The tests, by the question you are asking
- Compare two groups: t-test (independent Welch, paired, one-sample), Mann-Whitney U, Wilcoxon signed-rank
- Compare several: one-way ANOVA with Tukey HSD, Kruskal-Wallis
- Count tables: chi-square, Fisher's exact, McNemar
- Relationships: Pearson, Spearman, and Kendall correlation
- Models: linear, binary logistic, log-binomial, conditional logistic, Poisson, Cox proportional hazards, multinomial
- Time to event: Kaplan-Meier curves, log-rank test, life tables, cumulative incidence
- Diagnostics: ROC with AUC, sensitivity and specificity, likelihood ratios, ICC
- Many hypotheses: Bonferroni, Holm, and Benjamini-Hochberg corrections
Built to be checked
Every result reports the test statistic and p-value, with a confidence interval where the method has one, plus the formula it came from and a citation for your methods section. The engines are cross-checked against R, and each tool ships a worked example so you can see the expected data shape before you paste your own.
From spreadsheet to figure
Paste from Excel or Google Sheets, or upload CSV and Excel files. The tests run in your browser, so pasted data stays on your device. Results become publication figures: survival curves, ROC curves, forest plots, and more than 40 recipes in Figure Studio, with SVG output for journals.
Where it fits
Covexe is built for health research and evidence synthesis, so the statistics sit beside meta-analysis, network meta-analysis, GRADE, and PRISMA tools. It is not a replacement for R or SAS on advanced modeling: no mixed models, no survey weights, no factor analysis. For the tests above, it is free, with nothing to download or install.