Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.

Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 12

12
CHAPTER 2
Results and Analysis of
Ruggedness Experiments
2.1 Analysis Approach while Table 12 and Table 13 present results for tests in
FHWA's laboratory using the IPC Global equipment. These
Linear regression is an efficient method for analyzing the tables present p-values indicating the significance of the re-
ruggedness data. For each combination of mixture, laboratory, gression coefficients for each of the factors included in the
temperature, frequency, and confinement, the ruggedness ruggedness experiment. The p-value is the probability of re-
test data can be fit to a linear model of the form: jecting the null hypothesis when it is in fact true. For this
analysis, it is the probability that the regression coefficient for a
Y = B0 + B1X1 + B2X2 + B3X3 + B4X4 + B5X5
particular ruggedness factor is zero when the analysis indicates
+ B6X6 + B7X7 + Error (1)
that it is either greater or less than zero. Thus, low p-values
where: indicate the regression coefficient is statistically significant
and the ruggedness factor affects the results of the test.
Y = measured value The key to analyzing the ruggedness test in this manner is
X1, X2, X3, X4, X5, X6, X7 = seven factors included in the selecting critical p-values above which the regression coeffi-
ruggedness testing cient is not significant, and it can be concluded that the
B0, B1, B2, B3, B4, B5, B6, B7 = model coefficients ruggedness factor does not affect the test result over the range
Error = model error tested. It is important to keep the objective of ruggedness test-
ing in mind when selecting critical p-values. The objective of
From this analysis, the statistical significance of the model ruggedness testing is to identify those controllable factors that
coefficients can be determined. For statistically significant likely affect a test, and to establish levels for their control. This
factors, the model coefficients can then be used to estimate is different from the usual objective of regression analysis,
values for each of the factors that will keep their effect below which is to develop a model to predict an outcome. A predic-
a specified level. tive model should only include variables that are highly re-
lated to the predicted outcome, so a very low p-value of 0.05
or less is normally used to detect significant variables for pre-
2.2 Dynamic Modulus Test
diction models. However, for analysis of ruggedness test data,
The results of the dynamic modulus ruggedness testing are selecting a very low p-value may result in the erroneous con-
presented in Appendix A. The dynamic modulus ruggedness clusion that one or more of the ruggedness factors does not
experiment included the factors listed in Table 3. The re- affect the test result over the range tested and controlling
sponses measured in the dynamic modulus ruggedness ex- that factor is not important. For this analysis higher critical
periment are listed in Table 9. These include the measured p-values than used in regression modeling should be selected.
dynamic modulus and phase angle, and the computed data A critical p-value of 0.10 was used. In Table 10 through
quality indicators. Table 13, factors with p-values less than or equal to 0.10 are
Regression equations of the form of Equation 1 were de- shown in bold. The analysis was not performed for the un-
veloped for each parameter listed in Table 9. The results are confined 40°C data for the dense graded mixture tested in
summarized in Table 10 through Table 13 for the dynamic the IPC device because the quality of the data was poor.
modulus and phase angle. Table 10 and Table 11 present re- The modulus measured in the equipment for this condition
sults for tests in AAT's laboratory with the ITC equipment was below the calibrated limit of the machine.