### Table 14 shows the effect sizes of the three studies. It can be seen that, in fact, the 2nd replication has the lowest effect size value. The low effect size for the 2nd replication explains why the result of the statistical test turned out to be not statistically significant.

"... In PAGE 32: ... Table14 : Effect Sizes of the Cost per Defect Ratio for the Overall Inspection. For comparing and combining the effect sizes, we first checked the effect size homogeneity by calculating Q.... ..."

### Table 14 shows the effect sizes of the three studies. It can be seen that, in fact, the 2nd replication has the lowest effect size value. The low effect size for the 2nd replication explains why the result of the statistical test turned out to be not statistically significant.

1999

"... In PAGE 32: ... Table14 : Effect Sizes of the Cost per Defect Ratio for the Overall Inspection. For comparing and combining the effect sizes, we first checked the effect size homogeneity by calculating Q.... ..."

### Table 2: Simulation Results: low correlation case (C =0.1) forboththeDSOEandSSOEdgp; T =100, 500; No. of replications =1000

2007

"... In PAGE 16: ... In the averaging of the RBIAS results across parameters, the absolute value of the bias figures is taken before averaging, in order to avoid the cancellation of negative and positive biases. The results in Table2 indicate that for the low correlation setting (C =0.1)andfor T =100, the parameters of the SSOE model are estimated with much larger relative bias than those of the DSOE model, but with similar precision, as measured by MRAD.... ..."

Cited by 1

### Table 1: The gene expression states of 4 genes in 24 (6 mice, 4 replicates) assays, with flve possible levels: Very High (VH), High (H), Very Low (VL), Low (L) or Normal (N)

"... In PAGE 9: ...4716 G-protein coupled receptor protein sig- naling pathway 0.3 Table1... ..."

### Table 4: Component placement strategies (on a 12-node cluster) used for measurements in Section 5.1. The all replication strategy is not shown.

"... In PAGE 11: ... We call this strategy low replication. Table4 lists the three placement strategies except all replication. Figure 17 illustrates the measured system throughput under the above four placement strategies.... ..."

### Table 4: Component placement strategies (on a 12-node cluster) used for measurements in Section 5.1. The all replication strategy is not shown.

"... In PAGE 11: ... We call this strategy low replication. Table4 lists the three placement strategies except all replication. Figure 17 illustrates the measured system throughput under the above four placement strategies.... ..."

### Table 3: QMC results: Simulated prices of the Asian basket options using randomized low discrepancy se- quences. The values in parentheses denote the standard errors based on 10 independent replications.

"... In PAGE 8: ...619(0.035) [50] In Table3 , the same set of examples and the same tech- niques are compared. The only difference is that the input is drawn from the randomized Sobol0 low discrepancy sequences proposed by Owen (1995), instead of a ran- dom sequence.... ..."

### Table 1. Resource utilization. L is a low end, H a high-end machine (see text for description). Percentages are for the system. M/S is Master/Slave replicated database

2006

Cited by 5

### Table 1. Resource utilization. L is a low end, H a high-end machine (see text for description). Percentages are for the system. M/S is Master/Slave replicated database

2006

Cited by 5