Two
Background and Research Design

The universe of private standards is massive and mysterious. There are literally tens of thousands of these uncelebrated standards. In contrast, public standards seem better understood and more commonplace. Agencies such as the Consumer Product Safety Commission, whose standards often generate significant media attention, have been the subject of countless studies. Most private standards-setters operate in relative obscurity. This chapter presents some general background about both universes, public and private. It also provides a précis of several important, but not well known, private standards-setters. The chapter concludes with an explanation of the research design of this study, consisting of four comparative case studies of setting safety standards.

Private Standards and Public Interests

According to the most extensive directory of standards-setting organizations, compiled in 1983 by the National Bureau of Standards, based on information submitted by private organizations, approximately 420 nongovernment organizations maintain thirty-two thousand standards in the United States.^[1] These standards facilitate commerce in various ways, but most are not particularly important to public policy. Some set forth definitions such as size of screw threadings. Others facilitate the interchangeability of items such as flashlight batteries and automobile parts. An ANSI standard specifies the minimum requirements for the

permanence of paper for printed library material (see the copyright page of this book). A small portion, consisting of at least one thousand standards, are infused with more significant implications for the public interest.

The largest component of this group involve matters of public health and safety. Private standards affect the public interest in other areas, of course, such as finance and communications. For example, the Financial Accounting Standards Board, an entirely private organization, develops the influential Statements of Financial Accounting Standards (SFASs) that form the bases of "generally accepted accounting principles."^[2] Similarly, the National Advertising Review Board and, to a lesser extent, the National Association of Broadcasters privately regulate advertising. But this study focuses on private safety standards, a field that roughly parallels the distinct domain of public regulation often placed under the rubric of "environment, health, and safety" regulation.

The subset of private standards that directly concern public health and safety is fairly well delineated. The best measure comes from the American National Standards Institute, an organization that certifies standards written by other groups. For reasons discussed in later chapters, most, but not all, private standards-setters seek ANSI approval. ANSI classifies approximately 900 of its 8,500 certified standards under the rubric of "Safety and Health."^[3] These standards cover a fantastic array from the obviously important (Criteria for Accident Monitoring Functions in Light-Water-Cooled Nuclear Reactors) to the seemingly trivial (Safety Standard for Christmas Tree Lights). Some of ANSI's safety and health standards are procedural (Storage and Handling of Mixed Fluid Fertilizers); others are substantive (Safety Requirements for Baling Equipment). A few are massive in scope. The Boiler and Pressure Vessel Code, for example, fills twenty-four volumes. Others take up just a page (Safety Standard for Agricultural Equipment), although obviously there is no way to tell from the name alone.

The ANSI list omits some significant private organizations involved in environment, health, and safety regulation. Building code organizations, for example, play a critical role in public safety. Building codes, which vary by region and by type of construction, make reference to countless ANSI standards. But the codes themselves are not certified by ANSI. Nor does ANSI list many safety standards developed in connection with the insurance industry. The Factory Mutual Research Corporation, for example, develops loss control standards for industrial and commercial policyholders insured by the Factory Mutual System. This

extensive private regulatory system, founded in 1835, consists of hundreds of engineers and technicians, along with a cadre of inspectors who implement FM's "Approval Standards." In short, the ANSI estimate of nine hundred private "health and safety" standards includes many, but not all, of the relevant standards.

Private standards-setting occurs under several institutional arrangements. Four basic forms of organization account for nearly all private standards: (1) trade associations, (2) professional societies, (3) general membership organizations, and (4) third-party certifiers. These organizations have different forms of governance, and they rely on a variety of administrative procedures. But a few core concepts permeate these organizations: one is "consensus" decisionmaking; the other is due process.

