**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

**Suggested Citation:**"CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION." National Academies of Sciences, Engineering, and Medicine. 2009.

*Precision Estimates of AASHTO T 242*. Washington, DC: The National Academies Press. doi: 10.17226/23017.

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

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 4 CHAPTER 2- RESULTS OF ANALYSIS AND ESTIMATES OF PRECISION This chapter provides statistics of the friction measurements of TTI and TRC friction test centers. Also provided in this chapter are the precision estimates for AASHTO T 242 computed based on the statistics of the TTI Final and TRC Departure State System data. 2.1 METHOD OF ANALYSIS The analysis of friction data in this study was based on ASTM E 691 test method. Prior to the analysis, any partial sets of data were eliminated by following the procedures described in ASTM E 691 in determining repeatability (Sr) and reproducibility (SR) estimates of precision [9]. Data exceeding critical h and k values were eliminated as described in Section 15.6 of the test method. Once identified for elimination, the same data were eliminated from any smaller subsets analyzed. Test data from this study were displayed graphically using box plots. The box plot is a graphical data analysis technique for determining if differences exist between various levels of a 1-factor model. The box plot is in fact a graphical alternative to a 1-factor ANOVA. It is also a useful technique for summarizing and comparing data from two or more samples. A box plot is structured in the following manner. The bottom x is the data minimum and the top x is the data maximum. The bottom of the box is the estimated 25 percent point and the top of the box is estimated 75 percent point. The middle x in the box is the data median. 2.2 ANALYSIS OF TTI FRICTION DATA 2.2.1 Analysis of Initial State System Measurements The Initial State System measurements were made upon arrival of the systems to the test center prior to any adjustments. Data obtained using the initial state are provided in Appendix A and shown on box plots in Figure 2-1 with the state systems identified alphabetically from A to M. The data were collected from 12 replicate runs of eight state systems operated at 40 mph (64.4 km/h) on three different surfaces (pads). The h- and k- statistics for the Initial State Systems are provided in Table A-1 of Appendix A and displayed in Figure 2-1. As indicted in the table and figure, the k-statistics of Pad 1 using State Systems A, the k-statistics of Pad 2 using State System F, and the k-statistic of Pad 3 using State System B exceeded the critical k-value and were eliminated from the analysis. The remaining data were re-analyzed according to E 691 method to determine the Sr and SR standard deviations shown in Table 2-1 and Table 2-2.

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 5 2.2.2 Analysis of Final State System Measurements The Final State System measurements were made after the adjustments were applied to the state systems to put them into compliance with ASTM E 274. The Final State Systems were operated on Pad 1, Pad 2, and Pad 3 at 30, 40, and 50 mph (48.3, 64.4, 80.5 km/h). The Final State Systems data are provided in Appendix B and are identified alphabetically from B through N. The data obtained using the Final State Systems at 30, 40, and 50 mph (48.3, 64.4, 80.5 km/h) are shown on box plots in Figure 2-2, Figure 2-3, and Figure 2-4. Data were collected from 12 replicate runs on the three surfaces using 12 state systems. The h- and k- statistics of the Final State Systems at the three speeds and on the three pads are provided in Tables B-1 through B-3 of Appendix B and are also displayed in Figure 2-2, Figure 2-3, and Figure 2-4. At 30 mph, the k-statistic of Pad 1 and Pad 2 using System B and the h- and k-statistics of Pad 3 using System H exceeded the critical h- and k- values and were eliminated from the analysis (Table B-1 and Figure 2-2). At 40 mph, the k- statistic of Pad 2 using System H and k-statistic of Pad 3 using System E exceeded the critical k value and were eliminated from the analysis (Table B-2 and Figure 2-3). At 50 mph, the k-statistic from Pad 1 using System B exceeded the critical k value and was eliminated from the analysis (Table B-3 and Figure 2-4). The remaining data were re- analyzed according to E 691 method to determine the Sr and SR standard deviations shown in Table 2-1 and Table 2-2. Table 2-1- Number of data sets, averages, and standard deviations for friction numbers using different TTI measurement systems on three surfaces (Pads) at three speeds Table 2-2- Coefficients of variations for friction numbers using TTI measurement systems on three surfaces (Pads) at three speeds

