### Table 9.10.14.11. Minimum Construction Requirements for Exposing Building Faces Forming Part of Sentence 9.10.14.11.(1)

### Table 1: Coverage accuracy of likelihood-based intervals (100,000 simulations).

"... In PAGE 12: ...Some selected results on the coverage accuracy of the likelihood-based con - dence intervals (12) are presented in Table1 , with the likelihood ratio statistic,... In PAGE 14: ... This is largely due to treating the standard error as constant in . Con dence intervals based on the Wald test use essentially the same construction but treat the standard error as a function of and, although we do not present the results here, generally perform nearly as well as the likelihood- based intervals presented in Table1 . The general conclusion from Table 2 is that straightforward application of asymptotics associated with parameter estimators is not a very reliable or accurate method for constructing con dence intervals.... In PAGE 15: ...proximation has the highest power in these comparisons. The power is related to the length of the con dence intervals presented in Table1 with higher power being associated with shorter intervals. Of course, there are other aspects of the interval methods which a ect the power, but Table 3 does provide an indirect exploration of the lengths of the intervals.... ..."

### Table 1: Coverage accuracy of likelihood-based intervals (100,000 simulations).

"... In PAGE 12: ...Some selected results on the coverage accuracy of the likelihood-based con - dence intervals (12) are presented in Table1 , with the likelihood ratio statistic,... In PAGE 14: ... This is largely due to treating the standard error as constant in . Con dence intervals based on the Wald test use essentially the same construction but treat the standard error as a function of and, although we do not present the results here, generally perform nearly as well as the likelihood- based intervals presented in Table1 . The general conclusion from Table 2 is that straightforward application of asymptotics associated with parameter estimators is not a very reliable or accurate method for constructing con dence intervals.... In PAGE 15: ...proximation has the highest power in these comparisons. The power is related to the length of the con dence intervals presented in Table1 with higher power being associated with shorter intervals. Of course, there are other aspects of the interval methods which a ect the power, but Table 3 does provide an indirect exploration of the lengths of the intervals.... ..."

### Table 2: Upper and lower non-coverage probability of central intervals (100,000 simulations).

"... In PAGE 13: ...The upper and lower tail coverage accuracy of central con dence intervals is presented in Table2 . The standard asymptotic approximation leads to the use of... In PAGE 14: ... Con dence intervals based on the Wald test use essentially the same construction but treat the standard error as a function of and, although we do not present the results here, generally perform nearly as well as the likelihood- based intervals presented in Table 1. The general conclusion from Table2 is that straightforward application of asymptotics associated with parameter estimators is not a very reliable or accurate method for constructing con dence intervals. Finally, the power of likelihood ratio and REML tests against the null hypoth- esis that = 0 using the two asymptotic approximations is explored in Table 3.... ..."

### Table 2: Upper and lower non-coverage probability of central intervals (100,000 simulations).

"... In PAGE 13: ...The upper and lower tail coverage accuracy of central con dence intervals is presented in Table2 . The standard asymptotic approximation leads to the use of... In PAGE 14: ... Con dence intervals based on the Wald test use essentially the same construction but treat the standard error as a function of and, although we do not present the results here, generally perform nearly as well as the likelihood- based intervals presented in Table 1. The general conclusion from Table2 is that straightforward application of asymptotics associated with parameter estimators is not a very reliable or accurate method for constructing con dence intervals. Finally, the power of likelihood ratio and REML tests against the null hypoth- esis that = 0 using the two asymptotic approximations is explored in Table 3.... ..."

### Table 16: A potential NDP 1997; data from Table 10 and 14: all values expressed in millions of year 2000 dollars.

### Table 5. Candidate 10/14 bit sync words.

### Table 2. Probit of employment of children 10-14 years old

2005

"... In PAGE 6: ... V. Results Table2 reports the basic estimation for the probability that a child works. In both the urban and rural cases, we see, albeit to different extents, similar relationships between the probability of a child working and the independent variables.... ..."