Organizational Forms of Private Standards-Setting

Trade associations are probably the best known and the least trusted form of private standards-setting. Since trade associations are created to advance the interests of their (usually homogeneous) memberships, it is widely assumed that their standards will be anticompetitive or otherwise against the public interest. But only some trade association standards have significant implications for the public interest. Most facilitate commerce in a rather mundane fashion. The Diamond Walnut Growers, for example, develop standards for the size, color, and grade defects for in-shell walnuts. Other trade associations develop a full array of such relatively innocuous standards. The American Petroleum Institute, for example, maintains 350 standards concerning the transportation, refining, production, measurement, and marketing of petroleum products. Trade association standards are usually financed directly by the membership, reinforcing the concern that narrow private interests will capture the process. But trade associations account for only a small portion of ANSI-certified health and safety standards.

Professional societies and general membership organizations bring together a broader spectrum of participants than trade associations. Rather than being tied to one economic interest, the memberships of these organizations are diverse. There are numerous professional societies organized around the specialties of engineering. For example, the Society of Automotive Engineers (forty-four thousand members) maintains over a thousand standards for ground vehicles and several thousand more for aerospace applications. Similar organizations include the

American Society of Agricultural Engineers (167 standards), the American Society of Heating, Refrigeration, and Air-Conditioning Engineers (60 standards), and American Society of Lubrication Engineers (22 standards).^[4] These organizations generally develop technical standards (for example, definitions, specifications, and tolerances), few of which are very controversial or significant to the public interest.

Membership organizations are broader in constituency, but not necessarily in purpose. They are more open than trade associations, often including participants from various professions and competing aspects of industry. The American Society for Testing and Materials, often called "the world's largest source of voluntary consensus standards," has almost thirty thousand members and seven thousand standards. Most of its standards resemble those of an engineering society. In other words, very few have significant implications for public policy. The National Fire Protection Association (NFPA), by contrast, has a similar number of members and about one-thirtieth the number of standards (approximately 250), but all NFPA standards concern public safety. One important feature that membership organizations have in common is loose reliance on the market demand for their standards: 77 percent of ASTM's budget comes from publication sales; NFPA derives two-thirds of its income from the sale of standards.

The final form of private standards-setting is so different from the others that it is often given a separate name: certification.^[5] "Third-party certifiers" test products against standards. They collect a fee for certifying compliance, which is usually signified by affixing a label or seal to the product. While some testing laboratories rely on standards developed by other organizations, the most significant product certifiers are also standards-setters. The most prominent of these organizations is Underwriters Laboratories, a nonprofit organization with the motto Testing for Public Safety. In-house engineers develop UL's standards. Less well known organizations engaged in third-party certification include the National Sanitation Foundation (which certifies restaurant equipment) and the International Association of Plumbing and Mechanical Officials (a model building code organization that certifies compliance of various products with the code). Some trade associations, such as the Association of Home Appliance Manufacturers and the Snowmobile Safety and Certification Committee, provide specialized certification services, which can generate significant income for them.^[6]

A few administrative procedures characterize almost all of these organizations. One is "consensus" decisionmaking. The other, surpris-

ingly, is due process. These tenets form the basis for ANSI certification of private standards and are incorporated into the by-laws of many private standards-setting organizations. Although the extent of due process protections varies by organization, and the real-world implications of "consensus" decisionmaking are not clear, these concepts obviously play an important role in how these organizations define themselves.

Profiles of Prominent Private Standards-Setters

Lacking a more detailed description of the universe of private safety standards, perhaps the best way to understand this territory is through its major landmarks. Some of the most prominent private standards-setters in the field of public safety are described briefly below. These organizations vary significantly in age and size (see table 2). Four of them—ANSI, AGA Labs, NFPA, and UL—are examined in detail in later chapters.

ANSI . The American National Standards Institute is unique among these groups, acting as an overall coordinator and certifier of the so-called voluntary national standards system. It is the trade association of the standards-writing industry. ANSI does not write standards. Other organizations, including ANSI committees—groups "accredited" by ANSI—submit their standards for approval as American National Standards. There are approximately 8,500 ANSI-approved standards (1,000 of which were developed by ANSI committees). The requirements for ANSI certification, discussed later in this chapter, are essentially procedural, not substantive. The Board of Standards Review hears complaints from anyone who objects to certification of a "national consensus standard." Appeals are infrequent, except in the area of "health and safety." Approximately 900 ANSI-certified standards are in this category. ANSI's membership consists of industry representatives and standards-setting organizations (including professional and technical societies, trade associations, and government agencies).