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 6 2.2.3 Comparison of TTI Initial and Final Precisions The comparison of the initial and final statistics would indicate if the calibration process improves the precision of the frictional measurements. The statistics in Table 2-1 and Table 2-2 of initial and final measurements at 40 mph shows that 5 out of 6 repeatability and reproducibility standard deviations or coefficient of variations have improved upon calibration of the systems. This indicates that the calibration process in addition to providing adjustment to the friction measurements would improve the precision of the collected friction data. 2.2.4 Selection of Form of Precision Estimates A review of statistics of Final State System in Table 2-1 indicates that there are no significant correlations between averages and standard deviations. Therefore, standard deviations will be statistically tested if they can be combined for the precision estimate development. Statistical F-test was performed to examine the significance of the differences in variances of friction measurements from different surfaces and different speeds as explained in the following sections. 2.2.4.1 Test of Significance on Standard Deviations from Different Surfaces Statistical F-test was performed to examine the significance of the differences in variances of friction measurements from different surfaces. The results of the F-test are provided in Table 2-3. The comparison of the computed and critical F values in the table indicates that for 1% level of significance the repeatability and reproducibility values of the three surfaces are not significantly different. Therefore, the standard deviations in Table 2-1 for the three surfaces were combined as reported in Table 2-4.

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 7 Table 2-3- Computed F- and critical F- values for comparison of precision estimates for different surfaces of TTI measurements Table 2-4- Combined standard deviations for friction numbers of different surfaces (Pads) at TTI 2.2.4.2 Test of Significance on Standard Deviations from Different Speeds The evaluation of differences in variability of TTI friction measurements at different speeds would determine if statistics from different speeds could be combined. An F-test was performed to examine the significance of the differences. The results of the F-test are provided in Table 2-5. The comparison of the computed and critical F values in the table indicates that for 1% level of significance the repeatability and reproducibility of the friction numbers are not significantly different when measured at speeds of 30, 40, or 50 mph (48.2, 64.4, and 83.3 km/h). Therefore, the standard deviations of the three speeds were combined as reported in Table 2-6.

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 8 Table 2-5- Computed F- and critical F- values for comparison of the precision estimates for different speeds of TTI friction measurements Table 2-6- Combined repeatability and reproducibility precisions for frictional properties of TTI measurements

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 9 Figure 2-1- Box plots and h and k statistics for friction numbers of TTI Initial State System at 40 mph

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 10 Figure 2-2- Box plots and h and k statistics for friction numbers of TTI Final State System at 30 mph

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 11 Figure 2-3- Box plots and h and k statistics for friction numbers of TTI Final State System at 40 mph

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 12 Figure 2-4- Box plots and h and k statistics for friction numbers of TTI Final State System at 50 mph

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 13 2.3 ANALYSIS OF TRC FRICTION DATA 2.3.1 Analysis of Arrival State System Measurements The Arrival (ARR) State System measurements were made upon arrival of the systems to the TRC test center prior to any adjustments. The ARR data available for analysis were collected using ribbed 501 tire. The ARR Data are provided in Appendix C and shown on box plots in Figure 2-5, Figure 2-6, and Figure 2-7 with the state systems identified numerically from 1 to 12. The data were collected from 12 replicate runs on three different surfaces (pads) using 12 state systems operated at 20 mph (32.2 km/h), 40 mph (64.4 km/h), and 60 mph (96.6 km/h). The h- and k- statistics for the Arrival State Systems are provided in Tables C-1 through C-3 of Appendix C and displayed in Figure 2-5, Figure 2-6, and Figure 2-7. At 20 mph, the k-statistic of Pad 4 and Pad 6 using System 10 exceeded the critical k- values and were eliminated from the analysis (Table C-1 and Figure 2-5). At 40 mph, the h-statistic of Pad 6 using System 9 exceeded the critical h value and was eliminated from the analysis (Table C-2 and Figure 2-6). At 60 mph, the k-statistic from Pad 6 using System 1 exceeded the critical k value and was eliminated from the analysis (Table C-3 and Figure 2-7). The remaining data were re-analyzed according to E 691 method to determine the Sr and SR standard deviations shown in Table 2-7 and Table 2-8. Table 2-7- Averages and standard deviations of TRC Arrival friction measurements using 501 tire Table 2-8- Coefficient of variation of TRC Arrival friction measurements using 501 tire

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 14 Figure 2-5- Box plots and h and k statistics for friction numbers of TRC Arrival State System at 20 mph

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 15 Figure 2-6- Box plots and h and k statistics for friction numbers of TRC Arrival State System at 40 mph

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 16 Figure 2-7- Box plots and h and k statistics for friction numbers of TRC Arrival State System at 60 mph

