Usage Note 22224: What exact and Monte Carlo methods are available?
 |  |  |  |
Monte Carlo simulation to estimate exact p-values is useful for problems that are so large that exact computations require a great amount of time and memory, but for which asymptotic approximations might not be sufficient. Exact and Monte Carlo methods are available as shown below. See the indicated procedure documentation for details.
| Release | Procedure | Exact Methods available |
| Prior to SAS 6.11 | UNIVARIATE | Sign test; Wilcoxon signed-rank test for N≤20 |
| FREQ | Fisher's exact test for 2x2 or RxC tables | |
| PHREG | Exact conditional method for ties in continuous failure times | |
| MULTTEST | Fisher's exact test; exact permutation versions of Cochran-Armitage trend test and Peto test; limited exact CMH tests | |
| SAS 6.11 TS020 | NPAR1WAY | Two- or multi-sample exact tests based on Wilcoxon, median, Van der Waerden, and Savage scores |
| SAS 6.11 TS040 | FREQ | Exact tests for the Pearson, likelihood ratio, and Mantel-Haenszel chi-square statistics, and the Pearson and Spearman correlation coefficients; exact Jonckheere-Terpstra, Cochran-Armitage, and McNemar tests |
| SAS 6.12 (and SAS 6.09 TS450 or later on MVS, CMS, and VMS) | FREQ | Exact tests and confidence intervals for kappa, weighted kappa, and the odds ratio |
| SAS 7 TS P1 | FREQ | Exact tests for kappa, weighted kappa, binomial probability, and the odds ratio in 2x2 tables; exact confidence interval for the odds ratio and binomial probability |
| NPAR1WAY | Exact tests for scale (variability) differences; Monte Carlo estimates of exact p-values | |
| SAS 8.0 (TS0M0) | FREQ | Monte Carlo estimates of exact p-values |
| SAS 8.1 (TS1M0) | LOGISTIC | Exact logistic regression |
| SAS 9.0 (TS0M0) | GLM | Exact multivariate tests (Wilk's lambda, Hotelling-Lawley trace, Roy's greatest root) |
| NPAR1WAY | Exact two-sample Kolmogorov-Smirnov tests | |
| LOGISTIC | Stratified exact conditional logistic regression; Monte Carlo estimation | |
| POWER | Exact power computations for various tests | |
| SAS 9.1 (TS1M0) | NPAR1WAY | For two-sample problems, exact one-sided tests for the Kolmogorov-Smirnov D+ and D- statistics |
| FREQ | Exact confidence limits and test for the common odds ratio | |
| CASECONTROL ALLELE FAMILY HAPLOTYPE |
Monte Carlo estimates of exact p-values for case-control tests, disequilibrium tests, and family-based tests | |
| SAS 9.2 (TS1M0) | FREQ | Unconditional exact confidence limits for the proportion difference; Zelen's exact test for equal odds ratios |
| NPAR1WAY | Exact confidence limits for Hodges-Lehmann estimate of location shift for two-sample data; Exact tests based on Conover scores | |
| POWER | Power for exact equivalence test for binomial proportion |
Operating System and Release Information
| Product Family | Product | System | Reported Release | Fixed Release* |
| SAS System | SAS/STAT | All | n/a | |
| SAS System | Base SAS | All | n/a |
| Type: | Usage Note |
| Priority: | low |
| Topic: | Analytics ==> Categorical Data Analysis SAS Reference ==> Procedures ==> FREQ SAS Reference ==> Procedures ==> GLM SAS Reference ==> Procedures ==> LOGISTIC SAS Reference ==> Procedures ==> MULTTEST SAS Reference ==> Procedures ==> NPAR1WAY SAS Reference ==> Procedures ==> PHREG SAS Reference ==> Procedures ==> POWER SAS Reference ==> Procedures ==> UNIVARIATE Analytics ==> Multivariate Analysis Analytics ==> Nonparametric Analysis Analytics ==> Power and Sample Size Analytics ==> Survival Analysis |
| Date Modified: | 2008-07-22 14:03:49 |
| Date Created: | 2002-12-16 10:56:36 |