Agriculture Reference
In-Depth Information
Because the types of crops and livestock can differ widely across a state, land is
divided into homogeneous strata: intensively cultivated land, urban areas, and range
land. The general strata definitions are similar for the different states. However,
minor adjustments may be made to the definitions depending on the specific
requirements of a state. Each stratum is further divided into substrata by grouping
areas that are agriculturally similar. Within each substratum, the territory is divided
into PSUs.
A PSU completely covers all agriculture activities occurring within it, and,
consequently, all farmers in the state. Each PSU is divided into segments, which
are roughly a square mile in area. A sample of PSUs is selected, and one segment is
randomly selected from each selected PSU.
Field interviewers divide all the selected segments into tracts, where each tract
represents a unique land operating arrangement. Each tract is labeled as agricultural
or non-agricultural. A tract is considered as agricultural if it has qualifying agri-
cultural activity either inside or outside the segment. Otherwise, it is labeled as
non-agricultural. An agricultural tract is classified as farm if its entire operation in
sales (or potential sales) consists of at least
$
1,000. All non-agricultural tracts, and
agricultural tracts with less than
1,000 in sales, are classified as non-farms.
Each year, approximately 3,500 segments are selected for inclusion in the
sample. A selected segment is in the sample for 5 years. Thus, each year, the
sample contains approximately 11,000 segments.
The JAS design is a probability-based sample. Each tract has a first-order
inclusion probability
$
π
i
. Within each farm tract,
a proportion of a farm is observed and defined as
t
i
(tract acres/farm acres). The
current JAS estimate for the number of farms (
NoF
) is defined (Lamas et al.
2010
)
π
i
, and an expansion weight
d
i
¼
1/
n
ij
X
l
X
s
i
X
NoF
¼
d
ijk
a
ijk
;
ð
:
Þ
2
1
i¼
1
j¼
1
k¼
1
where
i
¼
1,
...
,
l
is the stratum,
j
¼
1,
...
,
s
i
is the substratum,
s
i
is the number of
substrata in stratum
i
,
k
,
n
ij
is the segment,
n
ij
is the number of segments in
substratum
j
that are in stratum
i
,
d
ijk
is the expansion weight (see the HT estimator
¼
1,
...
x
ijk
X
t
ijkm
,
x
ijk
is the number of farm tracts in the given segment,
m¼
1
and
t
ijk
is the proportion of observed farm tracts.
During the pre-screening, the survey personnel visit each newly defined tract to
determine if it is a farm. In June, those tracts that have been determined to be a farm
during pre-screening (approximately 35,000) are revisited, and crop and livestock
information is collected through personal interviews.
This knowledge is subsequently used to provide state, regional, and national
estimates for crop acreages, livestock inventories, and other agricultural items.
Naturally, the procedures used to develop and sample area frames affect the
precision and accuracy of the survey statistics.
Search WWH ::
Custom Search