|
Chapter 6 Dissemination of Information
Dissemination is the distribution of information to the
public, in any medium or form, including press releases, reports, data files,
or web products. These standards cover
releasing information (Section 6.1) and ensuring the accuracy and interpretability
of different types of BTS information products:
tables, graphs, and maps (Section 6.2), text (Section 6.3), and micro
data (Section 6.4). The standards also cover
issues affecting all information products:
data protection (Section 6.5), rounding (Section 6.6), and revisions (Section
6.7). Finally, the public documentation
standard (Section 6.8) provides for the transparency and reproducibility of the
information disseminated by BTS.
6.1 Releasing Information
Standard 6.1: Procedures for the release of information
products to the public must receive predissemination reviews (disclosure, content
matter, statistical and methodological) and must include provisions for
ensuring fair access to all users.
Key Terms: information
product, peer review, sensitive material
Guideline 6.1.1: Release Schedules
To provide fair access to the public,
major information products should follow published release schedules.
- Provide the schedule
for the release of information products to the BTS public affairs office for
release.
- Protect information
to be published against any unauthorized pre-release or disclosure in advance
of the publication schedule.
Guideline 6.1.2: Ease of Accessibility and Understanding
Information products should be made
accessible to the public.
- All information products
disseminated through the Internet should comply with the requirements for
Section 508 of the Rehabilitation Act of 1973, as amended.
- Codes, abbreviations, and acronyms should be used sparingly and defined in accordance with the BTS Guide to Style and Publishing Procedures. Provide definitions to the user in the product.
- As appropriate, information products should also include definitions of any subject-matter-specific or otherwise technical terms.
Guideline 6.1.3: Formal Pre-Dissemination Review
All information products require
pre-dissemination review to ensure compliance with OMB and DOT Information
Quality Guidelines, and BTS standard procedures.
- Before
sending an information product outside the originating office for review,
the product manager should:
- Verify
compliance with all applicable BTS standards and guidelines (BTS 2002,
2003, 2004, 2005),
- Double-check
facts,
- Proofread
text, and
- Clearly
mark the product as a draft for review only, and not for attribution or
further distribution.
- All
information products require a confidentiality protection review (BTS
2004).
- Verify
calculations through an independent recalculation of a random selection of
statistics in the information product.
- Persons
not directly involved in preparing the information product should
proofread the text and verify that numbers in tables, graphs, maps, and
text are consistent.
- All
information products require a subject-matter review by someone,
preferably from within BTS, who is familiar with the topic area and with
the techniques used. The
information product may require a separate review of the statistical
methodology.
- If
the topic may be of interest to another DOT organization, industry group,
or others in general, ask formally for review by those deemed most
interested.
- Publication
specialists should edit text products to ensure consistency and
readability.
- The
appropriate office director should review and clear all information
products before submitting the products to the Director or the Director’s
designee. The Director or designee will
review the product and determine whether the product needs further review
within the Research and Innovative Technology Administration (RITA) prior
to final dissemination approval.
- Information
products to be posted on the web require review for compliance with BTS
web guidelines. Each information
product should be assigned to at least one of the program web pages.
Guideline 6.1.4: External Peer Review
If an external peer review process is
used:
- Select
peer reviewers primarily on the basis of necessary technical expertise,
- Any
non-government peer reviewers paid by BTS must disclose to DOT any prior
technical/policy positions they may have taken on the issues at hand and
their sources of personal and institutional funding (private or public),
- Conduct
peer reviews in an open and rigorous manner, and
- Consider
all relevant technical comments, although outside reviews are not binding
on BTS.
Guideline 6.1.5: Contact Information
All information products must include a contact
reference to BTS customer service. As
part of the dissemination process, inform the information service of new
products and provide background information so that the information service
staff can appropriately respond to, or forward, inquiries regarding the
information product and its data sources.
Related Information
Bureau of Transportation Statistics (BTS). 2002. Section 508 Compliance Plan, Version 1.0. Washington, DC.
__________. 2003. BTS Guide to Style and Publishing Procedures. Washington, DC.
__________. 2004. Confidentiality
Procedures Manual. Washington,
DC.
__________. 2005. BTS
Statistical Standards Manual. Washington,
DC.
