Table A2.1 The difference between two overall average MPRAB(I) values within a row (Region) in Table 4 is declared significant (based on \(\alpha =0.03\)) if it exceeds 2.58 in absolute value. The primary results of the post-hoc analysis for differences in overall average MPRAB(I) values are listed within this table

1)

In Chicago, we cannot distinguish the performance between the Methods for any Size.

2)

When the Size = 360, we cannot distinguish between the Methods in Phoenix and Pittsburgh. In Atlanta, VBEST-SYS is significantly better than Recent while in Baltimore, VBEST-SYS and Uniform are better than Recent.

3)

At Size = 540, Recent has the worst performance in Phoenix and Pittsburgh. Recent is worse than VBEST-SRS in Atlanta and VBEST-SRS and VBEST-SYS in Baltimore.

4)

With the Size, 720, Recent has significantly worse performance in Pittsburgh than VBEST-SRS and VBEST-SYS and it is worse than VBEST-SYS in Atlanta. Baltimore shows an anomaly with Recent having the lowest mean MPRAB(I), although not significantly better than VBEST-SRS and VBEST-SYS with Uniform being significantly worse than Recent.

5)

At Size = 540, the mean MPRAB(I) for Recent is not significantly better than those from any other Method using Size = 360. Similarly, Recent with Size=720 is no better than the others with Size = 540 (with the exception of the Baltimore Region).

6)

The MPRAB(I) means for a given Method with Size = 540 fall between those values of the Method using Sizes of 360 and 720 respectively. This result generally holds for each of the Methods across the Regions.