Field Trial ANOVA Calculator: RCBD, CRD, Latin Square and Split-Plot with LSD, Tukey and Duncan Letters
Analyse your field or pot experiment the way agronomy theses and journals expect: RCBD, CRD, Latin square or split-plot. Enter your data (or upload your sheet) and get the ANOVA table with F-table values, CV%, SEm±, SEd, CD (LSD) at 5% and 1%, and mean separation by LSD, Tukey HSD and Duncan’s multiple range test with letters and a chart.
Type one value per plot. In a CRD you can leave cells empty when treatments have unequal replications.
- PriceFree, no login or sign-up needed
- Your dataProcessed in your browser, never uploaded
- Works onPhone, tablet or computer, any modern browser
- Made byDr. Muhammad Aammar Tufail, Codanics
- Published16 November 2026
- Last updated16 November 2026
📚 How to cite this tool
This tool is free. If it helped your thesis, paper, report or assignment, please cite it. It keeps the tools free and helps other students find them.
Tufail, M. A. (2026). Field trial ANOVA calculator (RCBD, CRD, latin square, split-plot) [Online tool]. Codanics. Retrieved 16 November 2026, from https://codanics.com/tools/field-trial-anova-calculator/
In-text citation: (Tufail, 2026) · Narrative: Tufail (2026)
The access date is today’s date, filled in automatically. Author: Muhammad Aammar Tufail. Published 16 November 2026.
How to use the field trial anova calculator (rcbd, crd, latin square, split-plot)
- Choose the design: RCBD if your plots were arranged in replications/blocks (most field trials), CRD for pot, lab or greenhouse experiments without blocking, Latin square when you blocked in two directions (rows and columns), or split-plot when one factor (e.g. irrigation or tillage) was applied to large main plots and a second factor (e.g. nitrogen or variety) to sub-plots inside them.
- Set the size of the trial and type one value per plot (for a Latin square also the treatment code of each plot; for a split-plot the levels of both factors, separated by commas), or use Paste from Excel / Upload CSV / Excel in long format. Each design loads its own example when you switch to it.
- Pick the mean separation test for the chart letters (LSD, Tukey HSD or Duncan’s DMRT) and the significance level, then click Run ANOVA.
- Copy the ANOVA table, precision values (SEm±, SEd, CD) and means with letters into your thesis, and download the bar chart (grouped bars for a split-plot) with letters.
CRD, RCBD, Latin square or split-plot: which design is yours?
| CRD | RCBD | Latin square | Split-plot (in RCBD) | |
|---|---|---|---|---|
| Layout | Treatments assigned completely at random to all plots | Field divided into blocks; every treatment appears once in each block | t × t grid; every treatment once in each row and once in each column | Blocks → main plots (factor A) → sub-plots (factor B) |
| Typical use | Pot, lab and greenhouse experiments with uniform conditions | Field trials with one fertility or slope gradient | Gradients in two directions (e.g. slope and irrigation channels), 4–8 treatments | A factor that needs big plots (irrigation, tillage, sowing date) combined with one that fits small plots (nitrogen, variety) |
| ANOVA sources | Treatments, Error | Replications, Treatments, Error | Rows, Columns, Treatments, Error | Replications, A, Error (a), B, A × B, Error (b) |
| Error df | N − t | (t − 1)(r − 1) | (t − 1)(t − 2) | (r − 1)(a − 1) and a(r − 1)(b − 1) |
How the split-plot analysis works
In a split-plot the main-plot factor A is randomised to whole plots within each block, and the sub-plot factor B is randomised to smaller plots inside each main plot. A is therefore compared with the larger Error (a) (the block × A interaction), while B and the A × B interaction are compared with the smaller Error (b). Using a single error term, as a two-way ANOVA would, gives wrong F values and wrong LSDs. The tool gives a separate CV for each error, and SEd and CD values for the four comparisons in Gomez & Gomez (1984):
| Comparison | SEd | t for CD |
|---|---|---|
| Two A means | √(2Ea / rb) | t at Error (a) df |
| Two B means | √(2Eb / ra) | t at Error (b) df |
| Two B means at the same level of A | √(2Eb / r) | t at Error (b) df |
| Two A means at the same or different levels of B | √(2[(b − 1)Eb + Ea] / rb) | weighted t′ = [(b − 1)Eb·tb + Ea·ta] / [(b − 1)Eb + Ea] |
In the A × B table, small letters compare B levels within each level of A and capital letters compare A levels at each level of B. Report the interaction means when A × B is significant, and the main-effect means when it is not.