Office of Management and Budget (OMB). 2000. Electronic
and Information Technology Accessibility Standards, Final Rule. Federal Register, Vol. 65, No. 246, pp.
80500-80528. Washington, DC. December 21.
__________.
2002. Guidelines for Ensuring and Maximizing the Quality, Objectivity,
Utility, and Integrity of Information Disseminated by Federal Agencies. Federal Register, Vol. 67, No. 36, pp. 8452-8460. Washington, DC. February 22.
__________. 2005. Final Information Quality Bulletin for Peer
Review, Final Bulletin. Federal
Register, Vol. 70, No. 10, pp. 2664-2677.
Washington, DC. January 14.
__________.
2005. Standards for Statistical Surveys (Proposed), Sections 6.1 (Review
of Information Products) and 7.1 (Missing Data). Washington, DC. July 14.
U.S. Department of Transportation (DOT). 2002. The Department of Transportation Information Dissemination Quality Guidelines. Washington, DC. Available at http://dms.dot.gov/ombfinal092502.pdf as of January 19, 2005.
Approval Date: May 4, 2005
6.2 Tables, Graphs, and Maps
Standard 6.2: Tables, graphs, and maps in BTS information
products must accurately and effectively convey the information intended.
Key Terms: external source,
statistical map, weighted average
Guideline 6.2.1: Identifying Content and Sources
As far as possible, tables, graphs, and maps
should be interpretable as stand-alone products.
- Titles
for tables, graphs, and maps should be clearly worded and identify the
content. Include the timeframe and
geographical limitations. The axis
units may serve to identify the timeframe for graphs.
- All
tables, graphs, and maps must have a complete source note (BTS
2003b). Include information not
immediately evident from the main body of the presentation, such as
definition of codes, acronyms and special terms, and anything else that
would not be obvious to the general reader.
- Source
references should be sufficiently detailed for a reader to identify the
data used. Source notes in all
products must give a full citation for the actual source from which the
data were taken, even if that source merely collected data from other
sources.
- Since
databases and documents may be updated, the “as of” date for the source
should also be noted. Web links
should include the URL and date accessed.
Even a report featuring results entirely from one source should
have the source note with each table, graph, or map, in case they are
separated from the report.
- When
presenting estimates that are calculated using data from external sources, note each source and add a statement
describing how the calculation was done.
If the calculation is complex (e.g., a weighted average constructed
from raw data and weights), include a description of the methods used or a
reference to where they are described.
Cite BTS as the source of the calculations, based on the external
source.
- Use footnotes
to clarify data illustrations, tables, graphs, and maps regarding
particular points, abbreviation symbols, and general notes.
Guideline 6.2.2: Consistency of Presentation
To
facilitate comparability, be consistent in constructing tables, graphs, and
maps within an information product that cover similar material and use similar
units.
- Tables,
graphs, and maps within the same information product should use similar
fonts, units, spacing, and line thicknesses. Symbols and codes should also be similar
throughout an information product.
- For
comparability across BTS products, tables and graphs must comply with BTS
style and formatting guidelines (BTS nd, 2003a, 2003b).
Guideline 6.2.3: Tables
Each cell in a table must have a number, a zero indicator, or a symbol
indicating the reason that data are not displayed. Numbers in tables must comply with the
following criteria:
- All
values in a vertical list of numbers must have the same number of decimal
places. Use no decimal if all of
the values in a vertical list are integers.
- Use
no greater precision than is warranted by the data (see section 6.6).
- Only
display zeros for values that are true zeros. If a value rounds to zero, use alternate
symbols (BTS 2003b), such as “--,” to indicate that the estimate rounds to
zero in the units being presented.
- For sample-based
zero estimates, use alternate symbols to indicate that the estimates are
negligible, but possibly non-zero, in the population.
- All
tables that should logically sum to either 100 percent or some other numeric
total must provide a note if the summation is affected by independent rounding
or missing data.
Guideline 6.2.4: Graphs
Design graphs to maximize clarity and comparability within the
information product and with other BTS products.
- Design color graphs to show sufficient contrast if printed in
black and white or viewed by a colorblind user. Web graphs need appropriate alternative text for use by screen readers.
- Graph titles and axis labels should be clear with no
unexplained or undefined acronyms or industry jargon. In graphs
with axes, indicate well-defined variable names and units for each axis.
