Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 35
3
Experience with Public Posting of Government Data
INTRODUCTION
The concept of publicly posting government-generated data to provide direct public
access to information is not new. In response to calls for increased transparency and increased
provision of information, several government agencies, including regulatory agencies
responsible for protecting human health and safety, regularly post detailed data on the Internet.
In some cases, the data are related to individual firms or facilities; in other cases, they are
specific to commodities or products or to events. Although one could argue that government data
are public by default and only under special circumstances should they be restricted, the
committee began its assessment from a more neutral ground and by considering potential
benefits of and concerns with releasing data, as this was the task given to the committee.
This chapter briefly summarizes several examples of public posting of detailed (disaggregated or
establishment-specific) data. It also reviews some of the literature on the use and effects of data
releases. Currently, there are no empirical data on the effects (both positive and adverse) of
releasing establishment-specific FSIS data on the Internet. Therefore, the committee reviewed
the existing evidence on the benefits and costs of public release of data by other government
agencies. The review is meant to be illustrative rather than exhaustive. We focus on the posting
of data that stem directly from regulatory activities (Category 1) but also briefly discuss the
public posting of information derived from some prior analysis of data
35
OCR for page 36
(Category 2) and the posting of voluntarily provided data (Category 3). 20 In addition to the
examples discussed in this chapter, there are numerous other examples of public release by
government agencies of safety-related data on products or firms; some of these are briefly
summarized in Box 3-1.
Box 3-1
Examples of Sharing of Safety-Related Data on the Internet
Airborne Contaminants. The Occupational Safety and Health Administration posts some of its
compliance-monitoring information on airborne contaminants released from personal, area, and bulk
samples in industrial sites. URL: http://www.osha.gov/opengov/healthsamples.html.
Hospital Measures of Outcome of Care. Medicare publishes hospital-specific rates of outcome of care,
which indicate what happened after patients with particular conditions were treated in the hospital. URL:
http://data.medicare.gov/dataset/Hospital-Outcome-Of-Care-Measures/f24z-mvb9.
Safety in the Transportation Industry. The Bureau of Transportation Statistics publishes multiple
datasets on transportation accidents and exposure to safety risks (for example, measured in aviation
incidents, accidents, or fatalities). URL: http://www.bts.gov/programs/safety/index.html.
Safety of Nuclear Plants. The Nuclear Regulatory Commission posts plant-specific safety-inspection
reports and licensees’ performance indicators. URL:
http://www.nrc.gov/NRR/OVERSIGHT/ASSESS/index.html.
Safety of Motor Vehicles and Equipment. The National Highway Traffic Safety Administration posts
data on safety for the consumer, such as ratings of cars and tires
(http://www.safercar.gov/Vehicle+Shoppers/5-Star+Safety+Ratings/2011-Newer+Vehicles
and http://www.safercar.gov/Vehicle+Shoppers/Tires/Tires+Rating)
and children’s car seats (http://www.nhtsa.gov/Safety/Ease-of-Use); a list of all vehicle, equipment, and
tire safety-recall campaigns from 1966 to the present (http://www-odi.nhtsa.dot.gov/recalls/); consumer
complaints related to the safety of motor vehicles and motor-vehicle equipment (http://www-
odi.nhtsa.dot.gov/complaints/); and investigations of specific vehicles, tires, and equipment (http://www-
odi.nhtsa.dot.gov/cars/problems/defect/defectsearch.cfm).
20
As described in Chapter 2, Category 1 data arise from the activities of agencies as part of their normal enforcement
and compliance efforts. Category 2 data arise from the outcomes of enforcement and compliance efforts that have
been interpreted by others for use by end users. Category 3 data are collected by agencies from voluntary programs
not in conjunction with normal enforcement and compliance efforts but nonetheless intended to provide information.
36
OCR for page 37
EXAMPLES OF COLLECTION AND RELEASE OF DATA BY REGULATORY
AGENCIES
US Department of Labor
As part of the broader Open Government initiatives of the Obama administration, various
agencies of the US Department of Labor (DOL) have expanded direct public access to their
inspection and enforcement data, which are posted on a comprehensive Web site. 21 The data
underlying the site arise primarily from the enforcement activities of the agencies. Each agency
offers different types of information and levels of detail to the public, which reflect differences in
agency mission, nature of the regulatory process, sophistication of data collection, and
administrative processes, such as case-review procedures. 22 The agencies provide a variety of
information, including details about the inspected entity (such as industry, firm and
establishment size, and single-plant vs. multiplant status), characteristics (such as time spent and
type of inspection activity) and outcomes of the investigation (such as standards violated,
severity of violations, and penalties assessed), and related administrative processes (appeals and
their results). Accordingly, the data on the site are primarily in Category 1.
The site is regularly expanded and improved. Prior updates have focused on making it
easier for users to search by common criteria, such as company name or industry grouping. That
potentially provides information about the compliance behavior of a specific employer or
industry for a range of workplace laws. DOL is planning a number of future updates to increase
usability, including display of data through maps and interactive “dashboards” and engaging
public users of the data in finding “innovative ways of using DOL’s enforcement data to promote
worker’s safety and protect worker’s rights”. 23
In addition to the information on the comprehensive DOL Web site, some of the
individual agencies in DOL post detailed facility-specific safety data. For purposes of
illustration, we focus here on the Mine Safety and Health Administration (MSHA). MSHA is
responsible for the enforcement of health and safety standards for underground and surface metal
and nonmetal mines. Compliance with detailed health and safety requirements is determined
through physical inspection of mining facilities, interviews with mine operators and with
workers and their representatives (in unionized mines), and review of administrative information.
Inspectors also sample dust and air.
The most extensive mine-level data available to the public are published on the MSHA
Web site. 24 Those data originate in the electronic information systems maintained by the agency.
The data are stored in 16 linked databases that provide information on inspections, citations,
penalties, and abatement requirements. The site also provides mine-level data on fatalities and
21
See http://ogesdw.dol.gov/ (accessed June 7, 2011).
22
With respect to the latter dimension, agencies vary according to when the results of completed inspections and
investigations are publicly posted. The Wage and Hour Division posts only cases that are considered “closed” (for
example, all appeals of the investigators’ findings have been resolved). In contrast, the Mine Safety and Health
Administration posts inspection data even when a mine operator or other party is appealing parts of a decision, such
as penalties.