For a two-factor factorial without main plots (both factors randomised together) use the two-way ANOVA calculator. For a simple comparison of groups without blocks, the one-way ANOVA calculator also gives Welch’s ANOVA and Games–Howell.
Formulas used
LSD, Tukey or Duncan?
- LSD / CD is the most common in agronomy theses in South Asia. It is the most liberal test and is best used only when the F-test is significant and for planned comparisons.
- Duncan’s multiple range test (DMRT) is popular in agricultural journals. It sits between LSD and Tukey.
- Tukey HSD is the most conservative and the one most statisticians and international journals prefer when you compare all pairs.
Check that the residuals are roughly normal with the normality test calculator, and make a publication chart with the chart maker. Learn to run these designs yourself in the R with Aammar course, or read about types of data analysis.
Frequently asked questions
What is CD in agricultural statistics?
CD (critical difference) is the least significant difference (LSD): CD = t(α/2, error df) × SEd. Two treatment means that differ by more than the CD are significantly different. Thesis tables usually give SEm±, CD at 5% and CV% under the means.
What is a good CV% for a field experiment?
It depends on the crop and trait, but for grain yield a CV below about 10–15% is usually considered good and above 20% suggests high experimental error. Lab and greenhouse traits often have lower CVs.
Should I use LSD, Duncan or Tukey?
Tukey HSD controls false positives best and is preferred by many international journals. Duncan’s test (DMRT) is widely used in agricultural journals, and LSD is common in theses. Whatever you choose, apply it only when the treatment F-test is significant, and state the test and probability level.
Can I analyse unequal replications?
Yes, choose CRD and leave the missing cells empty. SEm±, SEd, CD and the Duncan ranges then use the harmonic mean of the replications, while LSD and Tukey letters use the exact replications of each pair. An RCBD, Latin square or split-plot with missing plots needs missing-plot estimation or a mixed model, which this tool does not do.
Do the results match R, SAS or OPSTAT?
Yes. The ANOVA, CV, SEm, CD, Tukey and Duncan values were checked against R (aov, aov with Error(block/A) for the split-plot, TukeyHSD, qtukey and the Duncan formula used by the agricolae package). Letters can be named differently by different programs, but the grouping of treatments is the same.
How do I analyse a split-plot design?
Choose Split-plot (in RCBD), type the levels of the main-plot factor and the sub-plot factor, and enter one value per sub-plot for each block. The tool tests the main-plot factor against Error (a) and the sub-plot factor and interaction against Error (b), and gives CV (a), CV (b), and the SEd and CD for all four standard comparisons.
What is Error (a) and Error (b) in a split-plot ANOVA?
Error (a) is the variation between main plots (the block × main-plot factor interaction) and is used to test the main-plot factor. Error (b) is the variation between sub-plots within main plots and is used to test the sub-plot factor and the interaction. Error (b) is usually smaller, so sub-plot effects are estimated more precisely.
How do I analyse a Latin square design?
Choose Latin square, set the number of treatments (the square is t × t), and copy your field layout: the treatment code and the value of each plot. The tool checks that each treatment occurs once in every row and column, removes the row and column effects, and compares treatment means with LSD, Tukey or Duncan’s test.
Can it analyse a strip-plot or a factorial RCBD?
Not yet. Use the two-way ANOVA calculator for a two-factor factorial in a CRD. Strip-plot and factorial RCBD designs are planned for a future version.
Is my data uploaded?
No. All calculations run in your browser, including files you upload. Nothing is sent to or stored on our server.
Do I need to log in or create an account to use the field trial anova calculator (rcbd, crd, latin square, split-plot)?
No. There is no login, no sign-up and no email required. Open the page and start using it straight away, for free, on any device.
How do I cite this tool in my thesis or paper?
Use the citation box on this page. In APA 7th edition: Tufail, M. A. (2026). Field trial ANOVA calculator (RCBD, CRD, latin square, split-plot) [Online tool]. Codanics. Retrieved [date you used it], from https://codanics.com/tools/field-trial-anova-calculator/. BibTeX, RIS (EndNote, Mendeley, Zotero), MLA, Harvard, Chicago and Vancouver formats are also available.
Free guides and courses from Codanics
More free tools
See all free Codanics tools → · Codanics on YouTube · WhatsApp channel

