Validated desktop statistics

Sigma
Statistics

Run guided statistical analyses from descriptive summaries to survival and competing risks — without coding or uploading sensitive research data.

●Download on theMac App Store ⊞Get it fromMicrosoft

One-time purchase · No monthly subscription

11 documented modulesFull-precision validationLocal processingDOCX + XLSX

Sigma Statistics

Statistical tools

Focused methods.
Transparent outputs.

Eleven structured workflows cover routine summaries, inference, regression, matching, time-to-event analysis and competing risks.

Descriptive StatisticsMetric and categorical summaries, grouped tables, missingness, percentiles, skewness and kurtosis.

Normality TestingShapiro–Wilk, Lilliefors, Anderson–Darling and D’Agostino–Pearson K².

Chi-Square & Fisher’s Exact2×2 inference with expected counts, effect measures and low-count warnings.

Unpaired t-TestWelch and pooled-variance comparisons with confidence intervals and effect sizes.

Paired t-TestComplete-pair analysis, paired differences, confidence intervals and standardized effects.

One-Way ANOVAClassical and Welch ANOVA, Brown–Forsythe, effect sizes and robust post-hoc comparisons.

Univariable Logistic RegressionOdds ratios, Wald inference and explicit separation warnings.

Multivariable Logistic RegressionAdjusted associations, convergence diagnostics and events-per-variable information.

Propensity Score MatchingATT-oriented 1:1 nearest-neighbor matching, overlap and balance diagnostics.

Survival AnalysisKaplan–Meier, log-rank, RMST/RMTL and Cox regression with robust inference.

Competing-Risk AnalysisCumulative incidence functions, Gray’s test and Fine–Gray regression.

Independent reference workflows

Validation reported at full precision.

Production calculation cores were compared against independently implemented R workflows using deterministic datasets, targeted edge cases and module-specific acceptance criteria.

256,822 / 256,822descriptive-statistics comparisons passed

104,770 / 104,770ANOVA comparisons passed

3,833,940 / 3,833,940survival-analysis comparisons passed

1,080 / 1,080PSM runs with exact pair identity

2,089,198paired cumulative-incidence time points

Sigma distinguishes stable results from malformed, degenerate, separated, singular and otherwise non-estimable cases instead of silently presenting unreliable inference.

Review the complete validation compendium →

Data sovereignty

Your data never needs to leave your computer.

Sigma is a standalone desktop application. CSV data is imported and analyzed locally; sensitive clinical or scientific datasets are not uploaded to a GBM Consulting calculation server.

Defined preprocessing rules, full-precision calculation cores, explicit warnings and exportable results support traceable research documentation. Users remain responsible for study design, method selection, assumptions and scientific interpretation.

Read the complete methodology →

Scientific transparency

Methods you can inspect.
Results you can verify.

Sigma methodology ↗Sigma validation ↗Phi methodology ↗Published references ↗

Start analyzing with Sigma.

One purchase for your offline macOS or Windows workflow.