23
See http://ogesdw.dol.gov/coming_soon (accessed June 7, 2011).
24
See http://www.msha.gov/drs/drshome.htm (accessed June 7, 2011).
37
OCR for page 38
injuries, air sampling results, such mine-level characteristics as geology and type of mining
technology, and detailed information on ownership and management of mining activities.
The publicly available data span from 1983 to the present and are updated weekly.
MSHA also provides information on closed and active inspections, including cases in which
mine operators have contested penalties or abatement orders. Data can be searched by any of the
characteristics of mine operation, ownership, inspection finding, and so on; the data can be
downloaded as extracts; and data from the various databases can be combined by using a
common mine-level identifier system.
US Environmental Protection Agency
In 2003, the US Environmental Protection Agency (EPA) launched its Enforcement and
Compliance History Online (ECHO), 25 a Web-based platform that provides easy access to EPA
and state data on environmental compliance and performance of over 800,000 individual
facilities in the United States. The Web interface draws on an underlying dataset, the Integrated
Data for Enforcement Analysis (IDEA), which integrates data from five enforcement and
compliance history datasets. Users can use ECHO to search for facility-specific data by location
(ZIP code) or other identifiers. It is designed primarily for situations in which a user is interested
in information on a relatively small number of facilities, but users who want to review larger
amounts of data can access the raw data from IDEA. The data are updated monthly.
The content of ECHO and its user capabilities have evolved. ECHO now allows Web-
based access to the following types of facility-specific data:
Inspection, violation, and enforcement data, including the number and dates of
individual inspections, compliance status by quarter, and penalties imposed
during the preceding 5 years.
Data on EPA enforcement cases.
Data on violations of the Safe Drinking Water Act, including the publication of a
list of all water suppliers deemed to be “serious violators”.
Toxics Release Inventory (TRI) data, which are mandatory, self-reported releases
of designated toxic chemicals by facility.
National Emissions Inventory data with information about estimates of air
pollutants from point, nonpoint, and mobile sources in the United States, for
example, data from state and local agencies, data on on-road sources from the
Federal Highway Administration, and fuel-use data from the Department of
Energy.
Detailed water-quality reports on facilities that have permits under the Clean
Water Act and information on noncompliance with effluent limits.
In addition to detailed facility-level data, ECHO includes more aggregated summary reports that
provide information about trends and state-level analyses.
25
See http://www.epa-echo.gov/echo/about_data.html (accessed June 8, 2011).
38
OCR for page 39
EPA reports that in its first year ECHO provided information in response to over a
million search requests. 26 It identifies members of the public, corporations, investors, and
researchers as possible user groups. In addition, EPA notes that provision of data to the public
creates an incentive for government agencies to improve the reporting of violations and for
facilities to take steps to correct violations.
Although EPA has worked continuously to enhance the usefulness of the data in ECHO,
some of the reported data are in “raw” form and can be difficult for users to interpret. For
example, the TRI data are reported in pounds released annually with no direct means of
converting the releases to a more useful measure, such as associated health risk. Efforts have
been made to convey health risk to end users, but the current information is not easy to find and
is not detailed and quantitative enough for end users to use to estimate the risk to which a person
might be exposed. Data disclosure, however, is likely to evolve and improve once shortcomings
are identified.
Food and Drug Administration
The Food and Drug Administration (FDA) has several databases that are available to the
public. They include data on inspections and enforcement (Category 1) and voluntarily reported
information on actual or potential adverse events (Category 3). FDA also collects some
microbiological sampling and testing data, but these are not generally available to the public.
Although FDA has posted summary data for many years, it announced in May 2011 that
it would disclose additional inspection information on FDA-regulated food products, including
the compliance status of specific firms as determined by FDA inspectors during inspections and
followup reviews for compliance. FDA made that information available to the public through a
searchable database on the its Web site, which includes the names and addresses of inspected
facilities, the dates of inspection, the types of FDA-regulated products involved, final inspection
classification, and a summary of the most common inspection observations, although not in great
detail. 27 The information is substantially equivalent to the information now provided by the US
Department of Agriculture (USDA) Food Safety and Inspection Service (FSIS) on administrative
actions (USDA FSIS, 2010; see Appendix C). FDA also provides access through its Web site 28
to facility-specific information, such as letters that warn firms that violations have been
identified and must be corrected, and enforcement reports that contain information on actions,
such as recalls, taken in connection with regulatory activities. Aggregated information about
enforcement activities is also found on the FDA Web site. FDA justified the disclosure of the
information on the grounds that it would help to provide the public with a rationale for the
agency’s enforcement actions, help consumers and industry stakeholders to make informed
choices in the marketplace, encourage industry compliance, and generally improve transparency
of agency actions to be consistent with administration policies.
In addition to Category 1 enforcement and compliance data, FDA collects other safety-
related data from health-care professionals, public-health officials, consumers, and the food
26
See
http://yosemite.epa.gov/opa/admpress.nsf/b1ab9f485b098972852562e7004dc686/e6bf84f19616f3b985256de30055a
fcd?OpenDocument (accessed June 8, 2011).
27
See www.accessdata.fda.gov/scripts/inspsearch/ (accessed July 25, 2011).
28
See http://www.fda.gov/AboutFDA/Transparency/TransparencyInitiative/ucm254426.htm (accessed July 25,
2011).
39
OCR for page 40
industry. 29 For example, information on potential or actual adverse events is collected by FDA
through the Reportable Food Registry (RFR) 30 for foods and through the Adverse Event
Reporting System (AERS) 31 for drugs, biologics, and dietary supplements. Industry must report
adverse events to the RFR. Raw data from the RFR are not released to the public, but FDA has
posted two reports since RFR was implemented: a first-7-months report (September 2009–March
2010) and the annual report (September 2009–September 2010) with summary information
aggregated in various forms, such as total entries by commodity or by commodity and hazard.
FDA also posts quarterly data files from AERS on its Web site and summary statistics for each
year. Although the information is not company-specific or facility-specific, the release of
information about adverse events can affect individual firms and entire industries whose
production is linked in some way to the events.