ACGIH . The American Conference of Government Industrial Hygienists publishes an annual compilation of Threshold Limit Values—recommendations on airborne contaminants and physical agents in workplaces—for approximately six hundred chemical substances. These values influence industrial practice in the United States and in a host of countries abroad. Despite the implication of its name, ACGIH is a private organization with no formal links to the public sector. The organization was founded in 1938 by federal, state, and local health officials. Its committees now include academics and industry representatives. A recent study of "Corporate Influence on Threshold Limit Values" concludes that there are "104 substances for which important or total reliance was placed on unpublished corporate communications."^[7]

AGA Labs . The American Gas Association Laboratories, a division of the larger trade association, is sometimes referred to as "the UL for gas appliances." Founded in 1918, AGA Labs provides third-party certification for all gas appliances. It currently maintains sixty-five standards. Unlike UL's, these standards are not developed by in-house engineers. They are developed by ANSI-sponsored committees consisting of various representatives, largely from industry.

ASME . The American Society of Mechanical Engineers is a nonprofit educational and technical organization with 110,000 members

and nearly six hundred active standards. ASME's reputation is based almost entirely on one standard: the Boiler and Pressure Vessel Code, a compilation of safety and performance requirements for power and heating boilers, nuclear reactors and power plants, and pressure vessels, which is widely incorporated into law throughout the United States and Canada. ASME's image was tarnished in 1983 when, after it refused to settle an antitrust case involving a blatantly anticompetitive interpretation of the Boiler and Pressure Vessel Code, the Supreme Court upheld a $9.5 million judgment against the organization. ASME argued that it should not be responsible for the acts of volunteers acting in bad faith.^[8]

ASTM . The American Society for Testing and Materials, founded in 1898, is a nonprofit organization "to develop standards on characteristics and performance of materials, products, systems, and services." A staff of two hundred oversees the maintenance of 7,218 standards, most of them standards for uniformity. A few committees act in a more overtly regulatory fashion. The F-15 Committee (consumer products) has developed a dozen standards for products such as high chairs, cigarette lighters, and bathtub grab bars. ASTM standards are written by "volunteer" committees and subject to the approval of ASTM's thirty thousand members. ANSI has long been an organizational rival of ASTM'S. The organizations trade allegations of "turf grabbing," and ASTM no longer submits its standards for certification as American National Standards.

Building Code Organizations: ICBO, BOCA, SBCCI . Building codes, enforced by local building inspectors, are one of the most visible forms of government safety regulation. They are largely written, however, by a complicated web of overlapping private standards-setters. Four model code organizations dominate the field. The International Conference of Building Officials (ICBO) promulgates fourteen comprehensive codes covering various aspects of construction. These codes make reference to nearly one thousand ASTM, UL, NFPA, and ANSI standards. The best known, the Uniform Building Code, is enforced in jurisdictions from Michigan and Indiana to Alaska and Hawaii. The Southern Building Code Congress International (SBCCI) and the Building Officials and Code Administrators International (BOCA) each promulgate similar sets of building, plumbing, mechanical, gas, fire, and housing codes. The SBCCI codes have been adopted in over 1,600 communities in the Southwest, South, and Southeast; the BOCA codes

cover fourteen states and 3,000 local governments in the East and Midwest. Finally, the Council of American Building Code Officials (CABO), created by the three major building code organizations, attempts to coordinate activities, particularly with regard to product certification. There are myriad other actors in the field of building codes, many providing specialized standards that are incorporated into building codes. The CABO One-and-Two Family Dwelling Code, for example, mandates, among other things, compliance with ANSI Z21.11.2 (the private safety standard for unvented gas space heaters, examined in chapter 6). The International Association of Plumbing and Mechanical Officials, author of fifty-seven product and installation standards, develops the Uniform Plumbing Code, adopted in the building codes of over 2,500 jurisdictions.