Both axes of a graph should be
labeled with the names of variables, except where the axis label “years”
is unnecessary because the years are shown.
- Graphs
that users are likely to compare should have similar scaling to facilitate
the comparison.
- Gridlines
can be helpful to users if kept inconspicuous.
- Minimize
non-data clutter.
- Minimize
use of stacked bar or line graphs.
They tend to present minimal information and are usually harder to
interpret than simple tables or line graphs.
- Do
not use 3D graphs to present two-dimensional data
- When using time intervals, spacing should be equidistant only
if the intervals are equidistant.
- In
graphs, a vertical numerical axis should normally include zero or a break
indicator (two slashes). If adding
such a break is not reasonable due to software restrictions, add a note
that the vertical axis is not zero-based.
- For graphs showing relative quantities such as an index, zero
is not a meaningful reference point.
In such graphs, use the natural basis (such as 100) as a reference
line in the graph.
Guideline 6.2.5: Statistical Maps
Statistical maps must comply with graph
standards where applicable.
- Use
shadings for a statistical map that can be easily distinguished, even if
reproduced in black and white.
- Design
category intervals to minimize the differences within classes and maximize
the differences between classes. Limit
the number of intervals to show better contrasts of shades. Three to five intervals should suffice.
- Take
care that the statistical map displays the data intended.
- Use
a scale consistent with the statistical information displayed.
- Note
that statistical area maps tend to emphasize geographic area versus other
factors.
- Provide
an accuracy statement when appropriate.
Note when data displayed in statistical maps have been collected
locally and reflect varying methods of data collection.
- Provide
a distance scale and a legend that defines symbols and other graphic
devices used in the map.
- Map
symbols and categories should be consistent throughout an information
product and throughout series of related maps.
Related Information
Bureau of Transportation Statistics (BTS). nd. BTS Web Software Guidelines. Washington, DC.
__________. 2002. Section 508 Compliance Plan, Version 1.0. Washington, DC
__________. 2003a. BTS Excel Table Standards. Washington, DC.
__________. 2003b. BTS Guide to Style and Publishing Procedures. Washington, DC.
Energy Information Administration (EIA). 1998. EIA Guidelines for Statistical Graphs. Washington, DC. Available at http://www.eia.doe.gov/neic/graphs/preface.htm as of April 19, 2005.
U.S. Department of Transportation (DOT). 2002. The Department of Transportation Information Dissemination Quality Guidelines, Section 5.1 (Publications and Disseminated Summaries of Data). Washington, DC. Available at http://dms.dot.gov/ombfinal092502.pdf as of January 19, 2005.
Approval Date: May 4, 2005
6.3 Text Discussion
Standard 6.3: Information should be presented clearly and
objectively to the public, including a full disclosure of source(s).
Key Terms: confidential,
objectivity, reliability, significant, time series, variance
Guideline 6.3.1: Presentation
Documents should be well organized with language that clearly
conveys the message to the intended audience. Text discussion in information products must
be consistent with accompanying tables, graphs, and maps, whether they are
adjacent to the text or in other areas of the product. Include tables with text wherever possible.
Guideline 6.3.2: Sources
Data presented in the text
that do not refer directly to the tables, graphs, or maps in the text
must have a source reference (see Section 6.2.1).
- Information
used in BTS information products should come from known reliable sources.
- Sources
for which methodological information is unavailable (such as proprietary
data) must include advisories indicating the lack of source and accuracy
information.
Guideline 6.3.3: Data Discussions
Discussions of data should be objective
and make statistically appropriate statements.
- Fundamental
changes within time series should be fully discussed. These changes may include, but are not
limited to, changes to how the data were collected, changes in
definitions, changes to the population, or changes in processing methods.
- Statistical
interpretations should indicate the amount of uncertainty. Only discuss differences or changes if the
appropriate statistical tests verify their statistical significance. Terms such as “confidential,” “reliability,”
“significant,” and “variance” should only be used in the statistical
sense.
- Avoid
statements that imply a specific cause and effect relationship where one
has not been established. Speculative
statements about possible causes are acceptable if worded as speculation
and not fact, and if supported by legitimate research citations.