State and Local Public-Health Agencies
Regulation of restaurant hygiene falls under the jurisdiction of public-health officials in
state, county, or city governments. In particular, local governments establish and implement
food-safety standards for institutional food-service establishments, restaurants, retail food stores,
and other retail food establishments; FDA, through its issuance of the Food Code, advises them
on food-safety guidelines (FDA, 1993;,1997;, 2001; 2005; 2009), inspector training, and
foodborne-illness risk factors (FDA, 2000; 2004; 2009). Those regulatory activities play a
critical role in ensuring food safety. 32
29
FDA collaborates with the National Institutes of Health (NIH) to administer a database of federally and privately
supported clinical trials at ClinicalTrials.gov. The database contains 108,486 trials sponsored by NIH, other federal
agencies, and private industry. Studies listed in the database are conducted in all 50 states and in 174 countries.
Users can access information on current clinical trials, including participant flow, baseline characteristics, outcome
measures and statistical analyses, adverse-events information, administrative information, and study results when
available.
30
The RFR is a new database administered by FDA. Required by Congress, it is an electronic portal for the food
industry and public-health officials to report when there is a reasonable probability that an article of food will cause
serious adverse health consequences. URL: www.fda.gov/Food/FoodSafety/FoodSafetyPrograms/RFR/default.htm
(accessed July 25, 2011).
31
AERS contains over 4 million reports of adverse events from 1969 to the present.
32
Restaurant hygiene has been linked to foodborne disease, so restaurant inspection has been studied as a tool to
reduce the occurrence outbreaks. For example, FDA (2000) checked 895 food establishments across the United
States and found that restaurants and retail store delicatessens were in compliance with the five risk factors
emphasized in the 1997 FDA Food Code only 60–74% of the time. In comparison, the average compliance record
for other food establishments (hospitals, nursing homes, elementary schools, and other departments of retail food
stores) ranged from 76 to 83%. Two followup reports (FDA, 2004, 2009) indicated that some of the risk factors
identified in the 2000 report (such as improper food temperature, poor personal hygiene, and contaminated
equipment) remained in need of attention despite some improvement (FDA, 2010). In 2007, the Center for Science
in the Public Interest examined over 530 inspection reports in 20 cities and found that over 66% of restaurants had at
least one high-risk violation (CSPI, 2008). Those statistics suggest that restaurant hygiene is an important
contributor to outbreaks of foodborne disease.
40
OCR for page 41
Public access to restaurant hygiene-inspection outcomes (Category 1 data) varies greatly
among regions and over time. The most traditional way to share data is “available on request”. In
some cities (such as Pittsburgh and Washington, DC, before 2011), inspection outcomes are
available only through the Freedom of Information Act (FOIA), which requires written requests
and can take up to 6 months for receipt of a final report (CSPI, 2008). In other places (such as
Atlanta and San Francisco), restaurants are required to keep copies of the most recent inspection
reports and provide them on request by consumers. Alternatively, disaggregated inspection
outcomes can be posted on an on-line searchable database; access to these data requires
consumers to initiate an on-line search. Many states and large cities—including Virginia,
Florida, Boston, Chicago, Denver, Philadelphia, and Washington, DC—have adopted on-line
posting.
Several jurisdictions have recently adopted methods that help to deliver restaurant
hygiene-inspection results directly to consumers at the point of sale—the front door or window
of the restaurant. For example, North Carolina, South Carolina, Tennessee, Las Vegas, St. Louis,
Los Angeles, New York City, and some international cities, such as Beijing and Toronto, require
storefront posting of hygiene information. The information can be in the form of a numerical
score of the most recent inspection (usually of a total of 100 points) or broad categories (A–B–C
or pass–conditional pass–closed) based on the numerical scores. The regulatory agency or some
other body must define how the raw inspection results or scores will be used to define the
relevant categories (for example, Category 2, for which the government agency interprets the
disaggregated hygiene data and provides them in a more actionable form for consumers).
REPORTING OF FOOD-SAFETY DATA BY NONREGULATORY AGENCIES
As noted above, although many of the detailed data related to food safety are collected
and reported by regulatory agencies as they engage in normal compliance and enforcement-
related activities (Category 1 data) or are interpreted for consumers (Category 2 data), some
agencies, such as the Centers for Disease Control and Prevention (CDC) and the USDA
Agricultural Marketing Service (AMS), collect and report food-safety data that are generally
reported to them by others, including consumers and health-care professionals. Because
reporting is often voluntary, those data would typically fall into Category 3. The CDC and AMS
data are not linked to individual firms or facilities and thereby differ from FSIS data. They have
the potential to be of benefit to the public but can also affect related firms and industries.
Centers for Disease Control and Prevention
Much of the public-health surveillance for foodborne disease, including outbreaks, is
conducted by state and local health departments. For multistate foodborne-illness outbreak
investigations, CDC plays a prominent role in surveillance and investigation. CDC does not have
authority to mandate that states report their surveillance data to it, but it has developed a system
whereby state and local health departments voluntarily report outbreak data to it. CDC maintains
aggregated and case-based disaggregated foodborne-illness surveillance data in multiple
databases. No personal identifiers are maintained by CDC. Some of the data are publicly
accessible, and others are available only through FOIA.
41
OCR for page 42
In 2009, CDC launched the Foodborne Outbreak Online Database (FOOD), 33 which is
designed to allow the public direct access to state-level information on foodborne outbreaks. The
database spans 1998–2008 and is updated periodically. FOOD enables the public to search and
download data on reported outbreaks as an XML file. It does not identify specific establishments
involved in outbreaks.
A recent report suggested that the FOOD data have several limitations. 34 For example,
state health departments may update the data at any time, so these entries are never considered
“final”. The rigor with which state health departments collect and report data can vary widely
and some users of the data have noted inconsistencies in the dataset. In addition, data are not
updated in real time, so the most current data available are usually several years old.
CDC also collects surveillance data through the Foodborne Diseases Active Surveillance
Network (FoodNet) program. FoodNet is a partnership of 10 state and local health departments,
CDC, FDA, and USDA that conducts population-based surveillance for laboratory-confirmed
infections commonly implicated in foodborne disease. CDC releases annual summaries of
FoodNet data in a published report but does not make the raw data publicly available, although
they can be requested through FOIA.
Agricultural Marketing Service
The Monitoring Programs Office of USDA’s AMS is responsible for managing the
Pesticide Data Program (PDP), 35 a voluntary program that was implemented in 1991 to test food
commodities for pesticide residues. The PDP is based on a sampling plan with a rigorous
statistical design to ensure the reliability of the data for use in exposure assessments. However,
the pesticide databases are commodity-specific rather than establishment-specific.
