t,s--0
I
,""
~i
,. FOR ADMINISTRATIVE USE
\~
Research and Development Branch
July 1965
United States Department of Agriculture
Statistical Reporting Service
Standards and Research Division
vIashington, D. C.
Comparison of 1964 Objective and Reported Corn Yields
From Weighed and Non-weighed Fields
This study was made to determine if an under reporting bias is present in farmer
reported yields for corn fields from which the harvested grain was weighed.
Using
1964 objective yield data, comparisons were made between the yield computed for the
ha~ested
sample units and the farmer reported yield (obtained by a post-harvest
interview) for the same field.
The hypothesis tested was that no differences exist between farmer reported yields
and objective yield estimates when the production from the sample field had been
weighed.
Since the number of sample fields actually weighed within any given State
was smaJ.l,the sample fields were pooled to a regional level.
Region 1 included
11 North central States and 12 Southern States were assigned to Region 2. There
was a total of 54 fields in Region 1 and 57 fields in Region 2, for which the farmer
indicated the field had been weighed out at an elevator.
The following table gives a summary of the comparisons made.
Yield Comparisons for Fields Weighed Out at Elevator
Number
Average
Average
Standard
Region
of
yield
yield
Difference
error
t
fields
objective
reported
of
value
samples
by farmer
difference
(Xl)
(~)
(Xl - ~)
S(xl - i2)
1
54
82.0
73.9
8.1
2.10
3.86**
2
57
62.9
64.9
2.0
3.09
0.65
**Highly significant.
It is apparent that a real difference exists in Region 1 between the objective
yield estimate and the farmer reported yield.
When earlier studies, which show
that the objective yield procedure has no more than a small amount of upward bias
are taken into account, it seems clear that the farmer reported yields are, based
upon weighed grain, low by about 10 percent in Region 1. It should be pointed out
that the yield computed from the objective samples assumes an 8-percent harve