- No
policy recommendations may be made regarding solutions to problems except
with regard to data requirements.
Related Information
Bureau of Transportation Statistics (BTS). 2005. BTS Statistical Standards Manual,
Chapter 5 (Data Analysis). Washington,
DC.
Office of Management and Budget (OMB). 2002. Guidelines for Ensuring and Maximizing the
Quality, Objectivity, Utility, and Integrity of Information Disseminated by
Federal Agencies, Public Comments and OMB Response (Applicability of
Guidelines). Federal Register, Vol. 67,
No. 36, pp. 8453-8454. Washington,
DC.
February 22.
Plain Language Action & Information Network. nd. Writing User-Friendly Documents. Available at http://www.plainlanguage.gov/handbook/index.htm as of February 9, 2005.
U.S. Department of Transportation (DOT). nd.
Plain Language Resource Page. Available at http://www.dot.gov/ost/ogc/plain.htm as of February 9, 2005.
__________. 2002. The Department of Transportation Information Dissemination Quality Guidelines. Washington, DC. Available at http://dms.dot.gov/ombfinal092502.pdf as of January 19, 2005.
Approval Date: May 4, 2005
6.4 Micro-Data Releases
Standard 6.4: Where confidentiality protections permit
their release, release micro data (unit-level data) in a manner that
facilitates its usefulness to the public.
Documentation must be readily accessible to customers, provide the
metadata necessary for users to access and manipulate the data, and clearly
describe how the information is constructed.
Key Terms: metadata, micro
data, record layout, standard error, variance, weight
Guideline 6.4.1: Software Accessibility
If micro data are released as an information product,
all micro-data products and documentation should be made accessible without
requiring the use of any one particular commercial product. Open source formats (ASCII text, space
delimited, comma delimited, etc.) must be made available in addition to any
others.
Guideline 6.4.2: File Description
Provide complete documentation for all
data files.
- Data
producers should determine what metadata standards are current at the time
data files are prepared and produce associated metadata for their files
that comply with applicable standards.
- Documentation
must include a description of the data files including the title, data
collection sources, tables that make up the set, inter-relation among
tables (e.g., keys), and record layouts for data files.
- Documentation must also include
descriptions for each variable in the data set that includes the variable
name, description, type (categorical, numerical, date/time, etc.), format,
entry restrictions (e.g., categories, range), and missing value
codes.
- Indicate changes made to previously released
data and the “as of” date of the data file.
Guideline 6.4.3: Information Quality Discussion
Micro-data files must include a
discussion of how the data were collected and the limitations of the data (see
Section 6.8).
Guideline 6.4.4: Items Needed for Variance Estimation
Datasets containing sample data must
contain appropriate weights and associated variables for accurate variance
estimation. A dataset that requires
weights and additional variables for the computation of estimates and standard
errors should not be released before these items become available.
Related Information
American Association for Public Opinion
Research (AAPOR). nd. Best Practices for Survey and Public Opinion
Research. Available
at http://www.aapor.org/default.asp?page=survey_methods/standards_and_best_practices/best_practices_for_survey_and_public_opinion_research#best12
as of April
29, 2005.
International Organization for Standardization.
2002-2003. ISO/IEC 11179, Information Technology --
Metadata Registries (MDR), (multipart standard). Available at http://metadata-standards.org/11179/ as of January
25, 2005.
Office of Management and Budget (OMB). 2005. Standards for Statistical Surveys (Proposed),
Section 7.4 (Documentation and Release of Public Use Micro Data). Washington, DC. July 14.
Approval Date: May 4, 2005
6.5 Data Protection Prior to
Dissemination
Standard 6.5: All information products must be released in
accordance with applicable Federal law and regulations in conjunction with any
confidentiality pledges given to data providers.
Key Terms: confidentiality,
disclosure limitation
Guideline 6.5.1: Non-disclosure of Confidential Data
For information collected under a
confidentiality pledge, employ statistical disclosure limitation procedures and
methods to protect any identifiable or other confidential data from disclosure
prior to public dissemination. BTS staff
must follow the established confidentiality procedures outlined in BTS’ Confidentiality Procedures Manual
(2004).
Guideline 6.5.2: Security of Disclosure Limitation Methods
The BTS confidentiality officer must
review and approve any descriptions of disclosure limitation methods prior to
their public dissemination.
