so starting new ones was quite
a challenge.”
Challenging Research Issues in Statistics To put into perspective just
The following research topics have been identified by the u.S. Census Bureau as high
how challenging these undertak-
priorities for ongoing research:
ings were, the ACS is now an
ongoing survey sent to about 3
Nonresponse: Demographic Surveys — Survey response rates have been declining, lead- million residential addresses annu-
ing to concerns about the accuracy of demographic survey estimates. Errors introduced ally that describes what the popu-
by unit nonresponse may bias survey estimates when nonresponse is high and those lation looks like and how it lives.
who participate in surveys are different from those who do not. Ultimately, it helps communities
determine where to locate services
Nonresponse: Economic Surveys — As with demographic surveys, reducing respondent
and allocate resources. “The ACS
burden is key to improving response. Ongoing research and assessments are under way,
was initially tested with a sample
using such techniques as cognitive testing, respondent debriefings, and usability testing
from four counties,” said Gordon.
for data collection instruments.
“Today, it surveys the population
throughout the U.S. and Puerto
Reporting Unit Research: Economic Surveys — A problem for economic surveys is
Rico, collecting information to
the potential mismatch between the bureau’s definition of statistical reporting units
meet the needs of federal programs
and the structural units of a company. Research is needed to evaluate the effect of
about age, race, income, commute
mismatches on published statistics.
time to work, home value, and vet-
Editing: Economic Surveys — Methods are needed to analyze the editing process and
eran status, among other topics.”
evaluate the quality of the edits and related review processes so it can improve specific
The LED provides details
edits, such as macro-editing and selective editing.
about America’s jobs, workers,
and local economies. It integrates
Measurement Error Research and Prevention — Errors of observation arising from the
existing data from state-supplied
interviewer, respondent, questionnaire, or mode of data collection are major sources of
administrative records with exist-
inaccurate and inconsistent data. The bureau needs to carry out research on the sources
ing information from censuses,
and magnitude of measurement errors, conduct developmental research and method-
surveys, and federal administrative
ological studies to support measurement error reductions through improved question-
records. “The prototype program
naire design and procedural improvements, and evaluate and further refine questionnaire
began with just one state, but it
pretesting methods.
has grown to include the District
of Columbia, Puerto Rico, the
Disclosure Avoidance Methods — The u.S. Census Bureau must protect the confidential-
U.S. Virgin Islands, and all but
ity of respondents’ information while releasing as much high-quality data as possible. To
two states,” said Gordon.
do so, the bureau needs to continue improving disclosure avoidance methods.
Today and the Future
Investigation of Links of Census Bureau Problems to Behavioral Science Research —
In 2005, Gordon moved into
understanding recent developments in the behavioral sciences—including psychol-
her current position and shifted
ogy, cognitive neuroscience, and linguistics—may help the bureau realize some of the
her focus from collecting data to
potential benefits to its work.
finding innovative ways to dis-
Research to Improve Population Estimates — Improvements in the annual measure -
seminate them, as well as directing
ment of international and internal migration will lead to improved population estimates
the development of the bureau’s
and survey controls, which can improve estimates from numerous demographic surveys,
strategic plan—which defines
especially the American Community Survey.
its mission, long-term goals, and
objectives. “It is harder now to
Research on Time Series Methods and Seasonal Adjustment — Economic surveys col-
release information about people
lect data for more than 1,400 monthly and quarterly economic data series and use time
due to confidentiality concerns.
series analysis methods for seasonal adjustment of more than 1,000 of them. However,
The information on the Internet
time series analysis and modeling techniques could also be used to assist in editing,
about specific, identified individu-
imputation, and estimation and to improve the bureau’s published measures of statisti-
als makes the disclosure avoidance
cal uncertainty.
requirements of statistical agencies
harder to meet,” said Gordon. At
Decennial Census Coverage — This research aims to reduce coverage errors and expand
first, the Census Bureau reduced
the bureau’s understanding of these errors by improving census coverage measurement,
the amount of data released, but it
looking forward to the 2020 Census.
was clear more sophisticated sta-
Detailed information can be found at
www.census.gov/srd/researchtopics.pdf.
tistical methods had to be devel-
oped so information the country
12 AMSTAT NEwS SEPTEMbER 2009
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76 |
Page 77 |
Page 78 |
Page 79 |
Page 80