How to Conduct Sensory Tests: Steps and Best Practices

Sensory testing is one of the most critical processes in food quality evaluation. It bridges the gap between what instruments can measure and what humans actually experience when they eat or drink. Whether you’re developing a new product, checking batch consistency, or studying consumer preferences, the way you conduct a sensory test determines the reliability of your results. This post walks you through the key steps involved in conducting sensory tests – from defining objectives to analysing data – along with best practices that keep your results accurate and actionable.

Why conducting sensory tests matters 🔗

Sensory evaluation is a scientific discipline that measures, analyses, and interprets human responses to food characteristics perceived through sight, smell, taste, touch, and hearing. It goes far beyond simply deciding if something tastes good. The process generates objective, reproducible data from human subjects who essentially serve as measurement instruments.

In the food industry, sensory tests support product development, quality control, shelf-life studies, and consumer acceptance research. They help food companies understand whether a reformulated product still meets expectations, whether a cheaper ingredient substitution goes unnoticed, or whether a new flavour profile resonates with target consumers. Without a systematic approach to conducting these tests, decisions become guesswork.

Step 1: Define the objective 🔗

Every sensory test begins with a clear question. What exactly are you trying to find out? The objective shapes every subsequent decision – the type of test, the number of panellists, the sample preparation, and the statistical analysis.

Common objectives include comparing two formulations, detecting the effect of a new ingredient, evaluating consumer preference among multiple products, or profiling the sensory attributes of a product in detail. For example, if you’ve replaced an expensive ingredient with a more affordable alternative, your goal might be to determine whether consumers can detect any difference. If you’re launching a new snack, you might want to know which of three flavour variants consumers prefer most.

A vague objective leads to poorly designed tests and unusable data. Being specific – for example, “determine whether panellists can detect a taste difference between formulation A and formulation B” – keeps the entire process focused.

Step 2: Select the appropriate test method 🔗

Once the objective is clear, the next step is choosing the right sensory test method. Sensory test methods broadly fall into three categories: discrimination tests, descriptive tests, and affective (hedonic) tests.

Discrimination tests 🔗

These tests determine whether a detectable difference exists between two or more products. They are commonly used in quality control and ingredient substitution studies. The main types include:

Triangle test: Panellists receive three coded samples – two identical and one different – and must identify the odd one out. It is highly effective for detecting subtle differences. However, it can cause panellist fatigue, so it’s recommended to limit testing sessions to a maximum of six samples.

Duo-trio test: Panellists receive a reference sample and two coded samples (one matching the reference, one different) and must identify which coded sample matches the reference. This method is easier for participants to understand than the triangle test, though it requires a minimum of 10 panellists for sufficient data.

Paired comparison test: Two samples are presented, and panellists indicate which has more of a specific attribute (e.g., sweetness or saltiness). This approach works well when the attribute under investigation is already known.

Descriptive tests 🔗

Descriptive analysis provides a detailed sensory profile of a product by identifying specific attributes and quantifying their intensities. Trained panels typically consisting of 6 to 15 well-trained panellists carry out these evaluations. Key methods include Quantitative Descriptive Analysis (QDA), Flavour Profile Analysis, and Texture Profile Analysis. These are particularly valuable for product development, competitive benchmarking, and understanding exactly how products differ from each other.

Affective (hedonic) tests 🔗

These tests measure consumer liking and preference. They typically involve untrained consumers – often 75 to 150 participants – who rate products on scales such as the nine-point hedonic scale, ranging from “dislike extremely” to “like extremely.” Affective tests answer the fundamental business question: will consumers enjoy this product?

Step 3: Select suitable panellists 🔗

The choice of panellists depends entirely on the type of test being conducted. Analytical tests (discrimination and descriptive) require panellists with demonstrated sensory acuity, while affective tests need consumers who represent the target market.

Panellist types 🔗

Trained panellists undergo rigorous preparation to recognise and assess specific sensory attributes with precision. According to a comprehensive review on panellist selection, training involves honing their abilities to identify and describe attributes following standardised procedures. Trained panels are ideal for discrimination and descriptive tests where subtle distinctions matter.

