.By October 1968, the BOB decided not to officially propose a data center but aimed for a "Budget Bureau Program for Improvement in Storage of and Access to Statistical Data," 115 as Bowman told the Director.The program excluded the centralization of files and the release of data to outside users, but still contained "pilot projects to integrate and collate data in selected existing files." 116Bowman wrote: "Opposition to the data center idea has focused almost exclusively on the dangers to privacy attendant to the centralization of data.This proposal would leave data decentralized for the time being." 117The earlier draft bill of the BOB never reached Congress (Kraus, 2013, p. 22).In the report of the House Committee, the authors rejected the proposals of a National Data Center under the current conditions, insisting that "no work to be done to establish the national data bank until privacy protection is explored fully and guaranteed to the greatest extend" (U.S. House of Representatives, 1968, p. 6).The election of a new administration put the final nail in the coffin.
In retrospect, the National Data Center proposals are mostly perceived as a threat to privacy, for instance, as sociologist Jerry Rosenberg said in 1969 in Death of Privacy: "Under the surveillance of a national data bank or center most of our actions could be documented, put into a permanent dossier and stored on tape with other vital data about us" (Rosenberg, 1969, p. 4f.).While scientist had all the way struggled with rules on confidentiality and disclosure, the privacy debate came as a surprise.Subsequent efforts for centralization also failed.According to a story of the Washington Post in June 1974, the General Service Administration planned the system Fednet providing "modern nationwide data facilities," but critics regarded the project as "murky and mysterious" (Editor, 1974, p. A14).President Gerald Ford rejected proposals for central information systems, and further reform projects like the Bonnen Commission under President Jimmy Carter were not successful either (Kraus, 2013, pp.30-33).The statistical system remained decentral with some guidance from the Executive Office.Likewise, in 1980, within the Office of Management and Budget, the former BOB, the Office of Information and Regulatory Affairs was funded, an office supervising the collection of information, policy concerning privacy of the citizens as well as statistical programs according to the White House (Office of Management and Budget, 2023a).The Chief Statistician oversees thirteen statistical facilities, three agencies mainly concerned with statistics as well as more than a hundred statistical programs (Office of Management and Budget, 2023b).The latest attempt to further integrate the statistical system was made in 2017 by the Commission on Evidence-Based Policymaking that considered the existing system as insufficient in order to provide data on policies.Therefore, the commission suggested to set up a "National Secure Data Service" for statistical purposes, referring to the plans of the 1960s: "Technology and privacy-protective approaches have advanced greatly over the past 50 years since the Data Bank proposal was rejected" (Commission on Evidence-Based Policymaking, 2017, p. 66ff.).

Conclusion
This article analyzed how scientists developed the idea to integrate government information in a Federal Data Center and how officials developed the proposals of a National Data Center further but ultimately failed.To make a long story short, an exploratory study on statistics by a low-profile committee was taken far too serious within the administration, and it backfired.Yet the story leads to further conclusions on the history of technology.Following Bruno Latour's approach to open "black boxes" of science as long as controversies are not yet settled (Latour, 2003, p. 14f.),, the article focused on the initial committee on economic data.The Committee on the Preservation and Use of Economic Data was eager to acquire so-called microdata regarding the federal government as a major source.But agencies would not release such data on the level of individual persons or companies for reasons of confidentiality and disclosure rules.Because the requests for data could not be realized on a small scale, the committee concluded to set up a data center on a large scale.Since the creation of an external facility seemed impossible, at this point, officials of the BOB encouraged the committee members to propose a government agency.The necessity and benefit of a data center for the federal government was taken for granted and rarely questioned.Within the status quo, numerous bureaus already provided statistics in a decentral structure.The OSS within the BOB were interested to integrate the information system and concentrate competences, getting into rivalries with the BOC and the DOC.Location and supervision of a possible center were contested.
The political debate and congressional hearings focused on concerns about privacy.These concerns were anything but new.From the very beginning, the committee mandated by the SSRC were confronted with rules of confidentiality.In the public debate privacy became a normative claim.As Westin defined at the time: "Privacy is the claim of individuals, groups, or institutions to determine for themselves when, how, and to what extent information about them is communicated to others" (Westin, 1967, p. 7).Besides serving for statistical analysis, the possible National Data Center was under suspicion to provide dossiers on individuals.Research suggests that the project failed due to privacy concerns as the first hypothesis stated.Simon Garfinkel argues in The database nation, the death of privacy in the 21st century that the proposals were about to transform into a "massive databank containing cradle-to-grave electronic records for every U.S. citizen" (Garfinkel, 2000, p. 13).Ironically, the idea of a Federal Data Center emerged as a way to solve problems of disclosure and confidentiality.Centralization was hard to realize within the decentral structure of federal statistics, as the second hypothesis stated.Several bureaus and agencies, most notably the DOC and the BOB, competed for competencies, and it remained unclear who should be responsible for a National Data Center or a National Statistical Data Center and where to locate it.The proposals for centralization were introduced by a research committee and did not evolve within the government as a meaningful way of reform.In this regard, interests between science and politics diverged, as the third hypothesis said, even though there were close ties between the two fields.
The question remains how to balance interests of the scientific community, the federal government and individual rights.One solution could be to establish procedures for creating samples of microdata, a technique that was already available at the time, instead of pooling data from different agencies.On the one hand, researchers and analysts were interested in shaping economic models and interpreting the social structure not in reading the characteristics of any individual person.On the other hand, new statistical methods evolved such as correlation between variables or the matching of data sets in a time line or across different sources, methods that relied on non-aggregated data, data that contained individual profiles.Parallel to this technological development, the discourse shifted to emphasizing the responsibility of the individual.For instance, while poverty had perceived as "a matter of social rather than individual morality" since the Progressive era, "individual, rather than social, morality" became the focus of neoliberalism (O 'Connor, 2001, p. 8, 10).Later, in the age of the internet, concepts as big data became prominent pooling information of different sources, and in addition, micro targeting became a mean to address individual persons based on their data.From a historical point of view, the debate on proposals for a National Data Center was at a threshold to data science and its ethics.