The PROC GLM (General Linear Model) in SAS was used to run ANOVA tests. Each
bid range was assigned a factor level: greater than $75,000=1, $15,000-$75,499=2, $14,999-0=3,
no bid=4. A class statement was used in PROC GLM to define the bid range variable as a
classification or categorical variable. The null hypothesis tested is that the mean value for a
variable (e.g., time spent preparing records) is the same for all bid ranges.
Table 5 shows the means for several record, farm, and farmer characteristics segregated
by the four bid categories. Results indicate that those individuals submitting the lowest bids
spend the least amount of time, on average, preparing their records. The hypothesis that the
mean time spent preparing records is the same for each of the four bid categories can be rejected
at the p=0.01 level according to an ANOVA test. Time spent analyzing records was roughly
increasing in the amount bid for financial records, but the means were not statistically different
across categories. Table 5 shows a pronounced difference in the value of farm records across
farm size. Farmers in the lowest bid category ($0 to $14,999) had gross farm sales 4.26 times
lower ($1,015,000 vs. $238,160), on average, than farmers in the highest bid category ($75,000
and greater). The hypotheses that mean gross farm sales and mean farm acreage are the same
across bid categories can be rejected at the p=0.01 level according to the ANOVA tests.
However, although older farmers tended to submit either lower or no-bids, and those submitting
bids in the highest bid category ($75,000 and greater) were the most educated on average, results
of the ANOVA tests could not reject the hypotheses that age and education were equivalent
across bid categories. The results in table 5 also provide some insight into the characteristics of
the individuals choosing not to bid in the auction. Such individuals tended to be the least
educated, the oldest, and while having relatively large farms in terms of acreage, their gross sales
15