NFPA . The National Fire Protection Association is a membership organization similar to ASTM. "Volunteer" committees write the 260 NFPA codes and standards, and the membership votes en masse at semiannual conventions. NFPA has over thirty-two thousand members, including architects, engineers, firemen, manufacturers, and representatives of insurance interests, labor, and government. NFPA standards are published as the National Fire Codes in a multivolume set consisting of over twelve thousand pages. Various NFPA standards are referenced by OSHA, the Coast Guard, the Veterans Administration, and the Department of Housing and Urban Development. Its best-known codes, the National Electric Code and the Life Safety Code, have the force of law in most jurisdictions. Other standards cover the fire risks at nuclear power plants, airports, storage tanks, and grain elevators.

UL . The primary business of Underwriters Laboratories is product safety certification. UL evaluates products and monitors the quality control of their production. Manufacturers pay for the service, and if their products comply, they are entitled to display the UL label. There are other testing laboratories—most prominently, the American Gas Association Laboratories in the field of gas appliances—but many simply certify compliance to UL standards. The most important feature of UL is that it writes the standards it uses in testing. There are over five hundred published UL standards, covering such diverse products as microwave ovens, life preservers, kerosene heaters, fire extinguishers, and automated teller systems. UL maintains membership in five hundred committees of other private standards-setting organizations.

The Universe of Public Safety Standards

Public standards seem both more visible and more controversial than private ones. In the areas of environment, health, and safety, most of the federal agencies that develop standards are practically household names: EPA, OSHA, FAA, FDA. Other federal agencies, involved in similar missions, have less name recognition, but they are nevertheless well known for what they do. The National Highway Traffic Safety Administration (NHTSA), the Nuclear Regulatory Commission (NRC), and the Consumer Product Safety Commission, for example, all write safety standards in areas where public regulation is widely recognized and generally supported.

These agencies vary considerably in size. It is difficult to compare the number of standards they develop because public standards usually come in packages, rather than individually. Unlike the private sector, where separate committees develop discrete standards on a project-by-project basis, public agencies take a range of actions from the adoption of single standards to the implementation of complex statutes. The Code of Federal Regulations is not divided by individual standards. For example, the safety standards of the NRC comprise 850 pages in one title of the code, which by one count represents approximately 350 standards. In contrast, the subchapter of the code concerning biological products regulated by the FDA accounts for approximately 325 standards.^[9] Lacking a comparable measure of actual standards-setting activity, the relative size of these agencies can roughly be gauged by their budgets and staff (see table 3).

These agencies also vary considerably in age, reflecting the three waves of public regulation this century. The Food and Drug Administration (FDA) and the Federal Trade Commission (FTC) are the oldest, reflecting the Progressive Era, when federal regulation began. The FDA was created in 1906, the FTC in 1914. The Federal Aviation Administration (FAA) evolved from a New Deal agency, the Civil Aeronautics Board (1938), created along with a host of agencies primarily engaged in economic regulation (for example, the Securities and Exchange Commission and the National Labor Relations Board). The most controversial federal agencies are also the youngest. These agencies spearhead the "new social regulation" movement, a grandiose agenda of environment, health, and safety objectives adopted about twenty years ago. The most prominent are the EPA, OSHA, NHTSA, NRC, and CPSC. Although other strategies would be possible, these agencies generally favor

"command and control" regulation, which relies on standards, inspectors, and penalties to achieve social ends. One important feature common to all of them is the Administrative Procedures Act (APA), which describes two kinds of rulemaking, formal and informal. Most agencies favor the latter but implement it in a "hybrid" fashion that incorporates some due process protections beyond those demanded by the APA.^[10]

Agencies concerned with environment, health, and safety regulation account for only a small portion of all federal standards. By one count, the federal government maintains more standards than the private sector. According to the National Bureau of Standards, there were forty-nine thousand standards in 1983.^[11] As in the private sector, however, most of these are not particularly weighty. The overwhelming number are procurement standards adopted by either the Department of Defense (38,000) or the General Services Administration (6,000). The number of federal standards with significant implications for the public interest is minuscule by comparison.

