Blind Sensory Testing of Frozen Ingredients: Codes, Temperature and Serving Order

Blind sensory testing can make a frozen-ingredient comparison more useful when participants receive neutral sample codes, comparable preparation and serving conditions, and a planned presentation order. The test still needs a defined question. Detecting a difference, describing texture and choosing a preferred sample are different tasks.

For a commercial trial, connect that question to the product and its intended use before preparing the samples. Keep the identity key, preparation record and individual responses together. A well-organised session should help the buyer decide what to investigate or approve next, within the limits of the test actually conducted.

Choose the sensory question before selecting the test

“Compare these two samples” leaves too much open. A product developer may want to know whether a candidate broccoli floret becomes noticeably softer in a ready meal. Purchasing may want to know whether the candidate can replace the current specification. A colleague may simply prefer its colour. Those observations can all matter, but they answer different questions.

Write the immediate question in one sentence. For example: “Under the agreed meal preparation, how do the two identified broccoli samples compare for perceived stem firmness?” This gives the trial team something more precise to design than an overall judgement of which supplier is “better”.

NC State’s sensory evaluation overview distinguishes consumer-response work from trained descriptive assessment. Consumer work can examine liking or other responses from the relevant audience; descriptive work characterises attributes and their intensity. A trained panel’s firmness profile and a consumer group’s preference are therefore different types of evidence.

A formal discrimination test asks another kind of question. In a triangle test, participants receive three coded samples, two from one product and one from the other, and identify the odd sample. The result requires analysis appropriate to that design. It does not, by itself, explain the nature of a difference or establish which product consumers prefer.

For an early internal review, a small technical team may be able to identify useful questions about the material. Describe such a session accurately. Notes from colleagues working on a recipe do not automatically represent the target market, and a handful of informal responses should not be relabelled as a validated consumer study.

Separate the next development decision from the eventual supply decision. The first session might decide whether a candidate deserves a larger application trial. Final approval may also depend on the written specification, manufacturing performance, relevant safety evidence, supply consistency and commercial requirements. A favourable tasting comment is only one contribution to that wider review.

Agree the expected output before anyone sees the codes. The team might need attribute comments, a properly designed difference assessment or another defined deliverable. If the request changes after tasting begins, record that change and decide whether a separate session is needed. Combining several new questions into the same hurried exercise makes its conclusion harder to defend.

Hide product identity while keeping sample records intact

Blinding concerns what the evaluator knows. It should not remove the coordinator’s ability to identify the material. Keep a controlled key connecting every serving code with its product, lot, sample reference and preparation record. The person arranging the trial needs that information even when the participant’s bowl carries only a neutral number.

For a proposed two-product session, codes such as 847 and 263 can replace supplier names on the serving vessels and response form. These are illustrative numbers, not reusable assignments for every trial. Avoid codes or visible notes that reveal the expected ranking, price, source or reason the sample was submitted.

The Wageningen University chapter on sensory study preparation describes noninformative coding and control of cues unrelated to the question. Expectations, contrast and the first sample encountered can influence later responses. Neutral labels and a suitable presentation plan reduce avoidable bias; they do not make the products physically identical.

Review the complete presentation. A supplier’s outer carton on the tasting table, a sample name written beneath a bowl or an organiser’s introduction can reveal information that the code was meant to hide. Prepare the coded servings away from the participants and keep the identity sheet with the coordinator.

Use serving vessels appropriate to the test and keep irrelevant differences under control. If one sample arrives in a deep bowl and the other on a flat plate, the presentation may no longer support the intended comparison. Record the vessel and portion arrangement so the next session can reproduce the agreed setup.

Do not conceal a characteristic that the test is supposed to examine. If visible floret shape is part of the buyer’s requirement, changing the lighting or hiding the sample may remove useful information. If appearance is outside a specific taste question, the sensory specialist can decide how to address it. The design should follow the question, rather than apply every possible form of concealment.

Check the key against the served samples before the session, using the agreed verification process. A perfectly completed score sheet is of little use if two codes were transposed during preparation. Keep any correction traceable and preserve the original entry instead of silently overwriting it after the responses are known.

