Correction to Tables 1 and 7 of Joe, H. (1995). JASA, v. 90, pp. 957-964. (March 1998) The correction is for method P3. There was a missing factor of 2 in one line of C code that does affect the calculation especially for cases of larger correlations. The overall conclusion from the new calculations is that method P3 is more accurate than what was previously indicated. However the conclusion that method P@ is the best has not changed. The cases were run on a faster computer than 1994, with more accuracy for the 4-dimensional multivariate normal rectangle probabilities needed for method P2. There are also some changes of +/-.00001 for method P2. Table 1. Comparison of Accuracy of Approximations in the Equicorrelated Case m rho w exact M-E P_1 P_2 P_3 5 0.1 0.0 0.05286 0.05286 0.05287 0.05287 0.05287 -0.2 0.09574 0.09576 0.09588 0.09575 0.09575 -0.4 0.15877 0.15881 0.15907 0.15878 0.15878 -0.6 0.24261 0.24268 0.24302 0.24262 0.24262 -0.8 0.34393 0.34401 0.34439 0.34394 0.34394 5 0.4 0.0 0.13419 0.13542 0.13423 0.13422 0.13429 -0.2 0.19603 0.19789 0.19697 0.19606 0.19608 -0.4 0.27216 0.27457 0.27405 0.27222 0.27209 -0.6 0.36017 0.36285 0.36283 0.36024 0.35985 -0.8 0.45575 0.45826 0.45888 0.45583 0.45508 9 0.1 0.0 0.00955 0.00953 0.00957 0.00957 0.00955 -0.2 0.02365 0.02363 0.02385 0.02367 0.02366 -0.4 0.05157 0.05156 0.05210 0.05160 0.05158 -0.6 0.09979 0.09984 0.10075 0.09984 0.09980 -0.8 0.17289 0.17301 0.17424 0.17295 0.17290 9 0.4 0.0 0.06876 0.06947 0.06888 0.06888 0.06878 -0.2 0.11100 0.11274 0.11274 0.11121 0.11094 -0.4 0.16887 0.17195 0.17269 0.16923 0.16859 -0.6 0.24277 0.24719 0.24878 0.24331 0.24202 -0.8 0.33080 0.33612 0.33867 0.33153 0.32928 ======================================================= Table 7. Standard Deviations over Permutations for Some Cases case w x exact P_1 (SD) P_2 (SD) P_3 (SD) 5.2 -(.1,.2,.3,.8,.9) oo .2436 .2444(.0015) .2436 (.0002) .2437 (.0004) 5.4 -(.1,.2,.3,.8,.9) oo .3923 .3956(.0070) .3921 (.0006) .3880 (.0016)