Every year, the AMS publishes on line a report summarizing the PDP data. It includes the
study design for data collection for the relevant year, details about how data are reported, a
summary of the results, the history of commodities tested, and the raw commodity-level data.
Requests for PDP information are received from many parties, including other government
agencies and various organizations, and the staff of the Monitoring Programs Office generates
specific reports for these queries. However, users can also import the data into database-
management software and conduct their own analyses.
The PDP was not designed for the purpose of enforcing regulations. The major user of
the PDP data is EPA, which uses them in its pesticide risk assessments and to estimate whether
human exposure exceeds safety standards. 36 FDA is informed of residues that exceed tolerances
33
See http://wwwn.cdc.gov/foodborneoutbreaks/ (accessed June 8, 2011).
34
See http://www.aei.org/docLib/REG-2011-02-g.pdf (accessed June 8, 2011).
35
See
http://www.ams.usda.gov/AMSv1.0/ams.fetchTemplateData.do?template=TemplateC&navID=PDPOviewBox2Link
1&rightNav1=PDPOviewBox2Link1&topNav=&leftNav=ScienceandLaboratories&page=PesticideDataProgram&r
esultType=&acct=pestcddataprg (accessed on June 23, 2011).
36
Other groups also use the data to provide information to the lay public about pesticide residues in foods (see, for
example, http://www.ewg.org/foodnews/), and researchers have conducted analyses with the PDP residue data (see,
for example, Punzi et al., 2005; Baker et al., 2002; Kuchler et al., 1996).
42
OCR for page 43
or that have no tolerances. The program is voluntary (AMS has no regulatory authority to require
participation in the program) and works with 12 state agencies that are responsible for sample
collection and analysis. Five states were selected initially because they were diverse geographic
areas; the list was later expanded to 12 states to increase the sampling data points (M. Lamont,
USDA AMS, Manassas, VA, personal communication, July 20, 2011).
REPORTED EFFECTS OF RELEASING ESTABLISHMENT-SPECIFIC DATA
The overview above suggests that other government agencies have already had
considerable experience with the release of detailed data. The academic literature also has
examined the pros and cons of information disclosure in many contexts, including disclosure of
establishment-specific regulatory information similar to the FSIS data and disclosure of product-
specific information that may be traced back to manufacturers. The committee reviewed the
many National Research Council and Institute of Medicine reports on data-sharing (for example,
NRC, 1985; NRC and the Social Science Research Council, 1993; IOM, 1996; NRC, 2000;
2001, 2005; NAS, 2009; IOM and NRC, 2010). The reports have a somewhat different focus,
and none addresses directly the issue of publicly releasing data gathered originally for regulatory
purposes; for example, NRC (1985) focuses on data-sharing among researchers. However, many
of the issues associated with data-sharing in other settings, as discussed in these National
Research Council reports and related documents, do address benefits and concerns related to the
process and point to conclusions that are similar to those discussed here and in Chapter 4. For
example, the report Sharing Research Data (NRC, 1985) considers the benefits and costs of
data-sharing among researchers. The noted benefits include promoting and improving research
that leads to better decisions and improving measurement and data-collection methods. The costs
include technical obstacles to sharing data and the costs of documentation and training. The
entities that own and disclose data may go beyond regulatory agencies (to manufacturers and
third-party certifiers), but lessons learned from the broader literature can help us to anticipate the
specific potential effects of releasing establishment-specific FSIS data.
As discussed in Chapter 2, the effectiveness of any transparency system is based on
whether the information that it provides is “embedded” in the action cycle of users (such as
consumers) and disclosers (such as businesses) of the information (Fung et al., 2007). On the
positive side, the literature suggests that information disclosure may enable consumers to make
more informed choices. Analysis of specific policies yields numerous examples of information
affecting consumer choice. The public posting of hygiene inspection outcomes has resulted in
increased sensitivity to restaurant hygiene (Jin and Leslie, 2003). In the health-care domain,
substantial patient sorting has been observed in response to the disclosure of cardiac-surgery
outcomes associated with hospitals and doctors (Dranove et al., 2003). Increased consumption of
fiber-rich cereals has occurred on disclosure of the nutrition content of food products (Ippolito
and Mathios, 1990).
In addition, there is evidence that disclosure of firm-specific or facility-specific
information can motivate firms to improve their performance, at least along the disclosed
dimensions. For example, evidence shows that public posting of restaurant hygiene information
led to better public-health outcomes in Los Angeles and Toronto (Jin and Leslie, 2003; Simon et
al., 2005; Serapiglia et al., 2007). EPA TRI disclosure led to substantial improvements in
environmental performance (Konar and Cohen, 1997). Likewise, for large Massachusetts water
43
OCR for page 44
suppliers, the mandatory public provision of information about violations of drinking-water
standards resulted in a 30–44% reduction in total violations and a 40–57% reduction in more
severe health violations (Bennear and Olmstead, 2008). A recent study of state voluntary site-
cleanup programs revealed that public disclosure of contaminated sites is an efficient tool for
promoting participation of property managers and developers in site remediation (Blackman et
al., 2010). Patten (2002) further argues that release of establishment-level TRI data generated
public-policy pressure that led to increased environmental disclosure by TRI firms. Those
improvements can be in response to consumer pressures of the type discussed in the previous
paragraph, pressures from input markets (such as investors and suppliers), and actual or
threatened regulatory pressures (Fung et al., 2007).
Release of establishment-level data has also generated research opportunities. For
example, researchers have used establishment-level enforcement data to examine the
effectiveness of inspection programs in a variety of contexts, including mine safety (Kniesner
and Leeth, 2004), occupational safety (Bartel and Thomas, 1985; Gray and Jones, 1991; Weil,
1996; 2001), nuclear safety (Feinstein, 1989), seafood safety (Alberini et al., 2008), air and water
pollution (Magat and Viscusi, 1990; Gray and Deily, 1996; Earnhart, 2004), and pharmaceutical
production (Macher et al., 2011). Establishment-level data have also been used to study issues
not directly related to enforcement, such as the link between air pollution and fetal or infant
mortality (Agarwal et al., 2010), the effectiveness of nonregulatory programs (e.g., Arora and
Cason, 1996), environmental justice (Daniels and Friedman, 1999; Dolinoy and Miranda, 2004),
interjurisdictional pollution effects (Helland and Whitford, 2003), and the effects of physician
prescription of drug combinations on competition among pharmaceutical firms (Lucarelli et al.