Tell participants only what they need to perform the task and participate appropriately. Necessary ingredient and allergen information must be handled through the trial’s safety arrangements; blinding is never a reason to withhold information needed for safe participation. Keep those arrangements separate from promotional descriptions that could influence an evaluation.

Control preparation, holding and serving temperature

A frozen ingredient changes state between the sample freezer and the tasting bowl. The trial request should identify which state the buyer needs to assess: an appropriately prepared ingredient, a component in a finished meal, or another defined application stage. “Taste at the same time” does not describe how the two servings reached that point.

Follow the product’s applicable cooking and handling instructions and the trial team’s safety procedure. Frozen does not automatically mean ready to eat. The Food Standards Agency’s freezing and defrosting guidance explains that freezing does not kill all bacteria and calls for attention to product instructions. Its domestic guidance is not a complete commercial tasting protocol.

For the proposed broccoli comparison, identify the preparation method, equipment, loading arrangement and endpoint agreed by the responsible team. Record deviations that could affect interpretation. If one batch was prepared differently, the comments may describe that preparation difference as well as the ingredients.

Specify the serving condition in the trial plan. There is no single temperature that answers every frozen-fruit, vegetable and mushroom question. A hot meal component and a chilled fruit preparation have different uses. The relevant target, acceptable working range, measurement method and handling limits need to come from the actual product and test design.

Record the time between preparation and serving where it matters to the question. If one set waits while another is freshly prepared, the team should know that before interpreting a difference in eating quality. A serving log can connect each code with its preparation group and actual presentation time without adding those details to the participant’s ballot.

The Society of Sensory Professionals’ reporting guidance includes preparation, serving size and temperature, holding time, order, breaks and the testing environment among the details needed to understand a sensory study. For a buyer’s trial, select the controls relevant to the agreed method and retain the actual preparation and serving records.

Check that the serving plan is workable before inviting the panel. If the kitchen cannot prepare and present the required samples under the agreed conditions, revise the session design first. A smaller, properly scheduled comparison can be more informative than a larger set with undocumented delays.

Match the comparison to the ingredient and its application

Decide whether the trial is comparing the supplied material under common preparation or comparing alternative ingredients in their intended finished recipes. Those are both useful development questions. They require different records, because a change in recipe or process can affect what the evaluator experiences.

For whole frozen strawberries, first define the intended use. A fruit preparation made for a filling may be assessed at a different stage from fruit pieces intended for a chilled dessert. Record the actual sample form and whether the fruit was prepared whole, cut or processed before assessment.

In a proposed filling trial, record which ingredient version entered each recipe and whether the formulation was held constant. If the developer changes the sugar level or preparation method for one candidate, that may be a reasonable development step. Report it as a comparison of the resulting formulations, rather than attribute every observed difference to the fruit alone.

The same distinction applies to vegetables. A buyer may compare two broccoli specifications under a common meal process, then separately assess an adjusted process for a promising candidate. Give those stages different trial references. This preserves the useful first comparison while allowing development to continue.

Define the words on the response form. “Soft” can refer to a stem, a floret’s outer portion or the whole bite. Ask the trial designer which attribute and sample part should be assessed, and what guidance evaluators need. An ambiguous scale can produce numbers that look comparable while describing different observations.

Keep preference comments separate from attribute descriptions. A participant may describe one sample as firmer and still prefer the other. For a ready-meal buyer, the preferred direction depends on the product brief and intended eating occasion. Combining firmness and liking into a single unexplained “quality” score hides that distinction.

If an approved reference is part of the comparison, identify its role and condition. Our guide to using a golden sample alongside a written specification covers reference custody and its relationship with the specification. A reference should enter the session through the agreed design, with its identity managed appropriately for the chosen task.

Before requesting another supplier sample, write down what would make it a useful addition. A different cut, product form or specification may require a revised application question. Adding an unrelated sample to the same session can increase work without resolving the original choice.

