About this project

TCC Tools grew out of a concrete and very common problem: students finish collecting data for a thesis or dissertation and stall at the analysis stage. Not for lack of effort, but because the established statistical package costs more than a student budget allows, has a steep learning curve, and returns raw output that still has to be translated into the formatted prose a reviewer expects.

Our aim is to close exactly that gap: you bring the spreadsheet, the site checks the assumptions, runs the appropriate test, produces a publication-resolution figure, builds the table to editorial standard and writes the results paragraph already formatted in APA style. From raw data to review-ready text in a few clicks.

Principles

• Privacy by architecture. All processing happens in your browser. Your spreadsheet never leaves your computer — this is not a policy promise, it is a technical property of the system. You can even disconnect from the internet after the page loads and the calculators keep working.
• Genuinely free. No sign-up, no usage limit, no premium tier. Costs are covered by unobtrusive advertising.
• Methodological transparency. We state which algorithm and which reference underpin every calculation, so that you can cite and defend the method.
• Educate, not just compute. Every result comes with an effect size and a confidence interval, not only a p value — because statistical significance is not the same thing as practical relevance.

Methods and references

Procedures are implemented from the consolidated statistical literature:

• Normality test: Shapiro-Wilk using Royston's (1995) AS R94 algorithm, including the p-value approximation via normalising transformation.
• Homogeneity of variance: Levene's test centred on the mean.
• Comparison of means: Student's t test and the Welch-Satterthwaite correction for unequal variances.
• Non-parametric tests: Mann-Whitney and Wilcoxon with normal approximation and tie correction.
• ANOVA: classical sum-of-squares decomposition, with multiple comparisons adjusted by Bonferroni.
• Correlation: Pearson with confidence intervals via Fisher's z transformation; Spearman by rank correlation.
• Effect size: Cohen's d, eta squared, epsilon squared and Cramér's V, with interpretive ranges following Cohen (1988).
• Sample size: normal-approximation formulas with the finite population correction, following Cochran (1977).
• Probability distributions: the regularised incomplete beta function via Lentz's continued fraction, and Acklam's algorithm for the normal quantile.

Validation

Every procedure is checked against R output (t.test, aov, shapiro.test, cor.test, wilcox.test) using reference datasets. Test statistics and p values agree to the third decimal place. Even so, and as a matter of principle, we recommend verifying critical results in reference software before final submission — guidance that also appears in our Terms of Use.

Found an error?

Discrepant results, unexpected behaviour or a suggestion for a new tool: get in touch through the contact page. Calculation corrections are treated as top priority.