2010).
The evidence reviewed suggests that public disclosure of establishment-specific data can
have important social benefits. However, as with all regulatory interventions, some parties may
be adversely affected by public data disclosure. Different parties have different perspectives on
what constitutes an adverse effect. In fact, a negative for one party might be viewed as a positive
by another or ultimately considered as a positive by the public at large. One potential adverse
effect is related to the market. For example, a body of literature demonstrates that some firms
suffered reductions in their stock prices immediately after public release of data (e.g., Hamilton,
1995; 2005). There are similar examples of the effect of food recalls on stock prices (e.g., Salin
and Hooker, 2001; Thomsem and McKenzie, 2001). Konar and Cohen (1997) reported that firms
more adversely affected by the release of TRI data were later more likely to reduce their toxic
releases; this suggests that the adverse effects of the data release motivated firms to improve
their performance (in response to pressure from investors in capital markets). Thus, the adverse
effect on some firms may ultimately generate benefits for the broader community.
Some researchers have raised concerns about potential adverse effects of disclosure due
to misinterpretation or lack of understanding of the data. If that occurs, the disclosure might not
have the intended effect. For example, the terrorist color-coded threat advisory system that was
enacted shortly after the 9/11 attacks and was in effect until early 2011 provided vague
information that tended to cause confusion, alarm, or eventually disregard by the public but little
evidence of reduction of risk to the public (Fung et al., 2007). The implication, however, is not
that the data should not be released but rather that the data should be provided in a meaningful
and understandable form.
Another concern raised in the literature is that information disclosure may encourage
firms to improve on the reported outcomes but reduce performance regarding unreported
44
OCR for page 45
outcomes, especially when the omitted outcomes are unreported because of measurement
difficulty rather than because of lack of importance. This type of distortionary behavior has
been documented in a number of contexts. For example, Khanna et al. (1998) showed that
release of TRI information reduced on-site releases of toxic substances but increased transfers
of the same substances to off-site locations. As a result, in that particular case, the authors
concluded that the overall effect on the amount of toxic waste generated was negligible. Similar
examples can be found in the context of medical outcomes (Dranove et al., 2003) and school
performance (Haney, 2000; Deere and Strayer, 2001; Jacob and Levitt, 2003; Hanushek and
Raymond, 2005; Jacob, 2005; Cullen and Reback, 2006; Figlio and Getzler, 2006). The
potential for distortionary behavior suggests that agencies contemplating public data disclosure
should anticipate such responses by firms and design information collection and disclosure
policies that will reduce that kind of behavior.
Information disclosure not only has the potential to distort firm behavior but can add
pressure on the people (such as inspectors) who generate data in the field. On the one hand,
inspectors may be under closer scrutiny and thus pressured to do their jobs in a more precise and
consistent way. For example, there is evidence that increased public attention after the Three
Mile Island accident increased inspector detection rates (Feinstein, 1989); this suggests that
public attention on inspection may motivate inspectors to do a better job. On the other hand,
firms identifiable in the disclosure data have incentives to ask for leniency of the inspectors who
are assigned to their facilities. Anecdotal evidence has shown inspector bribery after Los Angeles
County adopted restaurant hygiene report cards, and data plots raise concern about leniency
regarding the cutoffs of letter grades (90 for A and 80 for B, Jin and Leslie, 2003; 2005).
The concern about inspector bias and potential bribery brings up issues regarding
heterogeneity in inspector performance. In nuclear-safety inspection (Feinstein, 1989) and
inspection of pharmaceutical manufacturing plants (Macher et al., 2011), researchers have found
that inspector identity or inspector demographics, experience, or training is important in
explaining inspection outcomes. That suggests that inspectors vary in their ability to detect or
their preference in detecting violations. Although the committee is not aware of any published
study that documents the effect of data disclosure on inspector behavior, public attention after
data disclosure may highlight the existence of inspector heterogeneity and motivate the provision
of additional training and standardization to enhance inspection consistency.
Ironically, distorted firm and inspector behavior that occurs as a consequence of
information disclosure suggests that the disclosed data are useful at least in the perception of
primary data users. That highlights the importance of what to disclose, how to disclose it
(including how to protect the identities of individual inspectors), and what additional support
might be needed from FSIS to facilitate proper data use. The experience of reporting hospital
outcomes may be informative. Given the complexity of raw data and consumer demand for easy-
to-read information, hospital outcomes are often reported in averages. However, the precision of
an average measure varies greatly with sample size, frequency of the measured event, and the
pool of subjects that contribute to the sample. Researchers have shown that measurement
problems can compromise the usefulness of disclosed data (Iezzoni, 1997; Kane and Staiger,
2002), but disclosure brings measurement issues to the forefront and thereby promotes research
that can lead to improvements.
Restaurant hygiene report cards provide another useful lesson: that many factors
contribute to the effectiveness of data disclosure. For example, in addition to issuing grade cards,
Los Angeles County adopted an easy-to-read format for the grade cards, inspected some
45
OCR for page 46
restaurants more frequently, provided restaurant inspectors with additional training, and put in
more effort to educate restaurant owners and staff (LADPH, 2008). Several studies have found
that sanitary conditions in restaurants could be improved through more frequent inspections and
enhanced education efforts (Bader et al., 1978; Mathias et al., 1995; Cotterchio et al., 1998;
Allwood et al., 1999; Cates et al., 2009; Hislop and Shaw, 2009). The Los Angeles restaurant
hygiene report cards were motivated by a CBS 2 news program that revealed, through the use of
hidden cameras, the unsanitary conditions in restaurant kitchens. That TV exposé 37 increased
consumer awareness of restaurant hygiene, drew attention to the weakness of the existing
system, and intensified political pressure for regulatory change. Another factor that potentially
contributed to the success of Los Angeles grade cards is that the stringent inspection codes
matched specific violations to defined numerical point deductions, which minimized the
subjectivity of hygiene inspections. This system contributed to a more standardized evaluation
among restaurants and inspectors, and increased consumer confidence in the grade cards.
On-line posting has become a new norm for data disclosure because of its low cost and
the ease of user access. Research is needed to examine the advantages and disadvantages of on-
line posting relative to those of other methods of disclosure, such as posted restaurant report
cards or published hospital rankings. The Internet facilitates posting of large amounts of data
and allows user customization, but access to the Internet is probably skewed toward a set of the
population (those of higher income and those who are better educated) and often requires
expertise and effort by end users to analyze and interpret the data correctly. 38 However, the
investment of time and expertise required to analyze and interpret large datasets appropriately is
not peculiar to their release on the Internet, and the costs and knowledge necessary to obtain the
same data through FOIA requests are potentially even greater barriers to the dissemination of
the information.
