were evaluated for each test method: oven drying, natural
condition, and vacuum drying.
The effect of initial drying methods on test results was
analyzed in two steps. First, ANOVA and Tukey’s pairwise
comparisons among levels of material and drying method
were conducted to determine the effects of three initial drying
methods (oven drying, natural moisture condition, and
vacuum drying) on the test results (Gsa, Gsb, Gssd, and water
absorption). Tables 3-25 and 3-26 summarize the results of
the statistical analyses for AASHTO T 85 and 84, respectively.
Based on a significance level of 0.05, the effect of the three
initial drying methods was determined, as shown in the
“significance” column. Grouping of test results that were not
significantly different was conducted according to Tukey’s
test method. In the last three columns (note that Tukey’s
comparisons should be made across the rows of the table,
not down the columns), initial drying methods that did not
share a letter yielded statistically different aggregate properties.
For all materials except the gravel coarse aggregate and
natural sand fine aggregate, the results using the initial drying
methods were significantly different for at least one of the
measured aggregate properties. An analysis of the variability
of test results showed that the initial drying methods did not
significantly affect the precision of AASHTO T 85 and T 84.
Second, the research team used graphical comparisons to
illustrate the trends of the differences in the measured properties.