- Do
not publish the details about how disclosure limitation methods were used
to protect the data, if publication could jeopardize data confidentiality. For example, do not reveal information
on how noise may have been added to the data, what variables were used to
implement record swapping, or the parameter
values used to protect tabular data.
Guideline 6.5.3: Disclosure Review Requirements
All information products must be reviewed for compliance with the disclosure protection procedures stated under the section, “Disclosure Review Board,” in the BTS Confidentiality Procedures Manual (2004).
Related Information
Bureau of Transportation Statistics (BTS). 2004. Confidentiality Procedures Manual. Washington, DC.
Bureau of Transportation Statistics Confidentiality Statute,
49 U.S.C. 111(k) as amended by the Safe, Accountable, Flexible, Efficient Transportation Equity Act: A Legacy for Users. P.L. 109-59.
Confidential Information Protection and Statistical Efficiency Act (CIPSEA) of 2002. P.L. 107-347, Title V.
Privacy Act of 1974, as Amended, 5 U.S.C. 552a.
Approval Date: May 4, 2005
6.6 Rounding
Standard 6.6: Use consistent practices for rounding and
displaying numbers in text, tables, and figures.
Key Terms: precision, significant
digit
Guideline 6.6.1: Using Rounded Numbers
All calculations should be made before
rounding. In particular, tabulations to
produce summary data and computations performed for purposes of estimating
standard errors should be done on data as collected. No rounding should take
place before completing these kinds of tabulations.
- The
sum of the rounded numbers may not equal the rounded sum. In such a case, add a note indicating that
totals may not equal the sum of their individual components due to independent
rounding.
- To
allow users to make further calculations accurately, do not further round
estimates disseminated in a spreadsheet.
Guideline 6.6.2: Degree of Rounding in Text and Graphs
The degree of rounding for text
discussion and graphs should depend on the type of data (actual measure vs.
sample), the known or suspected accuracy of the data, and the differences being
discussed.
- Round
percentages appearing in text to whole numbers unless smaller differences
being discussed require decimal places and the accuracy supports it.
- Perform
rounding consistently for similar subjects throughout the information
product.
- In
multiplying or dividing numbers, the resulting precision cannot be more
precise than the least precise of the component numbers.
Guideline 6.6.3: General Rounding Rule
Consistent with BTS standard software
(BTS nd), the general rules for rounding are:
- If
the first digit to be dropped is less than 5, then do not change the last
retained digit (e.g., round 6.1273 to 6.127).
- If
the first digit to be dropped is 5 or greater, then increase the last
retained digit by 1 (e.g., round 6.6888 to 6.69).
Related Information
Bureau of Transportation Statistics (BTS). nd. BTS Web Software Guidelines. Washington, DC.
Energy Information Administration (EIA). 2002. EIA Standards Manual, Standard 2002-15
(Rounding) and Standard 2002-15 Supplementary Materials (Guidelines on the
Standard for Rounding). Washington,
DC. Available at http://www.eia.doe.gov/smg/Standard.pdf as of January 25, 2005.
National Center for Educational Statistics (NCES). 2002. NCES Statistical Standards, Standard 5.3
(Rounding). Washington,
DC.
Available at http://nces.ed.gov/statprog/2002/std5_3.asp as of January
25, 2005.
Approval Date: May 4, 2005
6.7 Information Revisions
Standard 6.7: A standard process for handling possible post-dissemination
data changes should exist and be documented.
Key Terms: external source, revision
Guideline 6.7.1: Scheduled Revisions
When
appropriate, establish a schedule for anticipated revisions and make it
available to users.
- Identify the first dissemination of a data value in an
information product as "preliminary" if revisions are
anticipated in a subsequent dissemination.
- Designate scheduled revisions to data values as
"revised" (or “final”) when disseminating the changes.
Guideline 6.7.2: Errors in Previously Disseminated Information
Actions taken when data errors are
discovered, or an external data source makes changes, are dependent on the
impact that the potential revision would have on previously disseminated estimates.
- Establish threshold criteria for making revisions. For example, the threshold criteria
might be to revise for changes exceeding five percent in smaller values or
exceeding one percent in larger values.