At a minimum, posting data on the Internet would make it easier for the public to know
what kinds of information have been collected and are available and perhaps to gain some initial
understanding of the quality, complexity, and potential usability of the data for specific purposes.
Hence, the public may avoid the current costs of obtaining data through a FOIA request that
would ultimately be unsuitable for their needs. However, posting on the internet may increase the
potential for misinterpretation, if only by virtue of the fact that releasing data more broadly (via
the Internet) will result in a higher number of users and uses. The experience of other federal
agencies in posting data suggests the benefit of providing that information in formats and with
37
Behind the Kitchen Door: Joel Grover Investigation. First Broadcast November 1997.
38
The original discussion of the "digital divide" arose as a result of a survey conducted by the National
Telecommunication and Information Administration (NTIA) of the US Department of Commerce in 1994 that
showed differences in use of the emerging Internet by income and demographic characteristics. A second survey by
NTIA in 1998 (with the subtitle "Defining the Digital Divide") provided further evidence of gaps in use. The
surveys precipitated studies in the United States, the United Kingdom, and elsewhere that look into the causes and
consequences of differential use (see, for example, Norris, P. 2001, Digital Divide, Civic Engagement, Information
Poverty, and the Internet Worldwide. Cambridge, UK: Cambridge University Press). How much the gap has
narrowed in recent years with respect to income, ethnicity, and age is controversial.
46
OCR for page 47
documentation that facilitate its analysis and interpretation. The committee believes that FSIS
would be best suited to determine how to address misinterpretation of data on a case-by-case
basis.
SUMMARY
A number of federal agencies (none with specific food-safety jurisdiction) release
detailed data that are directly linked to the performance of individual facilities or firms or to the
products that they produce. In many cases, the data originate in regulatory (compliance and
enforcement) activities. A substantial body of literature documents the effects of the public
release of data and their uses. The literature suggests that release of facility-specific performance
data can have both benefits and costs (or unintended adverse consequences).
Major benefits include enabling users to make more informed choices, motivating
facilities to improve their performance, and provision of data for use in research studies of
regulatory effectiveness and other performance-related issues. The possible costs of public
disclosure of information include adverse effects on profitability, but it is precisely this
possibility that creates an incentive for facilities to improve their performance. The literature has
also raised concerns about some perhaps unintended consequences, including the potential for
data misinterpretation, the incentive for establishments whose data are disclosed to “game the
system”, 39 and potential pressure on inspector performance.
Based on its review of the entirety of the extant literature, the committee concluded that
the potential adverse impacts, while possible, were largely anecdotal or speculative, and are not
backed up by any significant systematic evidence. On the other hand, the positive benefits are
more credibly backed up by the scientific literature. Therefore, the current evidence of adverse
effects is insufficient for predicting specific problems that would be inherent in the release of
establishment-specific FSIS data. The committee believes that the potential for adverse effects is
not necessarily insurmountable but highlights the need to pay careful attention to the design of an
information-disclosure strategy. For example, potential adverse effects may be minimized if the
disclosing entity (FSIS) is careful to ensure the integrity of the data and provides precise and
appropriate definitions of what is being quantified, adequate documentation of context, a means
by which to support analysis of the data by users, and precautionary measures to prevent the
linking of portions of the data in ways that would allow users to deduce confidential information
about particular establishments. It is clear that the most effective disclosure systems improve in
quality, quantity, and scope as users gain a better understanding of how the data might be used.
REFERENCES
Agarwal, N., Banternghansa, C., and L.T.M Bui. 2010. Toxic exposure in America: Estimating
fetal and infant health outcomes from 14 years of TRI reporting. Journal of Health
Economics 29(4):557-74.
Alberini, A., Lichtenberg, E., Mancini, D., and G. Galinato. 2008. Was it something I ate?
Implementation of the FDA Seafood HACCP Program. American
Journal of Agricultural Economics 90(1):28-41.
39
To adjust their strategies to minimize public disclosure.
47
OCR for page 48
Allwood, P.B., Lee, P., and P. Borden-Glass. 1999. The vital role of restaurant health
inspections. Journal of Environmental Health 61(9):25-28.
Arora, S., and T. N. Cason. 1996. Why do firms volunteer to exceed environmental regulations?
Understanding participation in EPA's 33/50 program. Land Economics 72(4):413-32.
Bader, M., Blonder, E., Henriksen, J., and W. Strong. 1978. A study of food service
establishment sanitation inspection frequency. American Journal of Public Health
68(4):408-410.
Baker, B.P., Benbrook, C.M., Groth, E. III, and K. Lutz Benbrook. 2002. Pesticide residues in
conventional, integrated pest management (IPM)-grown and organic foods: insights from
three US data sets. Food Additives and Contaminants 19(5):427-446.
Behind the Kitchen Door: Joel Grover Investigation. First Broadcast November 1997.
Blackman, A., Darley, S., Lyon, T.P., and K. Wernstedt. 2010. What Drives Participation in
State Voluntary Cleanup Programs? Evidence from Oregon. Land Economics, 86(4):
785-799.
Bartel, A.P., and L.G.Thomas. 1985. Direct and indirect effects of regulation: New look at
OSHA’s impact. Journal of Law and Economics 28:1-25.
Bennear, L.S., and S.M. Olmstead. 2008. The impacts of the 'right to know': Information
disclosure and the violation of drinking water standards. Journal of Environmental
Economics and Management. September 56(2):117-30.
Cates, S.C., Muth, M.K., Karns, S.A., Penne, M.A., Stone, C.N., Harrison, J. E., and V. J. Radke.
2009. Certified kitchen managers: Do they improve restaurant inspection outcomes?
Journal of Food Protection 72(2):384-391.
Center for Science in the Public Interest (CSPI). 2008. Dirty Dining: Have Reservations? You
Will Now. Available at www.cspinet.org/new/pdf/ddreport.pdf. Accessed June 8, 2011.
Cotterchio, M., Gunn, J., Coffill, T., Tormey, P., and A.M. Barry. 1998. Effect of a manager
training program on sanitary conditions in restaurants. Public Health Reports 113:353-
358.
