NO SAS LICENCE. NO R SCRIPT. NO SYNTAX ERROR AT 2 A.M.

The C.D. table your guide will actually accept.

Paste your trial data. Pick the design it was actually laid out in — CRD, split-plot, augmented, whatever your field looked like. KrishiStat works out the ANOVA, the right error term, the Critical Difference, and writes the interpretation paragraph the way your department expects to read it.

Start free — no card needed See a real result

7 days, every module unlocked. Cancel by simply not paying — there's nothing to cancel.

Analysis of Variance

RBD · 4 treatments · 3 reps

Source	df	SS	MS	F
Replication	2	4.19	2.10	—
Treatment	3	168.34	56.11	27.84**
Error	6	12.09	2.02	—
Total	11	184.62

C.D. (5%) = 2.61 · CV = 4.8% · Treatment significant at 1% (**) — this is one table out of a full report with means, letter groupings and a written summary.

WHY THIS EXISTS

Most trials aren't complicated. Writing them up is.

Three replications, a handful of treatments, maybe a covariate — the design itself is usually the easy part. What eats an evening is getting from the raw plot numbers to a table an examiner will actually sign: the correct error stratum for a split-plot, the right critical range for DMRT once the treatment count changes, the exact phrase for "significant at the 1% level" that doesn't come back with a red pen through it.

That's the part between the data and the table. That's what this does.

Nobody fails a viva over the F-value. They fail it over the paragraph underneath it.

WHAT'S ACTUALLY INSIDE

Six areas, built for the analyses agricultural research actually runs

Not a general-purpose stats package with agriculture bolted on — every method here exists because a field trial, a breeding programme or a survey dataset needed it.

Design of Experiments

Loading…

–

HOW IT WORKS

Three steps, no installation

Nothing to download, no licence key to email your department for. It runs in the browser you already have open.

Paste your data

Straight out of Excel — tabs, commas, a plain list, it doesn't matter. Column headers are read, not guessed at.

Tell it the actual design

CRD or RBD, how many factors, whether it was a split-plot or a strip-plot — the same choice you'd make filling in a lab notebook.

Get the whole write-up

ANOVA, means, C.D. with letter groupings, a chart where one's actually useful, and the interpretation paragraph — ready for Word, Excel or PDF.

BEFORE ANY OF THAT

The randomisation, done before the seed goes in

Half the analyses that go wrong were doomed at sowing. The layout tool draws the plan first, so the design you analyse is the design you actually planted.

◱

Twenty-one designs, properly randomised

CRD through split-split-plot, lattices, augmented, row-column, Box-Behnken. The lattices are built from orthogonal Latin squares rather than reshuffled, so the balance the analysis assumes is actually there.

A plot list you can carry

Plot number, replication, block, row, column, treatment — numbered serpentine, the way the field is walked. Export it, add a column per character, and the finished sheet pastes back into the matching module.

⟳

The same plan, years later

Every layout carries a short code. Write it in the field book and the identical plan comes back whenever you need it — for the thesis, the report, or a second season on the same ground.

It also prints the skeleton ANOVA with the degrees of freedom the design will yield. An error line in single figures is worth knowing about now, not after harvest.

NOT A MOCKUP

This is what a finished table looks like

A two-factor RBD, post-hoc separated by DMRT. This is the platform's real output style — the same table you'd paste into a manuscript.

Treatment Means & Grouping

DMRT · 5% level

Treatment	Grain yield (q/ha)
T2 — N₁₂₀P₆₀K₄₀	32.6^a
T4 — N₉₀P₆₀K₄₀	29.8^ab
T3 — N₁₂₀P₃₀K₄₀	27.1^bc
T1 — Control	21.4^c

Superscript letters are the DMRT grouping — means sharing a letter aren't significantly different.

"T2 recorded the highest mean grain yield (32.6 q/ha), significantly superior to T3 and T1, but on par with T4 (C.D._5% = 3.2)." — written automatically, not templated fill-in-the-blank.

NOT A DEMO

What's actually running on this server right now

Real analysis counts pulled live from the same database every run writes to — not a marketing estimate.

–

Analyses run to date

–

Modules available

–

User rating

Daily Analyses — Last 30 Days

PRICING

Priced for a research budget, not a corporate one

One thesis, one season, or a running project — pick what matches, upgrade later if it doesn't.

Loading current plans…

QUESTIONS

Things people actually ask before signing up

Do I need R, SAS or any plugin installed?

No. Everything runs on our server; your browser just shows the result. There's nothing to install and nothing that needs a licence file.

How is the Critical Difference actually calculated?

From the correct error mean square for the design you ran — the sub-plot error for a split-plot, the pooled error where designs call for it — using the studentized range or t-distribution as the method requires, not a flat approximation. LSD, DMRT, SNK and Tukey's HSD are all supported and computed the same way a statistics textbook would derive them.

Can I actually put this table in my thesis?

Yes — that's the point of it. Tables come out in a clean, unruled publication format by default, and paid plans export straight to Word, Excel and PDF if you'd rather not copy-paste.

What happens after the 7-day trial?

Your account doesn't disappear — it drops to the free tier, which still runs single-factor designs. Everything you've already saved stays exactly where it was.

How do I pay, and how fast does it activate?

UPI or a direct bank transfer. Upload the transaction reference on the subscription page and it's checked against the bank statement — usually the same day, not a support ticket queue.