In most areas involving the environment, health, or safety, there are far more private standards than public ones. Take consumer product safety, for example. Hundreds of private standards are developed by UL alone. In contrast, the relevant federal agency, the CPSC, had twenty-one active standards in 1984, many of which were carry-overs from old statutes.^[12] A few others seem absolutely frivolous. Not coincidentally, these standards (matchbooks, swimming pool slides) were overturned in court. This leaves only a handful of safety standards actually devel-

oped by the CPSC—most prominently, standards for lawn mowers, gas-fired space heaters, electrical toys, and citizens band base station antennas.

In most areas that government seeks to regulate there are already so many private standards that interaction between the two sectors is inevitable. Model codes and use codes, including most standards developed by the National Fire Protection Association, contain repeated references to "the authority having jurisdiction." In other words, they are written in anticipation of adoption by government (usually at the local level). The federal agencies that regulate safety issues are also intertwined with the private sector. Many federal agencies rely directly on private standards to accomplish public purposes. The FDA, for example, has adopted over three hundred standards from private organizations such as ASTM, the American Public Health Association, and the Association of Official Analytical Chemists. The National Highway Traffic Safety Administration, by contrast, uses private standards (developed by the Society of Automotive Engineers) for test methods and other largely technical matters, while leaving basic safety questions to the public realm. Most agencies participate in private standards-setting activities. The CPSC, which originally had a procedure for private organizations to "offer" standards for agency adoption, has participated in the development of hundreds of private standards. The Nuclear Regulatory Commission participates on more than 150 private standards committees. Although interaction and coexistence prevail, there remain distinct areas of public regulation.

Research Design

Making a direct comparison of public and private standards-setting practically requires that the two sectors be active in the same area. Otherwise, whether outcomes or procedures are evaluated, case studies in either sector will leave the same lingering question: compared to what? The criterion of economic efficiency, often a useful benchmark, probably has limited potential. "Determining the proper level of product safety is next to impossible," note Eads and Reuter in the introduction to their case studies of corporate safety efforts.^[13] Without independently conducting cost-benefit analyses, it is doubtful that existing information will facilitate anything more than a crude analysis of outcomes.

An obvious answer to the benchmark problem is to compare the

private sector directly with the public. After all, the most important public policy questions about private standards are comparative. What really matters in assessing possible government action is how that action compares to the actual alternatives. One can only speculate whether public standards would actually be stricter in the thousands of areas presently regulated exclusively by private standards. But in the areas where public and private efforts overlap, outcomes can be evaluated comparatively.

If appropriate "pairs" can be found, the comparative approach has two obvious advantages. First, it solves the benchmark problem. Without a comparative framework, it is difficult to imagine how to evaluate either public or private standards in a fashion that will shed light on the other sector. The exercise would be entirely hypothetical: guessing what the other sector might have done. Second, the comparative approach might also facilitate broader generalizations by "controlling" for the idiosyncrasies of issues. It is frequently alleged that general theories of regulation are fruitless because circumstances such as industry profitability and structure, the quality of available information, and the mixture of political interests vary significantly by issue. A related argument, relevant to the study of public and private regulation, is that only certain types of issues are regulated by government. Government, the argument goes, regulates the issues that industry will not, making the public cases unique. Both of these arguments fade if the public and private sectors are active in the same area.

