Www.ams.sunysb.edu



Summary of the power/sample size calculationsThis summary includes online software for power/sample size calculations. All the links of the software mentioned in this summary are listed at the end. First of all, several power/sample size calculators are introduced by the analysis type.1. Inference on proportion Russ lenth’s power and sample size calculation1 can determine the sample size by Confidence Interval (CI) for one proportion, also do the power analysis based on the test of one proportion, or test comparing two proportions.2 can determine the sample size based on the specified CI and power.MACORR5 can only calculate the sample size based on the CI for one proportion.2. Inference on population mean(s) Russ lenth’s power and sample size calculation1 can do the calculation for one-sample t test, two-sample t test. It also has a “balanced ANOVA” selection for a list of several popular experimental designs, include one-way, two-way, three-way ANOVA for multiple meansPS3 can do the calculation for paired and independent two-group t test.2 can only calculates the sample size for testing one population mean.SOCR4 only can do power analysis for one population mean which follows the normal distribution.3. Inference on variance(s) Russ lenth’s power and sample size calculation1 can do the calculation for F-test comparing two variances. 4. Inference related to linear regressionRuss lenth’s power and sample size calculation1 can calculate the power of testing one predictor in the presence of other predictors.PS3 provides the power analysis for testing the slop of a simple linear regression or comparing the slopes and intercepts of two independent linear regressions.5. Inference related to some specific testsRuss lenth’s power and sample size calculation1 can calculate the power for Chi-square test and poisson test.PS3 can calculate the power for Mantel-Haenszel test, also independent or matched chi-squared test and fisher’ exact test in case-control studies.6. Inference related to survival studiesPS3 provides the calculation based on the log-rank test, the alternative hypotheses are specified in terms of the hazard ratio for control subjects relative to experimental subjects or the median survival times for control and experimental subjects.Another free one called QUANTO6, which is also the power/sample size calculator but different with all above, is specifically applied in the field of association studies of genes, gene-environment interaction, or gene-gene interaction.And some paid software, NCSS, PASS, GESS7 and Studysize 2.08 are also included in the GCRC website. I can’t access these software to see how they works. Just from the online manuals, they all very powerful and not only do the power/sample size calculation also almost all kinds of common statistical analysis.Software link list1. Russ lenth’s power and sample size calculation (use true different, no effect size)2. of statistics software offered, all based on R modules, but with very little note to explain what it is and how it works3. PS . SOCR based software with lots of other functions. But it is not stable, the program get no response quit often. 5. MACORR . QUANTO . NCSS, PASS, GESS . Studysize2.0 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download