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 17 2.3.2 Analysis of Departure State System Measurements The Departure (DEP) State System measurements were made after the adjustments were applied to the state systems to put them into compliance with ASTM E 274 (AASHTO T 242). The DEP State Systems were operated on the three surfaces of Pads 4, 5, and 6 and at three speeds of 20, 40, and 60 mph (32.2, 64.4, 96.6 km/h). There were 50 sets of data available for analysis. The data were collected from left or right wheel of the 1st, 2nd, or both friction systems of 12 states collected over the past 5 years. The DEP State System data are provided in Appendices E and F and are identified numerically from 1 to 35 for ribbed 501 tire and from 1 to 15 for smooth 524 tire. 2.3.2.1 Data from Ribbed 501 Tire The DEP State System data using 501 tire at 20, 40, and 60 mph (32.2, 64.4, and 96.6 km/h) are shown on box plots in Figure 2-8, Figure 2-9, and Figure 2-10. There were 35 sets of data collected over the past 5 years from 12 replicate runs on three different surfaces using one or more friction measuring systems of 12 states. The h- and k- statistics of the DEP State Systems are provided in Appendix D and are also displayed in Figure 2-8, Figure 2-9, and Figure 2-10. As indicated from the tables and figures, based on the exceedance of h or k statistics from critical values, 4 sets of data were eliminated from 20 mph measurements (Table D-1 and Figure 2-8), 4 sets of data were eliminated from 40 mph measurements (Table D-2 and Figure 2-9), and 6 sets of data were eliminated from 60 mph measurements (Table D-3 and Figure 2-10). The remaining data were re-analyzed according to E 691 method to determine the Sr and SR standard deviations shown in Table 2-9 and Table 2-10. Table 2-9- Statistics of TRC Departure friction measurements for 501 tire Table 2-10- Coefficient of variation of TRC Departure friction measurements for 501 tire types

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 18 Figure 2-8- Box plots and h and k statistics for friction numbers of TRC Departure State System with 501 tire at 20 mph

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 19 Figure 2-9- Box plots and h and k statistics for friction numbers of TRC Departure State System with 501 tire at 40 mph

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 20 Figure 2-10- Box plots and h and k statistics for friction numbers of TRC Departure State Systems with 501 tires at 60 mph

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 21 2.3.2.2 Data from Smooth 524 Tire The DEP State System data from 524 tire at 20, 40, and 60 mph (32.2, 64.4, and 96.6 km/h) are shown on box plots in Figure 2-11, Figure 2-12, and Figure 2-13. There are fifteen sets of data available from 524 tire for analysis of precisions of friction measurements. The data were collected over the past 5 years from 12 replicate runs on three different surfaces using one or more friction measuring systems of 12 states. The h- and k- statistics of the DEP State Systems are provided in Tables E-1 through E-3 of Appendix E and are displayed in Figure 2-11, Figure 2-12, and Figure 2-13. As indicated from the tables and figures, based on exceedance of k statistics from critical value, 2 sets of data were eliminated from 20 mph measurements (Table E-1 and Figure 2-11), 1 set of data was eliminated from 40 mph measurements (Table E-2 and Figure 2-12), and 3 sets of data were eliminated from 60 mph measurements (Table E-3 and Figure 2-13). The remaining data were re-analyzed according to E 691 method to determine the Sr and SR precision estimates shown in Table 2-11 and Table 2-12. Table 2-11- Statistics of TRC Departure friction measurements for 524 tire types Table 2-12- Coefficient of variations of TRC Departure friction measurements using 524 tire

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 22 Figure 2-11- Box plots and h and k statistics for friction numbers of TRC Departure State Systems with 524 tires at 20 mph

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 23 Figure 2-12- Box plots and h and k statistics for friction numbers of TRC Departure State Systems with 524 tires at 40 mph

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 24 Figure 2-13- Box plots and h and k statistics for friction numbers of TRC Departure State Systems with 524 tires at 60 mph

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 25 2.3.3 Comparison of TRC Arrival and Departure Precisions The comparison of the ARR and DEP statistics would indicate if the calibration process improves the precision of the frictional measurements. Table 2-13 provides the pooled standard deviations of the ARR and DEP State Systems with 501 tire. The comparison of the standard deviations in the table indicates that 4 out of 6 DEP standard deviations are smaller than those of ARR. This shows that the calibration process in addition to making adjustment to the friction measurements would improve the precision of the collected friction data. Table 2-13- Pooled repeatability and reproducibility standard deviations of 501 tire Arrival and Departure friction data 2.3.4 Selection of Form of Precision Estimates A review of statistics of TRC Departure State System in Table 2-9 and Table 2-11 indicates that there are no significant correlations between averages and standard deviations. Therefore, standard deviations will be statistically tested if they can be combined for the precision estimate development. Statistical F-test was performed to examine the significance of the differences in variances of friction measurements from different tire types, different surfaces, and different speeds as explained in the following sections. 2.3.4.1 Test of Significance on Standard Deviations of 501 and 524 Tires The results of the F-test on significance of difference between standard deviations of 501 and 524 tires are provided in Table 2-14 and Table 2-15. The comparison of the computed and critical F values in the tables indicates that for 1% level of significance the repeatability and reproducibility standard deviations of the two tire types are not significantly different. Therefore, the standard deviations in Table 2-9 and Table 2-11 for the three surfaces and the three speeds were combined as reported in Table 2-16.

