Formulation, sensory tests and cooking

Begin with the dish it must serve

Suppose a developer is comparing two bean-based mixtures for a vegetable soup. The soup must pour from a ladle and feel smooth enough for its intended audience. A mixture that forms a strong gel may perform well in another food, but be unsuitable here. The target is a named use, not maximum firmness or the longest ingredient list.

Formulation means choosing ingredients and quantities to provide the required composition and behaviour. Write down what must be evaluated: flavour, texture, preparation, nourishment and cost. These can interact. Changing an ingredient to improve one property may alter another, so a single successful test does not finish the evaluation.

Choose the sensory question first

A difference test asks whether samples can be distinguished. An attribute assessment asks about properties such as sourness, aroma or thickness. An acceptance or preference test asks what people like or choose. These are different questions: a detectable change need not be preferred.

Trained assessors can describe a sensory profile using agreed attributes. Intended users can report liking in the relevant setting. Neither group automatically represents everyone. Select participants and the test according to the question rather than turning a small convenience group into a universal consumer verdict.

Keep the comparison interpretable

Compare samples prepared and served under matching conditions unless a particular condition is the variable being investigated. Portion, temperature and time after preparation may otherwise confound the result. Record the ingredients and preparation for each sample so the comparison can be reproduced.

Codes that conceal formulation names reduce a source of expectation. Varying or randomising serving order helps separate a formulation difference from an order effect. These choices support a fair design; they do not compensate for unsuitable samples or an unrepresentative audience.

For a focused development question, change a defined formulation factor and track the rest. If several ingredients and cooking conditions change together, a preference result may still compare the finished products, but cannot identify which individual change caused it.

Measure the food as well as the response

A flow or texture measurement can help explain whether the soup pours consistently. A sensory panel can assess grittiness, aroma or liking. Instrument readings and sensory results complement each other; an instrument does not simply measure enjoyment.

Plant-protein research illustrates the boundary: laboratory lubrication measurements can show differences between material structures while actual sensory evaluation remains necessary. For the soup, identify the physical test and conditions, then investigate whether its result relates to the experience being sought.

Food samples offered to participants also need appropriate ingredient and preparation controls. A sensory test is not a safety assessment or a way to test whether an unknown product is safe. Assess the actual product and hazards before interpreting any tasting as development evidence.

Count people and preserve the limits

In an invented preference exercise, 20 participants each compare A and B once. Twelve choose A and eight choose B. The observed proportions are 60% and 40%. The diagram preserves that denominator and ends with a bounded result for this group.

That count alone does not establish a population preference or statistical significance. Recruitment, uncertainty, serving conditions and the relevant use still matter. Repeated responses from the same person should not be silently counted as additional independent people.

Separate the observation from the explanation. If A is preferred here, the result does not prove that greater protein content, a particular aroma or a new production method caused that preference. More focused measurements or comparisons may resolve the cause.

Return to nourishment and practical use

Read the finished product's composition on the serving and preparation basis actually tested. Nutrient availability needs its own evidence; liking does not establish adequacy. Conversely, a favourable nutrient number does not guarantee that people will enjoy the food or use it as intended.

Compare cost per usable serving and include preparation inputs when the decision calls for them. Batch price alone can mislead when yields differ. Equipment, labour, energy and waste can change feasibility, even if an ingredient substitution looks affordable.

A development conclusion should identify the tested use, audience, conditions and remaining uncertainties. An acceptable prototype is progress. Reproducible preparation, nutritional evidence, safety and economic performance remain necessary parts of deciding whether it serves that use well.

Worked examples

Check your understanding

1. Assessors can detect a difference between two soups. Does that establish which is preferred?

   Show answer and reasoning

   No.

   Detection and preference answer different sensory questions.

2. Ten people each rate a soup twice. Are there 20 independent participants?

   Show answer and reasoning

   No; there are ten participants and 20 responses.

   Repeated ratings from one person are related observations, not new independently recruited people.

3. A revised soup changes flour, oil and heating together and is preferred. Has the oil alone been identified as the cause?

   Show answer and reasoning

   No.

   The comparison can support a finished-product result, but the combined changes do not isolate one ingredient's effect.

Key ideas

• Define the culinary use before choosing a formulation or test.
• Keep difference, attribute and liking questions separate.
• Combine physical measurements with bounded sensory evidence.
• Compare nourishment and cost on the prepared serving actually being evaluated.