--- title: "Scenario Guide" output: rmarkdown::html_vignette vignette: > %\VignetteIndexEntry{Scenario Guide} %\VignetteEngine{knitr::rmarkdown} %\VignetteEncoding{UTF-8} --- ```{r setup, include = FALSE} knitr::opts_chunk$set(collapse = TRUE, comment = "#>") library(ggpower) ``` Use this guide to pick a module and test for your study design. ## Implemented scenarios (48 tests) ```{r counts, echo=FALSE} tests <- ggpower_tests() knitr::kable(as.data.frame(table(tests$module)), col.names = c("Module", "Tests")) ``` ### Power Workspace (30 tests) Classical families: t, F, chi-square, exact proportions, z correlations/GLM, Wilcoxon rank tests. ### Biomarker Discovery (7 tests) ROC/AUC, diagnostic accuracy, log-rank, Cox, FDR screening, differential expression. ### Clinical Trials (11 tests) Superiority, NI, equivalence (TOST), Simon two-stage, cluster RCT, multi-arm ANOVA, Poisson counts, survival. ## Decision table | Study question | Module | Example test | |----------------|--------|--------------| | Two-group mean difference | Power Workspace or Clinical | `t_two_sample`, `rct_superiority_continuous` | | Paired / pre-post | Power Workspace | `t_paired` | | One-way ANOVA | Power Workspace or Clinical | `f_anova_one_way`, `multi_arm_superiority` | | Multiple regression $R^2$ | Power Workspace | `f_mreg_omnibus` | | Two proportions (Fisher) | Power Workspace or Clinical | `exact_fisher`, `rct_superiority_binary` | | McNemar paired proportions | Power Workspace | `exact_mcnemar` | | Correlation difference | Power Workspace | `z_corr_independent` | | Logistic / Poisson GLM | Power Workspace or Clinical | `z_logistic`, `count_endpoint_poisson` | | Biomarker AUC | Biomarker | `roc_auc_one` | | Diagnostic sens/spec | Biomarker | `diagnostic_acc` | | Survival / HR | Biomarker or Clinical | `cox_regression`, `survival_pmu` | | Non-inferiority trial | Clinical | `rct_noninferiority_continuous` | | Bioequivalence (TOST) | Clinical | `rct_equivalence_continuous` | | Oncology Phase II Simon | Clinical | `simon_two_stage` | | Cluster RCT | Clinical | `cluster_rct` | | Omics screening + FDR | Biomarker | `discovery_fdr` | ## Not yet implemented Document for planning only — use external tools or custom simulation: - Group sequential / interim analyses - Crossover or repeated-measures RCT - Dunnett pairwise multi-arm comparisons - Paired / correlated ROC comparison - Partial AUC, NPV/PPV as primary endpoints - Time-varying accrual and dropout in survival - Bayesian assurance - Simon optimal design search (minimax / optimal) ## Related - [Choosing a power analysis](choosing-a-power-analysis.html) - [Support matrix](support-matrix.html) - [Formula reference](https://yaoxiangli.github.io/ggpower/articles/formula-reference.html) (pkgdown only)