indicating their reliability for use in an Analysis of Variance
(ANOVA) among farm scales.
Each exploratory analysis illuminated commonalities among
variables within each variable category with eigenvalues greater
than one. The eFAs used were maximum likelihood using varimax
rotation and standard Pearson's r correlations. Factors were created
for each variable category in Table 2A and factor labels were
assigned based on the commonalities we found among the variables
contributing to each factor and our substantive knowledge of
produce safety and this Amish community. For example, Factor 1
was labeled “Field Risks Sources” because of the affinity among
variables relating to the spatial location where the presence of a
phenomena or activity characterized by a variable would be found
or took place (Pesticide Application, Plant Disease, Weed Infestation,
Insect Infestation etc.).
In the eFA of Pre- and Post-harvest contamination sources eight
factors with eigenvalues greater than one were created that explain
a cumulative total of 70% of the variance. Two of factors, Field Risk
Sources and Worker Sources, explain 33% and 10% of the variance,
respectively. Of the six contamination source variables correlated
with farm size in the previous section, two appear in the Field Risk
factor, and one in the Equipment Sources factor. These are identified
by bold italics in Table 4.