There are drawbacks, however, to studying only those cases in which public and private efforts can be paired. First, the universe of possible cases is very small. Public and private standards rarely regulate the same things. Moreover, when there is overlap, there is the attendant danger that the effects of their interaction will obscure the influences that might otherwise prevail. In other words, the observed outcomes might reflect "strategic behavior" more than they reflect the indigenous characteristics of either sector. If either sector acts strategically—that is, based primarily on the anticipated response from the other sector—then the pairs themselves might be idiosyncratic. In other words, special patterns of regulatory behavior might characterize such pairs. One likely pattern is "splitting the territory"—the strategy by which both sectors adapt their behavior to avoid direct overlap. This problem is not unique to actual pairs, however. Strategic behavior is considered a primary characteristic of most private standards-setting. The nature and significance of the pair problem is examined in chapter 8.

The possibilities for "paired" case studies are limited mainly by the comparatively narrow scope of government regulation. There are private safety standards in almost every field where there are government standards. With so little known about these standards and the organizations that write them, however, it is difficult to know how potential case studies from the private sector fit into the larger universe of standards-setting.

Several considerations guided the choice of cases for this study. First, in order to permit an actual comparison of public and private decisionmaking, the cases had to overlap in content but have some degree of independence in development. In some cases, similar content denotes a lack of independent effort. Sometimes, a government standard is nothing more than a private standard with the force of law. The HUD mobile home standard was originally written by NFPA. The FDA relies largely on UL to evaluate microwave ovens. In other cases, the overlap is in title only. With many of the automotive safety standards, for example, the private sector (the Society of Automotive Engineers) writes test methods in areas where government standards specify performance levels.

These considerations narrowed the apparent field of possible pairs considerably, leading to the second consideration: how "representative" the pairs would be of the public and private sector. Recognizing that it would be impossible to select truly representative pairs without knowing more about the universe, particularly on the private side, selection was based on reputation. A primary goal was to examine private organizations with the best possible reputations, a variation on deviantcase analysis. Examining "best case" examples should provide a basis for estimating the outer bounds of the private sector's potential. Conversely, it should avoid the criticism that the sample is tilted toward the "bad apples."

By this reasoning, the most suspect class of private standards-setting organizations—trade associations—was eliminated from further consideration. Trade associations seem least likely to advance public purposes because of the narrow scope of their interests. Although trade associations vary more than is often appreciated, few have sufficient independence from immediate membership demands to seek a more enlightened, long-term course.^[14] The remaining organizations—professional societies, general membership organizations, and third-party certifiers—are not necessarily better, but their form of organization suggests greater potential for something beyond parochialism.

Within this grouping, UL clearly had to be one of the private organizations studied. It is not only the oldest and best-established private standards-setter; according to knowledgeable observers, it is also widely thought to be more independent and public-spirited than other private organizations. In addition, UL is most active in consumer products, an area in which standards policy may be most significant. Not coincidentally, there is practically no overlap between UL's standards and the few that have been successfully promulgated by the Consumer Product Safety Commission. UL actively seeks to keep the CPSC off its turf. The CPSC has considered getting involved in many areas regulated by UL, but the only case of dual standards is that of woodstoves. Accordingly, that is the first pair in this study.

The overlap between UL and the CPSC is minimal in the woodstove case. Instead of developing a product standard for woodstoves, the CPSC supplemented UL's product standard with a labeling standard—an unusual form of CPSC regulation. Therefore, although the woodstove case offers an opportunity to examine UL, it appears to provide a peculiar view of the CPSC. Selection of the second pair was motivated by a desire to balance the woodstove case with one in which the CPSC adopted a full-fledged product standard—its more usual role in regulation. Given the CPSC's relative inaction in recent years, there were only two possibilities: lawn mowers and gas space heaters. The private sector's lawn mower standard was adopted by an ANSI committee under the sponsorship of the Outdoor Power Equipment Institute; the gas space heater standard was also adopted by an ANSI committee, but under the supervision of the American Gas Association Laboratories. The latter was chosen for two reasons. First, the gas space heater case is more recent, minimizing the chance that subsequent changes in either the public or private sector would render the findings outdated. Second, the lawn mower case has been studied many times; gas space heaters have not."^[15]

Desire for a "best case" example of public standards-setting motivated the selection of the remaining cases. The CPSC may provide close to the "worst case" example of government regulation. The agency's short history has been marked by failure and poor judgment."^[16] (An American Enterprise Institute study in 1983 suggested abolishing the CpSC.)^[17] In order to ensure that the pairs were more representative of the potential for government regulation, the remaining cases had to involve government agencies with better reputations.