Cullen, J.B., and R. Reback. 2006. Tinkering toward accolades: school gaming under a
performance accountability system. Pp. 1-34 in Improving School Accountability:
Check-Ups or Choice, Advances in Applied Microeconomics 14, T. Gronberg and D.
Jansen, eds. Amsterdam: Elsevier Science.
Daniels, G., and S. Friedman. 1995. Spatial inequality and the distribution of industrial toxic
releases: Evidence from the 1990 TRI. Social Science Quarterly 80(2):245-261.
Deere, D., and W. Strayer. 2001. Putting Schools to the Test: School Accountability, Incentives
and Behavior. Working paper 114. Department of Economics, Texas A&M University.
Dolinoy, D.C. and M.L Miranda. 2004. GIS Modeling of air toxics releases from TRI-Reporting
and Non-TRI-Reporting facilities: Impacts for environmental justice. Environmental
Health Perspectives 112(17):1717-1724.
Dranove, D., Kessler, D., McClellan, M., and M. Satterthwaite. 2003. Is more information
better? The effects of "report cards" on health care providers. The Journal of Political
Economy 111:555-588.
Earnhart, D. 2004. Panel data analysis of regulatory factors shaping environmental performance.
Review of Economics and Statistics 86(1):391-401.
FDA (Food and Drug Administration). Food Code. 1993. U.S. Public Health Service, U.S. Dept.
of Commerce. Technology Administration, National Technical Information Service. Pub.
No. PB94-113941AS. Springfield, VA.
48
OCR for page 49
FDA. 1997. Food Code 1997. Available at
http://www.fda.gov/Food/FoodSafety/RetailFoodProtection/FoodCode/FoodCode1997/de
fault.htm. Accessed June 8, 2011.
FDA. 2000. Report of the FDA Retail Food Program Database of Foodborne Illness Risk
Factors. Available at
http://www.fda.gov/Food/FoodSafety/RetailFoodProtection/FoodborneIllnessandRiskFac
torReduction/RetailFoodRiskFactorStudies/ucm123544.htm. Accessed June 8, 2011.
FDA. 2001. Food Code 2001. Available at
http://www.fda.gov/Food/FoodSafety/RetailFoodProtection/FoodCode/FoodCode2001/de
fault.htm. Accessed June 8, 2011.
FDA. 2004. FDA Report on the Occurrence of Foodborne Illness Risk Factors in Selected
Institutional Foodservice, Restaurant, and Retail Food Store Facility Types (2004).
Available at
http://www.fda.gov/Food/FoodSafety/RetailFoodProtection/FoodborneIllnessandRiskFac
torReduction/RetailFoodRiskFactorStudies/ucm089696.htm. Accessed June 8, 2011.
FDA. 2005. Food Code 2005. Available at
http://www.fda.gov/Food/FoodSafety/RetailFoodProtection/FoodCode/FoodCode2005/de
fault.htm. Accessed June 8, 2011.
FDA. 2009. FDA Report on the Occurrence of Foodborne Illness Risk Factors in Selected
Institutional Foodservice, Restaurant, and Retail Food Store Facility Types (2009).
Available at
http://www.fda.gov/Food/FoodSafety/RetailFoodProtection/FoodborneIllnessandRiskFac
torReduction/RetailFoodRiskFactorStudies/ucm224321.htm. Accessed June 10, 2011.
FDA. 2009. Food Code 2009. Available at
http://www.fda.gov/Food/FoodSafety/RetailFoodProtection/FoodCode/FoodCode2009/.
Accessed June 8, 2011.
FDA. 2010. FDA Trend Analysis Report on the Occurrence of Foodborne Illness Risk Factors in
Selected Institutional Foodservice, Restaurant, and Retail Food Store Facility Types
(1998-2008). Available at
http://www.fda.gov/Food/FoodSafety/RetailFoodProtection/FoodborneIllnessandRiskFac
torReduction/RetailFoodRiskFactorStudies/ucm223293.htm. Accessed June 10, 2011.
Feinstein, J. 1989. The Safety Regulation of U.S. Nuclear Power Plants: Violations, Inspections,
and Abnormal Occurrences. Journal of Political Economy 97(1):115-154.
Figlio, D., and L. Getzler. 2006. Accountability, Ability and Disability: Gaming the System.
Available at http://papers.nber.org/papers/w9307.pdf. Accessed August 29, 2011.
Fung, A., Graham, M., and D. Weil. 2007. Full Disclosure: The Perils and Promise of
Transparency. New York: Cambridge University Press.
Gerde, V.W., and J.M. Logsdon. 2001. Measuring environmental performance: Use of the
Toxics Release Inventory (TRI) and other U.S. environmental databases. Business
strategy and the environment 10:269–285.
Gray, W.B., and C.A. Jones.1991. Longitudinal patterns of compliance with Occupational Safety
and Health Administration health and safety regulations in the manufacturing sector.
Journal of Human Resources 26(4):623–653.
Gray, W.B., and M.E. Deiley. 1996. Compliance and enforcement: Air pollution regulation in
the U.S. steel industry. Journal of Environmental Economics and Management 31(1):96-
111.
49
OCR for page 50
Hamilton, J. 2005. Regulation through Revelation: The Origins, Politics, and Impacts of the
Toxics Release Inventory Program. New York: Cambridge University Press.
Hamilton, J.T. 1995. Pollution as news: Media and stock market reactions to the toxics release
inventory data. Journal of Environmental Economics and Management 28(1):98-113.
Haney, W. 2000. The Myth of the Texas Miracle in Education. Available at
http://epaa.asu.edu/ojs/article/view/432. Accessed August 29, 2011.
Hanushek, E.A., and M.E. Raymond. 2005. School Accountability and Student Performance.
Available at
http://edpro.stanford.edu/hanushek/admin/pages/files/uploads/HanushekRaymond.pdf.
Accessed August 29, 2011.
Helland, E., and A.B. Whitford. 2003. Pollution incidence and political jurisdiction: evidence
from the TRI. Journal of Environmental Economics and Management 46:403–424.
Hislop, N., and K. Shaw. 2009. Food safety knowledge retention study. Journal of Food
Protection 72(20):431-435.
Iezzoni, L.I. 1997. Assessing quality using administrative data. Annals of Internal Medicine
127(8):666-674. Available at
http://www.annals.org/content/127/8_Part_2/666.full.pdf+html. Accessed August 29,
2011.
