Statistics Analyzer
Run the right test on your table and read the result in plain words.
Try before you buy.
- Free preview: the first rows of each result table.
- Locked until you unlock it: download and copy.
- Unlock: Pro pass, ₹179 for 30 days or ₹79 for 1 week, a one-time payment that never renews. Unlock: Pro pass, $5.99 for 30 days or $2.99 for 1 week, a one-time payment that never renews.
Ways to unlock shows how to get the full result.
Printing this result is locked in the free preview.
1. Your data
One table with a heading row: a CSV or TSV file, an Excel or OpenDocument workbook, or text you paste. It is read in this browser and never uploaded.
Or paste a table (from Excel, Google Sheets or a CSV file)
2. Choose the test
3. Results
The Word report holds the reading, the tables and the notes; the workbook has one sheet per table with full figures.
Back up your data and settings
One file with your table and these settings, to pick up later. It is not encrypted: anyone who gets the file can read your rows. Nothing is stored online.
Locked in the free preview. Opens the ways to unlock this result.
Locked in the free preview. Batch runs unlock with a pass.
Locked in the free preview. Query results unlock with a pass.
About the Statistics Analyzer
Run the common statistical tests on a table of your own, in the browser: descriptive statistics overall and by group; independent, Welch’s and paired t-tests; one-way ANOVA with Tukey–Kramer comparisons; the chi-square test of independence with Cramér’s V; Pearson and Spearman correlations; linear regression with coefficients, intervals and R²; and the Mann–Whitney U, Kruskal–Wallis and Shapiro–Wilk tests.
Not sure which test fits? Choose Suggest a test for my data: it checks whether the values look normal and whether the groups’ spreads differ, picks the test and shows those checks with the result. Every result comes as clear tables with effect sizes and intervals, a reading in words in the style journals use, and notes on the assumptions and on the rows left out because a cell was empty or not a number. Your file is read in this browser and never uploaded.
How to use it
- Open a CSV, TSV or Excel file, or paste a table, with one row per person or item and a heading row. No file at hand? Try it with sample data.
- Choose the test, or let the page suggest one. The form shows only the columns that test needs: the column to compare and the group column, two measurements for a paired test, or the outcome and predictors for regression.
- Pick the significance level (0.05 is the usual choice) and choose Run the test.
- Read the result: the reading in words, the table of statistics, and the checks and notes under it. With a Pro pass, download the Word report, the Excel workbook (one sheet per table) or the CSV, or copy the tables.
Examples
The sample data: 40 students, two study methods (Rereading, Flashcards), their scores.
Independent t-test: the mean of “score” in “Rereading” minus “Flashcards” is -5.56 (95% interval -10.069 to -1.051), t(38) = -2.496, p = 0.017, Cohen’s d = -0.789.
With Suggest a test, the page first checks normality (Shapiro–Wilk) and equal variances (Levene); both pass here, so it picks the independent t-test.
Two number columns of the sample data: score and study_hours.
Spearman correlation of “score” and “study_hours”: ρ = 0.784, p < 0.001, from 40 pairs.
Common uses
- Checking whether two groups in an A/B test, a class or a clinic differ, with an effect size as well as a p-value.
- Comparing three or more groups with ANOVA and seeing which pairs differ.
- Tables in the style of a methods section for coursework, a thesis or a report.
- Testing whether two answers in a survey are related (chi-square).
- Explaining one measure from others with linear regression.
How the tests are worked out
Missing cells and text in number columns are left out and counted, never filled in. t-tests report the mean difference with its confidence interval and Cohen’s d; ANOVA reports η² and ω², with Tukey–Kramer intervals for each pair; chi-square reports Cramér’s V and warns when expected counts are below 5. Regression uses least squares on centred and scaled columns and reports R², adjusted R² and the condition number. Shapiro–Wilk follows Royston’s AS R94 method as SciPy implements it.
Sources: the NIST/SEMATECH e-Handbook of Statistical Methods and SciPy’s Shapiro–Wilk documentation.
Limitations
- Student-t and studentized-range probabilities come from numerical approximations.
- Shapiro–Wilk needs 3 to 5,000 values; a large p-value does not prove that the values are normal.
- A p-value does not show a cause, how much an effect matters, or whether the data suit decisions about people. Check how the data were collected.
Privacy
On this device; its files download once
Your file is read in this browser and never uploaded; no account is needed and nothing about your data is stored online.
Frequently asked questions
What do I get without a pass?
Without a pass, Statistics Analyzer shows the first rows of each result table. Until you unlock it, the result can’t be downloaded or copied. A Pro, Premium or Ultimate pass, a one-time payment that never renews, unlocks the full result. The pricing page lists the passes and their prices.
Which test should I use?
Choose Suggest a test for my data. With a group column it compares groups: a t-test or Mann–Whitney for two groups, ANOVA or Kruskal–Wallis for more, depending on the normality and equal-variance checks it shows. Without one it correlates number columns, or describes a single column.
What happens to empty cells?
Rows with an empty cell or text in a column a test needs are left out of that test, and the result says how many and which rows (counting from the first row under the headings).
What does the free preview show?
The first rows of each result table, with the figures of a one-row result hidden. A Pro pass opens the full tables, the reading in words and every download.
Can I use the results in a report?
Yes. The Word report holds the reading, the tables and the notes; the Excel workbook has each table on its own sheet with full figures.