popularity of the Add-On Program likely will focus on communities with populations
helped get funding for the National in the 20,000 to 64,999 range, for which ACS data
Study. The Add-On Program is a are being reported for the first time. She also offered
win-win, giving FHWA leadership a reminder that, because they reflect data collect-
in-household travel survey data and ed from 2005 through 2007, users cannot expect
a bigger survey with more stake- the three-year estimates to reflect the effects of the
holders. Participants get quality data recent economic downturn.
that are consistent across the nation Navarro recalled some ACS basics that are impor-
and partnership status in a high-profile tant to understanding the nature of the multiyear
program. Participants also get guidance and estimates. The ACS is not longitudinal—it has an
user support. independent sample every year, and the sample is
Add-On Program rules allow each participant up based on housing units, without regard to who the
to five additional questions. (Contrino commented occupants are. Navarro explained that the one-year
that, thankfully, not all do this.) Participants can also estimates reflect data collected over 12 months, with
do their own sampling plans within the constraints all months given equal weight. The multiyear esti-
of the overall methodology. The level of involvement mates are an extension of this logic, pooling 36 or
is left up to the Add-On Program participants. 60 months of data in the same way the one-year
Data collection for the 2008 NHTS began in estimates pool 12 months. The multiyear estimates
March 2008, and about 56,000 households have are averages of 36 or 60 months of data, not aver-
been recruited since. A weighted interim data file ages of one-year estimates.
was developed for January, and a final 2008 data file However, while the one-year estimates are con-
is due this fall. trolled to the U.S. Census Bureau’s county popu-
Future design goals call for improvements for cell lation estimates for July of the relevant year, the
phones and noncontacts, better institutionalizing multiyear estimates are controlled to the average of
the program within DOT, maintaining the Add-On the population estimates for the relevant years. For
Program, serving the user community, and annually example, the three-year estimates spanning 2005–
reporting key measures. 2007 are controlled to the average of the most recent
county population estimates for 2005–2007.
Presidential Transition
Some geographic definitions change over time,
David McMillen joined the meeting to briefly
so Navarro explained that the multiyear estimates
describe his work (with Terri Ann Lowenthal)
use definitions effective at the end of the estimation
with the team guiding the new administration’s
period. For example, a five-year estimate (reflecting
transition with regard to the U.S. Census Bureau,
2005–2009) for a city that annexed several blocks in
Economics and Statistics Administration (ESA),
2007 would reflect the city including the annexed
and BEA. As McMillen described it, their func-
blocks. There is no attempt to ‘average’ the geogra-
tion is to provide the incoming administration
phy over the multiyear period.
with concise descriptions of the most important
Inflation adjustments are another element to
aspects of the agencies in question. They are to
understand, as dollar-valued items are inflation-
identify issues and describe options for addressing
adjusted (based on the national CPI) to the most
them, but they are not to make recommendations.
recent year of the estimation period. For example,
When asked if they are identifying persons who
income data from 2005–2007 are adjusted to 2007
could serve as the next U.S. Census Bureau direc-
constant dollar values for the reporting of three-year
tor, McMillen said they can describe the types
averages for the period of 2005–2007.
of persons who might make a good director and
The presence of one-, three-, and five-year esti-
identify lists of persons who fit the description,
mates for some areas will require users to determine
but they are not empowered to recommend indi-
which estimates are best suited for their purposes.
viduals or approach anyone about the position.
Navarro noted that the answer depends on the
application and involves a trade-off between reli-
First Release of Multiyear Estimates ability and currency. One-year estimates have lower
Susan Schechter, Douglas Hillmer, and Alfredo
reliability because of the smaller sample (just one
Navarro of the U.S. Census Bureau reported on
year of data), but more currency because all data
the release of the first ACS multiyear estimates.
are recent. Multiyear estimates have greater reliabil-
Schechter noted the first estimates were released
ity because of the larger sample, but less currency
December 9. This is a major milestone that provides
because some of the responses are several years old.
data for more than 13,000 areas (twice the number
Navarro also described the issue of overlapping
for one-year estimates), which include about 95% of
data in the annual series of multiyear estimates.
the U.S. population. Schechter predicted the media
Three-year estimates released this year covering
26 AMSTAT NEWS MARCH 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 |
Page 81 |
Page 82 |
Page 83 |
Page 84 |
Page 85 |
Page 86 |
Page 87 |
Page 88 |
Page 89 |
Page 90 |
Page 91 |
Page 92