The possibilities included automobile safety (where there is some

overlap between NHTSA and the Society of Automotive Engineers), nuclear power safety, aviation safety, and grain elevator safety. The first two were eliminated for practical reasons—they appeared to require a researcher to have substantial engineering (and possibly physics) background. Accordingly, the remaining cases were selected as the third and fourth pairs. In aviation safety, the FAA's 1984 proposal concerning fire extinguishers and smoke detectors overlaps directly with an NFPA standard. Although the FAA is criticized occasionally for an inadequate inspection and maintenance program, the agency enjoys strong congressional support and a surprising level of industry support for its regulatory activities.

The grain elevator case involves an OSHA standard, also paired with an NFPA one. Although OSHA's reputation, unlike the FAA's, is only a little better than the CPSC'S, this case study seemed appropriate for three reasons. First, it looked at "the new OSHA"—that is, OSHA under President Reagan. The grain elevator standard is one of only a few OSHA standards promulgated under the Reagan administration, and as such it might not be prone to the perceived excesses of previous OSHA standards. Second, like the CPSC, OSHA addresses the kinds of issues in which questions about the relationship between public and private standards are most likely to arise. Finally, the grain elevator case offered a research opportunity unavailable on the private side of the other selected cases: to observe the revision process in action. The NFPA grain elevator standard was due to be revised at a two-day subcommittee meeting in July 1985, offering an opportunity to observe the committee responding to comments and deliberating over specific language. This seemed likely to make possible a more complete and realistic view of private standards-setting than the other case studies (based as they are on documents and after-the-fact interviews).

Notes on Fieldwork

The fieldwork for the case studies was conducted from September to December 1984 in Washington, D.C., New York, and Boston, and from May to July 1985 in Chicago, Cleveland, New York, Washington, D.C., and Minneapolis. Seventy-two detailed interviews were conducted, approximately half of which were tape-recorded and transcribed (a list of these is given in the Appendix). Additional information on the research method is contained in notes that precede the bibliographies for each case.

A major problem encountered during the fieldwork, but not anticipated in designing the research, was access. As it turned out, initial contacts with Underwriters Laboratories were met with suspicion and an assertion that UL would not permit its engineers to be interviewed. The NFPA, which professes openness, was also hesitant about providing certain documents and allowing attendance at subcommittee meetings. These organizations are more secretive than they appear. They are sued with regularity and are not anxious to talk about it. Moreover, they have been on the defensive since the FTC began its investigation of standards and certification. Shortly before this fieldwork began, Mother Jones magazine ran an article containing criticisms of UL.^[18] A few years earlier the "60 Minutes" television program did a critical story on the American Gas Association Laboratory.

In both cases, personal contact eventually resulted in access: UL's representative in Washington, D.C., arranged for access to UL's engineers, and after a lengthy personal meeting, an NFPA vice president provided assistance in obtaining documents. Even so, a few NFPA committee members were unwilling to discuss their work in detail, and several UL engineers were quite guarded in conversation. During the course of the fieldwork, the FAA standard on fire extinguishers and smoke detectors was approved. The OSHA standard on grain elevators was adopted two years after conclusion of the fieldwork.

The other unexpected development during the course of the fieldwork involved revelations about the relationship between product standards and installation codes. An apparent advantage of the initial case selection appeared to be the diversity of private standards-writers; the four pairs involved three different organizations (UL, AGA, and NFPA). It turns out that NFPA plays an important role in the standards of both UL and AGA Labs. Accordingly, the case studies contain much more information about the NFPA than originally expected. This points up a potentially important distinction between codes and standards that was not understood during the research design. The two are so different that they are probably best studied separately. However, this discovery bears out the decision to do case studies in order to identify such factors, instead of doing a broader survey.