Agricultural Science Editing and Proofreading Services

Wheat yield rose nine percent under a foliar treatment. The result came from one field, in one season, on one soil type, in a year when June rain ran far above average. Within a year an agronomist quotes it to a grower four hundred miles away, on lighter soil, in a dry spring. Agricultural results travel further than their evidence, and the reporting has to say how far they can go.

The documents we edit for Agricultural Science

Field trial reports and their site-year structure, agronomy and variety trial publications, treatment and management description sections, and the reporting of soil, weather, and season form the written record of a trial, and that record is our work for agricultural scientists. It also covers statistical analysis and its design basis, extension and advisory material from trials, grant applications and project reports, on-farm and participatory research documentation, livestock and grazing trial reporting, and guidance for growers and advisors. In each case, our editors attend to the reporting that says where a result applies.

What the editing involves

The site-year structure is what makes an agricultural result transferable or merely local, and it fails most often when one mean is reported across everything. A nine percent response averaged over four sites and three seasons conceals whether the treatment worked at every site or at one. Results therefore belong site by site and year by year before they are pooled, because the pattern across sites is the finding, and the pooled mean only summarizes it.

Seasons need describing against the long-term average in the terms that decide a response: rainfall in the critical growth window, degree days, and the timing of any dry spell. Where one year gave 40 millimeters of rain in the two weeks after flowering and the next gave 110, the two years answer different questions. A season is not a replicate, and treating three of them as replications of one experiment is the field's most common overclaim.

The soil should be described with the properties that determine a response: texture, organic matter, pH, and any constraint such as compaction or low potassium. Those are what an advisor matches against their own fields. The management around the treatment belongs in the report in full: the variety, the rate, the timing, the previous crop, and everything held constant. A response obtained after a legume break may reverse after a second wheat.

Any site-year in which the treatment failed deserves the same prominence as the successes, because those cases define the boundary of the result. Any recommendation should state the conditions it assumes. The honest measure of the evidence is the number of site-years rather than the number of plots: a trial described as having 96 plots may rest on three. Reports written this way are used correctly by the advisors who read them.

Confidentiality and the limits of our role

Everything you send us is treated in confidence, including trial data, unpublished results, and project documentation. We are editors, not agronomists or statisticians, and we offer no opinion on treatments, results, or recommendations. What we can do is make the reporting say where the result applies.