University of Wisconsin Digital Collections
Link to University of Wisconsin Digital Collections
Link to University of Wisconsin Digital Collections
Ecology and Natural Resources Collection

Page View

LeMasters, Gary S.; Doyle, Douglas J. / Grade A dairy farm well water quality survey
[DNR-058] (1989)

Methods,   pp. 3-10 PDF (430.5 KB)

Page 4

A stratified random sample of Grade A dairy farms was the most efficient
methodology available. The geographical strata captured some of the variation
in soil, climate and hydrogeology in Wisconsin. It also provided a means of
allocating the sample to regions of the state having most of the Grade A dairy
There are nine Agricultural Statistics Districts in Wisconsin, providing land
area coverage of the state (Figure 1). The sample was allocated proportionally
among these districts or strata based on the number of Grade A dairy farms in
the district as a fraction of the total number of Grade A dairy farms in the
state. At the time the stratified random sample was drawn, 23,543 Grade A
dairy farms were in the sampling universe across all nine strata. It was
estimated that no more than 10% of those dairy farms chosen for sampling
would be out of business by the time water samples would be collected.
Sample Size
When estimating means, totals and proportions for a target population, the
size of a stratified random sample is dependent on several factors including the
population of each stratum, the degree of precision desired for each estimate
(limited by the most important estimate to the research), the variance of the
estimate for each stratum, and the cost of obtaining an observation in each
stratum. In this instance, the primary objective was to estimate the proportion
of wells on Grade A dairy farms with any pesticide and/or N03-N at detectable
levels. In determining sample sizes for proportion estimators it is helpful to
have some indicator of the proportion of the target population exhibiting the
trait measured.
As a pilot study, this project enjoyed few indications of the factors discussed
above, except for the population size in each stratum. The Agricultural
Resource Management (ARM) Division provided some idea of the proportion of
wells in which we might expect to find detectable levels of pesticides and NO3-
N. Staff from ARM suggested that 10% of the wells might contain pesticides.
In addition, ARM felt that they needed to have a tolerance of plus or' minus
two percent of the proportion estimator. Based on this information and funding
limitations we selected wells on 550 Grade A dairy farms, 534 of which were
actually sampled. The number of Grade A dairy farms and sampled farms are
in Table 2. Sixteen dairy farms were no longer operating when visited by the
water collector.

Go up to Top of Page