1997 Crops Day Report
INTERPRETING STATISTICAL ANALYSIS
Field research involves the testing of one or more variables such
as crop varieties, fertilizers, tillage methods, etc. Field testing of such
variables are conducted in order to determine which variety, tillage method,
or fertilizer etc. is best for the particular area of production. The main
objectives of crop production research are to determine the best means of
producing the crop and how to maximize yield and economic return from
Agricultural researchers use statistics as a tool to help differentiate
production variables so that real and meaningful conclusions can be drawn
from a relatively large amount of data.
One of these tools is the Coefficient of Variability (C.V.). This
statistic gives an indication of the amount of variation in an experimental
trial. Trials conducted at Hettinger use four replications or repetitions of
the variable in question. For example, the variety Amidon HRSW appeared four
times (four replications) in the HRSW variety trial. In this case, the C.V.
for yield of the Hettinger HRSW variety trial on fallow was 8.1%. This C.V.
is a relative measure of how much the yield of all HRSW varieties varied
between replications. In other words, C.V. is a measure of the precision or
effectiveness of the trial and procedures used in conducting the trial. More
can be said about a field trial with a relatively low C.V. (10 or less) than
one with a C.V. greater than 10. Attempts are made to control human error
and some environmental conditions such as conducting field studies on a
uniform soil so that variability between replicates is minimized with a
resulting low C.V. value (10 or less). In summation, a trial with a C.V. of
8 is more precise and more can be concluded from it than a trial with a C.V.
Another important statistical tool is the Least Significant Difference or
LSD. If the yield of variety A exceeds variety B by more than the LSD 5%
value you can conclude that under like environmental conditions, variety A
will significantly out-yield variety B 95% of the time. The LSD value allows
you to separate varieties, tillage practices, or any other variable and
determine whether or not they are actually different. The LSD 1% value is
always larger than the value for LSD 5% and is used in the same manner. If
the yield of variety A exceeds variety B by more than the LSD 1% value you
can conclude that under like environmental conditions, variety A will
significantly out-yield variety B 99% of the time. Little confidence can be
placed in variety or treatment differences unless the results differ by more
than the LSD value.