IOM (Institute of Medicine). 1996. Resource Sharing in Biomedical Research. Washington,
DC: The National Academy Press.
IOM. 2000. Protecting Data Privacy in Health Services Research. Washington, DC: The
National Academy Press.
IOM/NRC. 2010. Enhancing Food Safety: The Role of the Food and Drug Administration.
Washington, DC: The National Academies Press.
Ippolito, P., and A. Mathios. 1990. Information, advertising and health: A study of the cereal
market. Rand Journal of Economics 21(3):459-480.
Jacob, B.A. 2005. Accountability, incentives and behavior: The impact of high-stakes testing in
the Chicago public schools. Journal of Public Economics 89:761-796. Available at
http://faculty.smu.edu/millimet/classes/eco7321/papers/jacob.pdf. Accessed August 29,
2011.
Jacob, B.A. and S.D. Levitt. 2003. Catching Cheating Teachers: The Results of an Unusual
Experiment in Implementing Theory. Pp. 185-209 in Brookings-Wharton Papers on
Urban Affairs 2003, W.G. Gale and J. Rothenberg Pack, eds. Washington, DC:
Brookings Institution Press.
Jin, G.Z., and P. Leslie. 2003. The effect of information on product quality: Evidence from
restaurants hygiene grade cards. The Quarterly Journal of Economics 118(2):409-451.
Jin, G.Z., and P. Leslie. 2005. The case in support of restaurant hygiene grade cards. Choices
20(2):97-102. Also available at http://www.choicesmagazine.org/2005-2/safety/2005-2-
02.htm. Accessed July 2011.
Kane, T., and D. Staiger. 2002. The promise and pitfalls of using imprecise school accountability
measures. Journal of Economic Perspectives 16:91-115.
Khanna, M., Quimio, W., Rose, H., and D. Bojilova. 1998. Toxics release information: A policy
tool for environmental protection. Journal of Environmental Economics and Management
36(3):243-66.
50
OCR for page 51
Kniesner, T. J., and J.D. Leeth. 2004. Data Mining Mining Data: MSHA Enforcement Efforts,
Underground Coal Mine Safety, and New Health Policy Implications. Journal of Risk and
Uncertainty 29(2): 83-111, 09.
Konar, S., and M. A. Cohen. 1997. Information as Regulation: The Effect of Community Right
to Know Laws on Toxic Emissions. Journal of Environmental Economics and
Management 32(1):109-24.
Kuchler, F., Chandran, R., and K. Ralston. 1996. The linkage between pesticide use and
pesticide residues. American Journal of Alternative Agriculture 11(4):161-167.
Los Angeles County Department of Public Health (LADPH). 2008. 10-year Review of
Restaurant and Food Facility Grading Program. Available at
http://file.lacounty.gov/bc/q1_2008/cms1_082885.pdf. Accessed June 8, 2011.
Lucarelli, C., Nicholson, S., and M. Song. 2010. Bundling Among Rivals: A Case of
Pharmaceutical Cocktails. NBER Working Paper Series. Available at
http://www.nber.org/papers/w16321.pdf. Accessed July 10, 2011.
Macher, J.T., Mayo, J.W., and J.A. Nickerson. 2011. Regulator heterogeneity and endogenous
efforts to close the information asymmetry gap. Journal of Law and Economics 54:25-54.
Magat, W.A., and W.K. Viscusi. 1990. Effectiveness of the EPA's regulatory enforcement: The
case of industrial effluent standards. Journal of Law and Economics 33(2):331-60.
Mathias, R.G., Sizto, R., Hazlewood, A., and W. Cocksedge. 1995. The effects of inspection
frequency and food handler education on restaurant inspection violations. Canadian
Journal of Public Health 86(1):46-50.
National Academy of Sciences. 2009. Ensuring the Integrity, Accessibility, and Stewardship
of Research Data in the Digital Age. Washington, DC: The National Academies Press.
NRC (National Research Council). 1985. Sharing Research Data. Washington, DC: The
National Academy Press.
NRC and the Social Science Research Council. 1993. Private Lives and Public Policies:
Confidentiality and Accessibility of Government Statistics. Washington, DC: The
National Academy Press.
NRC. 2000. Improving Access to and Confidentiality of Research Data: Report of a
Workshop. Washington, DC: The National Academy Press.
NRC. 2001. Studies of Welfare Populations: Data Collection and Research Issues.
Washington, DC: The National Academy Press.
NRC. 2005. Expanding Access to Research Data: Reconciling Risks and Opportunities.
Washington, DC: The National Academies Press.
Patten, D.M. 2002. Media exposure, public policy pressure, and environmental disclosure: An
examination of the impact of TRI data availability. Accounting Forum 26(2):152-171.
Punzi, J. S., Lamont, M., Haynes, D., and R.L. Epstein. 2005. USDA Pesticide Data Program:
Pesticide residues on fresh and processed fruit and vegetables, grains, meats, milk, and
drinking water. Outlooks on Pest Management 16(3):131-137.
Salin, V., and N. H. Hooker. 2001. Stock market reaction to food recalls. Review of Agricultural
Economics 23(1):33-46.
Serapiglia, T., Kennedy, E., Thompson, S., and R. de Burger. 2007. Association of food premises
inspection and disclosure program with retail-acquired foodborne illness and operator
noncompliance in Toronto. Journal of Environmental Health 70(1):54-59.
51
OCR for page 52
Simon, P., Leslie, P., Run, G., Jin, G.Z., Reporter, R., Aguirre, A., and J.E. Fielding. 2005.
Impact of restaurant hygiene grade cards on foodborne disease hospitalizations in Los
Angeles County. Journal of Environmental Health 67:32-36.
Thomsem, M.R., and A.M. McKenzie. 2001. Market incentives for safe foods: an examination of
shareholder losses from meat and poultry recalls. American Journal of Economics
83(3):526-538.
Weil, D. 1996. If OSHA is so bad why is compliance so good? RAND Journal of Economics
27(3):618-640.
Weil, D. 2001. Assessing OSHA performance: Evidence from the construction industry. Journal
of Policy Analysis and Management 20(4):651-674.
USDA/FSIS. 2010. Quarterly Enforcement Report. Available at
http://www.fsis.usda.gov/PDF/QER_Q1_FY11_Tables1-19.pdf. Accessed August 18,
2011.
52
