# One Crop, One Useful Question

Author: Bennett Cawthon
Published: 2026-09-04
Updated: 2026-10-16
Canonical URL: https://www.streamlinefarms.com/blog/world-biological-outcome-model
Article type: Proposed research protocol

## How should Streamline begin studying whether crop practices improve a neighborhood food outcome?

Begin with one crop and predefined outcomes, retaining cultivar, conditions, interventions, sample methods, saleable yield, and costs. The proposed lettuce study tests a specific growing comparison before broader transfer. Food composition estimates are not measured batch results, and composition changes do not establish human health benefits. The World Biological Outcome Model remains a research ambition.

Suppose a grower changes one lettuce practice. Does the crop produce more usable food, meet the kitchen's quality needs, or change the cost of a meal? Answering requires a comparison designed before anyone sees the result.

The study design below starts with one cultivar and one useful question. It is a research plan, not a completed trial. The value to a neighborhood farm is a production choice someone can repeat and a result someone can measure.

## A first research question

For one lettuce cultivar and production setting, compare a defined baseline practice with a specified alternative growing practice. Record saleable yield, accepted quality, production cost, and a preselected food-composition measurement.

The alternative practice and laboratory endpoint must be specified before the study. This is a proposed protocol structure, not a completed trial or a claim that a particular intervention improves nutrients.

| Record | Why it belongs |
|---|---|
| Cultivar and planting batch | Establish biological identity and comparability |
| Bed, crop stage, and environmental observations | Preserve the setting of the comparison |
| Assigned and completed intervention | Distinguish the plan from physical execution |
| Harvest and saleable yield | Count output meeting a stated standard |
| Sample identity, handling, and laboratory method | Connect the measured composition to the sampled product |
| Labor, inputs, and capital allocation | Determine whether the practice pays |

## Compare before explaining

Use comparable replicated units, an assignment method appropriate to the site, and a sample-size plan based on expected variability and the effect that would matter. Predefine exclusions and missing-data handling. A qualified research partner should design the actual study.

An observational model can forecast an association. Claiming that a deliberate intervention caused a difference requires a comparison that addresses alternative explanations. A result at this site then needs a separate transfer test.

## Nutrient demand has several layers

A household's preferences and nutrient goals guide menu planning. USDA's [FoodData Central](https://fdc.nal.usda.gov/) supplies composition data for estimating ingredient contributions. Those estimates do not establish the measured composition of our particular harvest or cooked meal.

If a crop-composition claim matters commercially, sample it with a declared protocol. If a health claim is proposed, it requires a different evidence program. The farm should not convert a compositional measurement into a promise about a person's health.

Keep the next use visible as well. A harvested ingredient and a prepared lunch are different sampled products: storage, preparation, recipe quantities, and cooking can change the composition of what is served. A crop study cannot silently become a claim about the finished meal. Follow accepted lots into the chosen recipe when that is the commercial question.

## What could compound

Shared definitions let another team compare crops, observations, actions, samples, accepted food, customer use, and costs. That is the practical ambition behind the World Biological Outcome Model. Broader predictions have to follow successful, repeatable studies of specific questions.

Keep participating growers' rights and contribution terms explicit. Reusable knowledge does not justify unrestricted access to private farm or household records.

Streamline's current greenhouse replay addresses an earlier task: preserve the records needed to investigate what happened. [Provenance](https://www.streamlinefarms.com/blog/sensor-provenance-agricultural-ai/) explains that record keeping; [the neighborhood thesis](https://www.streamlinefarms.com/blog/biology-first-autonomous-food-system/) connects the research question to food people buy and use.

[Propose a research collaboration](https://www.streamlinefarms.com/work-with-us?intent=research_collaboration) around one crop, one declared comparison, and an outcome worth measuring.