Quality of data has many faces. Primarily, it has to be relevant (i.e., useful) to its users. Relevance is achieved through a series of steps starting with a planning process that links user needs to data requirements. It continues through acquisition of data that is accurate in measuring what it was designed to measure and produced in a timely manner. Finally, the data must be made accessible and easy to interpret for the users. In a more global sense, data systems also need to be complete and comparable (to both other data systems and to earlier versions). The creation of data that address all of the facets of quality is a unified effort of all of the development phases from the initial data system objectives, through system design, collection, processing, and dissemination to the users. These sequential phases are like links in a chain. The sufficiency of each phase must be maintained to achieve relevance. This document is intended to help management and data system "owners" achieve relevance through that sequential process.
The 1991 Intermodal Surface Transportation Efficiency Act (ISTEA) created the Bureau of Transportation Statistics (BTS) within the Department of Transportation (DOT). Among other things, it made BTS responsible for: "issuing guidelines for the collection of information by the Department of Transportation required for statistics ... in order to ensure that such information is accurate, reliable, relevant, and in a form that permits systematic analysis." (49 U.S.C. 111 (c)(3))
A parallel requirement for developing guidelines emerged in the Paperwork Reduction Act of 1995. It tasked the Office of Management and Budget (OMB) to "develop and oversee the implementation of Government wide policy, principles, and guidelines concerning statistical collection procedures and methods; statistical data classification; statistical information presentation and dissemination; timely release of statistical data; and such statistical data sources as may be required for the administration of federal programs." (44 U.S.C. 3504 (e)(3))
Lastly, the Consolidated Appropriations Act of 2001, section 515, elaborated on the Paperwork Reduction Act, requiring OMB to issue guidelines ensuring the quality of disseminated information by 9/30/2001 and each federal agency to issue guidelines by 9/30/2002.
On 28 September 2001, OMB published a notice in the Federal Register (finalized as 67 FR 8452, February 22, 2002) that required agencies to issue guidelines for ensuring and maximizing the quality of information disseminated by federal agencies.
As defined in the OMB guidance, quality consists of:
Agencies were required to develop guidelines, covering all information disseminated on or after October 1, 2002, regardless of format. Agencies were also required to develop a process for pre-dissemination review of information, an administrative mechanism allowing the public to request correction of information not complying with the guidelines, and an annual report to OMB indicating how the public requests were handled by the mechanism.
These guidelines incorporate the statistical aspects of the OMB guidelines as a baseline and elaborate on its recommendations to produce statistical guidelines adapted for the Department of Transportation.
These guidelines apply to all statistical information that is disseminated on or after 1 October 2002 by agencies of the Department of Transportation (DOT) to the public using the "dissemination" definition in the OMB guidelines. That definition exempts a number of classes of information from these guidelines. Major types of exempted information are listed below. A more detailed list is provided in section IV of the DOT Information Dissemination Quality Guidelines, of which this document is a subsection.
DOT disseminated data contain a lot of information provided by "third party sources" like the states, industry organizations, and other federal agencies. These guidelines apply to that disseminated data unless exempted for other reasons discussed above. However, DOT guidelines indicating design, collection, and processing methods do not apply to data acquisition steps performed by non-federal sources. Steps performed by federal sources outside DOT before providing the data to DOT will be governed by the agency's own guidelines in accordance with this legislation. For data provided to DOT by third party sources, these guidelines primarily emphasize disseminating information about data quality, the DOT processing methods, and analysis of the data provided to the users.
The recommendations within these guidelines apply to a wide range of data collection types. They include reporting collections, surveys, and special studies.
Reporting collections are set up to be automatic delivery of data into the data system. They collect incident information from government (federal, state, or local) and industry sources and periodic information on transportation flow and volume from government and industry. Incident data tend to cover all incidents (e.g., fatal accidents), though some data may be sampled due to its sheer volume (e.g., highway injuries). Flow and volume collections are a mixture of 100% collection and sampled data. Surveys and special studies are more of an outreach form of data collection. Surveys and studies are usually conducted using some form of sampling.
Samples taken for any data collection may be selections of people or organizations from lists, samples of geographic areas or sections of highway, or samples of time segments.
The quality guidelines for statistical information are based on structured planning (section 2), sound statistical methods (sections 3 and 4) and the principle of openness (sections 5 and 6). Structured planning maintains the link between user needs and data system design. Sound statistical methods produce information (data and analysis results) that conforms to that design. Openness ensures that users of statistical information can easily access and interpret the information.
Each section begins with a statement of principles, which contain definitions, assumptions, and rules or concepts governing action. The principles are followed by guidelines, which are specific recommended actions with examples. Finally, each section concludes with references.
These statistical guidelines are a subset of the DOT Information Dissemination Quality Guidelines. Chapters 2 through 6 discussed above form section VI, paragraphs a - e in that document.