Plan serving order before the session starts

Neutral codes do not control order by themselves. If every participant receives the current ingredient first and the candidate second, product identity is still tied to a presentation position. The organiser should arrange the sequence as part of the design and retain the actual order served.

For a two-product comparison, changing the first sample is one part of the presentation plan. The trial designer should determine allocation across participants, sessions and any repeats according to the method and intended analysis.

In the proposed two-product broccoli review, the coordinator would prepare the serving list before the session. Each participant’s record would identify the order assigned and the order actually delivered. If a tray is delayed or replaced, the organiser would note the change rather than assume the planned sequence still describes the event.

Formal methods have their own presentation requirements. For a triangle test, the Society of Sensory Professionals identifies six possible arrangements of the two products across the three positions. The three visible codes must support that design without revealing which two servings originate from the same product. Copying a two-bowl example into a triangle test would miss that requirement.

Keep the task manageable. The sensory team should decide how many samples and questions belong in a session, considering the products and method. A schedule that works for one ingredient may be unsuitable for a more intense or demanding comparison. Do not use a universal sample count simply because it appeared in another trial report.

Agree any breaks and palate-cleansing arrangements with the method. Record them along with the serving plan. A pause introduced only after an unexpected result, or a different accompaniment for one candidate, can complicate the comparison. These details are easier to organise beforehand than reconstruct from memory.

Collect individual responses before group discussion where the test requires independent evaluations. In a technical meeting, a senior colleague’s first comment can quickly become the language everyone else uses. A later discussion may be valuable for planning development, but keep it distinguishable from each participant’s original response.

When repeating a session, preserve the relationship between the repeated observations and the original design. More rows in a spreadsheet do not necessarily mean more independent evidence. Have the responsible sensory or statistical reviewer decide how repetitions should be scheduled, identified and analysed.

Keep the blind results connected to a clear decision

Complete the response check before revealing identities. Confirm that the ballots can be linked to the correct session and serving codes and that missing entries or unusual events are recorded. Retain participants’ actual words; a coordinator’s later summary should not replace the underlying observations.

Interpret the outcome using the method chosen at the start. A difference test, an attribute assessment and a preference exercise should not share an improvised pass rule. If a formal analysis does not establish a difference, do not automatically conclude that the products are equivalent. Ask what the design and its sensitivity allow the team to say.

The Journal of Food Science’s sensory reporting requirements distinguish consumer and trained-panel work and call for appropriate detail about methods, training and reliability. Its publication requirements are not universal rules for an internal buyer trial. The useful lesson for a commercial report is to make the basis and limits of the evidence visible.

Unblind according to the agreed process and connect the results to the product key. The report should state the question, identified materials, preparation and serving conditions, method, participants or panel description, relevant deviations and supported conclusion. It should also identify what the session did not assess.

For the hypothetical broccoli candidate, the next step might be another application trial, clarification of a size specification or a review of preparation consistency. Describe the action that the evidence supports. A session involving one submitted lot cannot on its own establish performance across all future production.

In an XMG supply review, we need the sample reference and intended application to make sensory feedback actionable. Comments such as “too soft” become more useful when they identify the product part, preparation and comparison material. We can coordinate the relevant questions with the partner factory and review available product information.

Use our specification and sampling process to connect any next sample request with the requirement under review. Where observations need a quality follow-up, we will coordinate through the quality and inspection arrangements. The responsible trial team designs and interprets the sensory work.

Close the trial record with the decision, its owner and the next action. Keep any approval conditions with the relevant ingredient specification. This allows purchasing and product development to work from the same account when the discussion moves from a coded bowl to a commercial order.

References

• Society of Sensory Professionals. Guidance for sensory data publications; Triangle Test.
• NC State University. Sensory Evaluation.
• Wageningen University. Sensory Nutrition, chapter 11: Methodological Considerations When Setting up Research, 2025.
• Food Standards Agency. How to chill, freeze and defrost food safely.
• Institute of Food Technologists. Journal of Food Science Author Guidelines, Sensory Evaluation subsection.