- If the change does not exceed the threshold criteria, or
threshold criteria do not exist, then management will determine whether
the error is serious enough to warrant a revision.
- Document the error discovery and correction process.
Guideline 6.7.3: Documentation of Error Corrections
Document the nature of the changes, any corrective
action needed to fix an error, and provide this information to data users.
- Identify data values changed due to unscheduled revisions and
explain the reasons for these changes to data users.
- Document problems regardless of the scope of the error or the
decision whether or not to revise the data.
- Provide
error documentation to data users.
Ensure timely and wide dissemination of information product
revisions.
Guideline 6.7.4: Monitoring Revisions to Disseminated Data
Track the differences
between an initial release of estimates and the corresponding final
disseminated estimates for key data series.
- Examine the effect of revisions (number of times data are
revised and the magnitude of the revisions). Revision error information can help users
better understand the variability between initial estimates and final
estimates. For data systems that
are continuously updated, compare the initial estimates with estimates
obtained after a suitable period has elapsed.
- Some ways to present revision error information include the
average revision error, the maximum revision error, or the distribution
of revision errors during a specified time period.
- If revision error for a key data series shows an initial
release is an unreliable indicator of the final estimate, consider whether
publishing the estimate with a measure of revision error or withholding
the initial estimate is the best way to serve data users.
Related Information
U.S. Department of Transportation (DOT). 2002. The Department of Transportation Information Dissemination Quality Guidelines, Section 6.4 (Data Error Correction). Washington, DC. Available at http://dms.dot.gov/ombfinal092502.pdf as of January 19, 2005.
Approval Date: May 4, 2005
6.8 Public Documentation
Standard 6.8: Documentation for the public must include the
materials and tools (if applicable) necessary to properly interpret and
evaluate disseminated information.
Key Terms: archive, reproducibility,
transparency
Guideline 6.8.1: Source and Accuracy Information
Source and accuracy information should
provide summary information suitable for posting on the web, and should be regularly
updated to include methodological changes and the results of any quality
assessment studies. Source and accuracy
statements should summarize:
- Data
system objectives and frequency of information release,
- Target
population and coverage, geographic or other characteristic distribution and,
where applicable, sample selection methodology and sample size,
- Data
collection methodology and content of forms,
- Data
adjustments for missing data, nonresponse, coverage error, measurement
error, seasonality, and (if applicable) confidentiality protection,
- Estimation
methodology, including variance estimation methodologies for statistical
samples,
- Description
of major sources of error, including coverage of the target population,
missing data effects, and measurement error, and
- A
BTS point of contact for further questions and comments.
Guideline 6.8.2: Availability of Additional Documentation
To ensure the transparency of BTS
information products, additional documentation (as specified in Chapter 2, Section
3.3, Section 4.6, Chapter 5, and Guideline 7.1.4) should be made available to
customers upon request, unless such release would jeopardize confidentiality or
disclose the actual methods used to protect the data.
Guideline 6.8.3: Reproducibility
Data users should be able to reproduce
any publicly released information product to a reasonable degree of closeness. Information products that have been revised
should clearly indicate the “as of” date.
Guideline 6.8.4: Archive Requirements
To ensure reproducibility within BTS, the
product manager should establish criteria for retaining and archiving:
- All
electronic product files,
- Complete
information products, whether paper or electronic, representing a specific
continuing publication product or one-time report,
- The
data files and/or databases (at the most disaggregated level), which are
used to generate publicly released information products, and
- System
and model documentation and computer software/programs used to generate any
information product.
Related Information
Bureau of Transportation Statistics (BTS). 2005. BTS Statistical Standards Manual. Washington, DC.
Office of Management and Budget (OMB). 2002. Guidelines for Ensuring and Maximizing the
Quality, Objectivity, Utility, and Integrity of Information Disseminated by
Federal Agencies. Federal Register,
Vol. 67, No. 36, pp. 8452-8460. Washington,
DC.
February 22.
U.S. Department of Transportation (DOT). 2002. The Department of Transportation Information Dissemination Quality Guidelines, Section 5.3 (Source and Accuracy Statements). Washington, DC. Available at http://dms.dot.gov/ombfinal092502.pdf as of January 19, 2005.
Approval Date: May 4, 2005
|
|