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 26 Table 2-14- Computed F- and critical F- values for comparison of precision estimates of 501 and 524 tires for different speeds Table 2-15- Computed F- and critical F- values for comparison of precision estimates of 501 and 524 tires for different surfaces Table 2-16- Pooled repeatability and reproducibility standard deviations of 501 and 524 tires 2.3.4.2 Test of Significance on Standard Deviations from Different Speeds The evaluation of significance of differences in variability of TRC pooled friction measurements at 20, 40, and 60 mph (32.2, 64.4, and 96.6 km/h) in Table 2-16 would determine if statistics from different speeds could be combined. An F-test was performed to examine the significance of the differences. The results of the F-test are provided in Table 2-17. The comparison of the computed and critical F values in the table indicates that for 1% level of significance the repeatability and reproducibility of the friction numbers from pads 5 and 6 are not significantly different at 20, 40, or 60 mph. However, the repeatability of friction measurements of Pad 4 at 20 mph is significantly larger than those at 40 and 60 mph. In addition, the reproducibility of friction measurements of Pad 4

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 27 at 20 mph is significantly larger than that at 60 mph. Therefore, for Pads 5 and 6 the standard deviations of the three speeds were pooled and for Pad 4 the largest repeatability and reproducibility values would be selected for precision estimate determination as shown in Table 2-18. Table 2-17- Computed F- and critical F- values for comparison of precision estimates of different speeds Table 2-18- Combined standard deviations of the 3 speeds for TRC surfaces 2.3.4.3 Test of Significance on Standard Deviations from Different Surfaces The evaluation of differences in variability of TRC friction measurements from different surfaces would determine if those statistics can be combined. Statistical F-test was performed to examine the significance of the difference in variances of friction measurements from different surfaces in Table 2-18. The results of the F-test are provided in Table 2-19. The comparison of the computed and critical F- values in the table indicates that for 1% level of significance the repeatability and reproducibility values of Pads 4 and 5 are not significantly different but they are significantly different from those of Pad 6. Therefore, as reported in Table 2-20, the standard deviations of Pads 4 and 5 in Table 2-18 are combined and those of Pad 6 are reported separately.

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 28 Table 2-19- Computed F- and critical F- values for comparison of precision estimates of different TRC surfaces Table 2-20- Combined repeatability and reproducibility precisions for frictional properties of TRC measurements 2.4 COMPARISON OF TTI AND TRC PRECISIONS The comparison of the repeatability and reproducibility standard deviations of TTI (Table 2-6) and TRC (Table 2-20) friction measurements would indicate if the standard deviations can be combined. Statistical F-test was performed to examine the significance of the difference in variances of friction measurements from the two centers. The results of the F-test on pooled variances of TTI and pooled variances of Pad 4 and 5 of TRC are provided in Table 2-21. The comparison of the computed and critical F-values in the table indicates that for 1% level of significance the pooled repeatability of TTI measurements is significantly smaller than the pooled repeatability of Pads 4 and 5 of TRC. However, the reproducibility values are not significantly different. Since the same procedure should be followed for both repeatability and reproducibility, the larger of the standard deviations of the two centers that correspond to TRC surfaces would be used for the precision estimates development of AASHTO T 242.

NCHRP Web-Only Document 142: Precision Estimates of AASHTO T 242 29 Table 2-21- Computed F- and critical F- values for comparison of variances of TTI and TRC friction measurements 2.5 PRECISION ESTIMATES FOR AASHTO T 242 Since the variances of frictional measurements from TRC were significantly different from those of TTI test centers, the larger variances of the two sets were used to develop the precision estimates for AASHTO T 242. The pooled variances of Pads 4 and 5 of TRC were also kept separate from those of Pad 6. This pad, which is a hot mix asphalt surface, provided significantly larger variances than all other surfaces. Table 2-22 provides the proposed precision estimates for AASHTO T 242. Table 2-22- Repeatability and reproducibility precisions for frictional measurements of various pavement surfaces 2.6 COMPARISON WITH CURRENT PRECISION ESTIMATE IN AASHTO T 242 The current version of AASHTO T 242 includes only a repeatability standard deviation that can be compared with the repeatability standard deviation computed in this study. The repeatability standard deviation is reported as 2 FN (friction number unit) in AASHTO T 242-96 (2004), which is smaller than the proposed standard deviation for hot mix asphalt surface (2.87 FN) and larger than the repeatability proposed for other surfaces (1.56 FN). There is no reproducibility standard deviation reported in the current version of AASHTO T 242.