Chapter 1 were compared with the results of OMB's review of the corresponding submission. In answering the second question-How do OMB's policies and practices influence agencies' decisions regarding requests to collect information?-we used two methods. Our first method involved semistructured interviews with officials from a representative sample of 50 agencies. Specifically, using our longitudinal data base of the universe of actions, we selected a stratified random sample of 50 agencies based on agencies' current and prior experiences with the OMB's review process. We classified agencies according to their relative success at obtaining approval from OMB and changes over time. The categorization resulted in five types of agencies. The types were agencies with (1) low approval rates (less than 90 percent) in both the early (1982-84) and late (198587) periods, (2) a low approval rate in the early period but a high rate (over 90 percent) in the later period, (3) a high approval rate in the early period and low in the later period, (4) high approval rates in both periods, and (5) limited experience with the paperwork clearance process. Within each of the first four categories, agencies were classified as having regulatory, nonregulatory, or statistical missions. We completed interviews with 17 paperwork clearance officers responsible for 38 of the 50 agencies in our sample. We also interviewed a subsample of 17 agency managers. The purpose of these interviews was to understand the paperwork development policies and procedures in each of the agencies and the perceptions of these agency officials as to how the paperwork clearance process has affected data collection decisions. We collected documents, when they were available, describing policies and practices within agencies. Our second method involved the 17 case examples described above. Here we examined more closely the development and submission of these collections within the agencies. In order to answer the third general question-How have agency and OMB actions influenced the availability of information?—we used two methods. First, using the data in the reports management system for our sample of 50 agencies, we examined the submission practices of the agencies over time. Of particular interest were the changes in the types of submissions (research, evaluation or statistics, application for benefits, regulatory or compliance, and so on) made by statistical and nonstatistical agencies and by agencies with high and low approval rates Chapter 1 Strengths and Report Organization over time. Second, we coupled this information with information from the interviews with agency managers and clearance officers, as well as the interviews with OMB's desk officers and managers and additional interviews for the case examples, regarding the likely effects of the actions of agencies and OMB on the availability of information. Our study design contains several methodological strengths and at least one limitation. With regard to its strengths, unlike prior reports on OMB's practices, our longitudinal assessment provides a basis for examining changes over time. Further, we selected our samples to represent the diversity of experiences that agencies have had in their interactions with OMB. Thus, we are able to describe the prevalence of problems and strengths in the review process. Finally, we used at least two different methods to answer each of the committee's questions. This strategy has several benefits. First, each method provides data on different aspects of the same process. For example, the longitudinal data used to describe the actions of OMB or agencies provides evidence on general trends. Interviews with agency officials provide complementary information on how the trends may have come about, and the in-depth study of particular cases generates evidence on the processes that may be at work. While each source of evidence, by itself, contains limitations, when all are used in concert, they yield a more complete, objective, and balanced answer to the complex questions we have been asked. Our study design does contain a limitation. Namely, we have been asked to assess how OMB influences agencies' decisions to collect data and to ascertain the joint influence of agency and OMB actions on the likely availability of information. Many forces result in changes in organizational practices. Our design cannot account for the influences of them all. In discussing answers to our questions, we describe where appropriate influences as plausibly associated with the actions of OMB or agencies. Chapters 2 and 3 provide answers to the committee's first question. In chapter 2, we describe the formal and informal process by which information collection requests are handled within OMB, outcomes of the review process, and the criteria used in making decisions to approve or disapprove a request. In chapter 3, we present our findings on the timeliness and technical adequacy of OMB's reviews. Chapter 4 presents our Chapter 1 findings on the second question. That is, it examines the influence of OMB's paperwork review policy and practice on agencies' decisions to collect data. Chapter 5 presents our findings on the third question. Here we examine how OMB and agency actions have influenced the availability of information. In chapter 6, we offer conclusions and recommendations for achieving a reasonable balance between the need to collect data and the need to control the federal paperwork burden on the public. Chapter 2 How Requests to Collect Information Are Since the enactment of the Paperwork Reduction Act in 1980, OMB has annually received from federal agencies between 3,000 and 4,000 requests for approval to collect information from the general public, businesses, states, and localities. Over 200 federal departments and agencies have submitted requests (they are listed in appendix D). To handle this volume of submissions, OMB has implemented regulations for submitting information collection requests. It has also developed a formal process for soliciting input on each request, reviewing submissions, deriving decisions, and notifying the agencies of its actions. Although a seemingly orderly and formal review process has been developed, the main OMB reviewers, the desk officers, rely on a variety of informal practices that affect decisionmaking on individual requests. Further, although one of the main reasons for centralizing the review process within OMB was to ensure that information was not duplicative, the majority of desk officers we interviewed stated that they did not use or there were limitations in using the Federal Information Locator System to check for duplication.' Rather, personal experience or contacts with other desk officers served as the basis for making this determination. Part of the variation in review practices appears to stem from the fact that new desk officers received little on-the-job training, and most desk officers had minimal training in areas needed to judge the technical merits of an information collection request, such as methodology. Although practices varied across desk officers, the outcomes of the review process were fairly predictable. Most submissions (93-95 percent) were approved, while a small fraction of these (8-12 percent of those we sampled) were ultimately modified as a result of the review process. Those that were disapproved were most often questioned on the grounds that their practical utility had not been established or that the collection was not necessary for the agency's proper performance. This portrait suggests that the OMB review process is relatively straightforward and represents a minor obstacle for federal agencies. The overall picture, however, masks at least two important areas of concern: a small number of agencies have had persistent difficulties in securing approval to collect information and new submissions, in particular new The Paperwork Reduction Act of 1980 mandated the creation of Federal Information Locator System to provide data on existing federal data collections for the purpose of ensuring that proposed data collections do not duplicate information already available to the federal government and for provid ing public access to data. The system that was established in 1983 was in operation until September 1988. However, OMB reports that it and other agencies had problems with the earlier version of FILS of insufficiency in identifying likely duplications. OMB is currently implementing the requirement for a FILS through enhancements to OMB's reports management system. Volume of Requests Chapter 2 How Requests to Collect Information Are submissions directed at research, evaluation, or statistical data collection, were considerably more likely to be disapproved or withdrawn than were other types of submission. As seen in table 2.1, over 20,000 submissions were received from 211 2Not all collections labeled new in OMB's administrative data were in fact new. Some had been previously approved more than 3 years ago and were therefore treated as new collections. Others, particularly in 1982-84, were ongoing collections that had not previously required OMB's clearance. |