Field data → finished analysis

A statistics bench for the data on your desk.

StatLab reads a spreadsheet the way you already keep one — variable names across the top, observations underneath — and runs the tests a veterinary, agricultural, medical or social-science dataset usually needs: description, comparison, correlation, regression, analysis of variance. Everything runs in this browser tab; nothing is uploaded anywhere.

What a file should look like

Row one holds variable names. Every row below is one observation. Numeric and grouping columns can sit side by side — StatLab sorts out which is which.

Weight	Height	Group
25	70	A
28	73	A
32	76	B
30	75	B

On the bench

Descriptive statisticsN, mean, median, mode, SD, SE, CV, skewness, kurtosis, quartiles.

t-TestsOne-sample, independent-samples (with Levene's test), paired.

CorrelationPearson and Spearman, plus a full pairwise matrix.

RegressionSimple and multiple linear regression, with VIF for collinearity.

One-way ANOVAF-test with Bonferroni-adjusted pairwise post-hoc comparisons.

GraphsHistogram, box plot, scatter with fitted line, group bars, correlation heatmap.

01 — Upload data

Bring in a spreadsheet

Accepts .xlsx and .xls, up to 20 MB. The first sheet is read; its first row becomes variable names.

Drop a spreadsheet here, or click to choose one

.xlsx or .xls — processed locally in your browser

02 — Descriptive statistics

Describe your variables

Numeric variables

03 — t-Test

Compare a mean

Variable

Test mean (μ₀)

Numeric variable

Grouping variable (2 groups)

Variable 1

Variable 2

04 — Correlation

Relationships between variables

Method

Variable A

Variable B

05 — Regression

Model one variable from others

Dependent variable (Y)

Independent variable(s) — pick one for simple regression, several for multiple

06 — One-way ANOVA

Compare means across groups

Dependent variable

Grouping variable

07 — Graphs

See the shape of it

Variable

08 — Download results

Your session report

Every analysis you run is logged here in order, formatted like a lab report. Download it as a plain-text file, or copy the descriptive-statistics table as CSV.

No analyses run yet this session.

09 — Help & about

Reading the output

Interpreting p-values

Throughout StatLab, a result is called significant at p < 0.05. This is a convention, not a law — for small samples or exploratory work, treat borderline values (0.05–0.10) as worth a second look rather than a clean verdict.

What StatLab does and doesn't do

Everything — parsing, computation, charting — happens in your browser. No file or result leaves this tab. Missing or non-numeric cells in a numeric column are excluded from that column's calculations automatically; StatLab notes the count it dropped. Post-hoc comparisons after ANOVA use Bonferroni-adjusted pairwise t-tests rather than Tukey HSD — more conservative, and simpler to audit by hand.

Errors you might see

Message	Cause
Column is not numeric	Selected variable has too few numeric values to compute this test.
Grouping variable needs exactly 2 groups	Independent t-test requires precisely two categories.
Singular matrix	Predictors in a multiple regression are perfectly collinear.
Sample too small	Fewer observations than the test needs to compute degrees of freedom.