Expert panellists possess specialised knowledge and extensive experience in a particular field, such as wine or coffee tasting. They can make consistent assessments across a range of products and are typically involved in grading or quality certification tasks.

Consumer panellists are untrained individuals representing the target market. They are selected based on demographic factors like age, gender, dietary habits, and product usage patterns. Their role is to provide feedback on liking and preference rather than analytical precision.

Screening for sensory acuity 🔗

Before selecting panellists for analytical tests, candidates must be screened for their sensory abilities. Common screening procedures include colour vision tests (such as the Ishihara test), basic taste identification tests (sweet, sour, salty, bitter, umami), odour recognition tests, and texture discrimination exercises. The DLG practice guide for sensory panel training notes that the screening methods should be chosen based on the project objectives and the specific sensory methods the panel will use.

Factors like health status, allergies, smoking habits, and willingness to participate on a regular schedule are also important selection criteria. Panellists with conditions that could affect sensory perception – such as a chronic cold or specific food allergies – should be excluded from relevant evaluations.

Step 4: Train the panellists 🔗

Training is essential for analytical tests but not required (and should be avoided) for affective tests where spontaneous consumer responses are needed.

For descriptive panels, training involves developing a shared sensory vocabulary so that all panellists use the same terms consistently. Panellists learn to use reference standards – physical samples that represent specific attributes at defined intensity levels. For instance, a sucrose solution at a certain concentration might serve as the reference for “sweet,” while a caffeine solution represents “bitter.”

Calibration sessions follow, where panellists evaluate a set of products and discuss their ratings as a group. The panel leader facilitates these discussions to ensure consistency across assessors. According to ISO 8589 guidelines, regular monitoring of panellist performance is crucial – panellists who consistently underperform or show bias may need additional training or replacement.

For discrimination panels, training is simpler. Panellists practise the specific test method (e.g., triangle test) with known samples, learning the correct tasting procedure: rinsing the mouth between samples, taking small bites, and following randomised sample order.

Step 5: Prepare the samples 🔗

Sample preparation directly affects the validity of your results. Consistency is non-negotiable – every panellist must receive samples that are identical in every way except the variable being tested.

Key considerations include:

Temperature: All samples of the same product type should be served at the same temperature. Unprepared samples are generally tasted at room temperature, with exceptions for products like ice cream or hot beverages.

Portion size: Each panellist should receive a standardised amount. For trained panel evaluations of products like meat, samples are often cut into uniform cubes to ensure equal portions.

Coding: Samples must be labelled with random three-digit codes to prevent identification bias. Using sequential numbers or letters that might suggest an order should be avoided.

Presentation order: The sequence in which samples are presented should be randomised for each panellist to control for order effects. This ensures that no sample benefits from being tasted first (or suffers from being tasted last, when palate fatigue sets in).

Neutralisation: Samples should be transferred to neutral, identical containers so that packaging or brand cues cannot influence judgements. The preparation area must be physically separated from the testing room to prevent cooking smells or visual cues from reaching panellists.

Step 6: Set up the test environment 🔗

The testing environment has a significant impact on data quality. External factors – ambient odours, inconsistent lighting, noise, uncomfortable temperatures – can distract panellists and skew their assessments.

Facility layout 🔗

An ideal sensory evaluation facility includes three separate areas: a testing room, a sample preparation area, and an administrative office. This separation, recommended by the ISO 8589 standard for test room design, prevents food preparation odours and administrative distractions from influencing panellists.

Individual testing booths are standard in professional sensory labs. These booths are neutral-coloured and odour-free, with dividers high enough to prevent visual and auditory distractions between panellists. Each booth connects to the preparation area through a pass-through hatch, allowing samples to be served without direct contact between panellists and test administrators.

Environmental controls 🔗

Lighting: Uniform illumination is essential. Standard white lighting at around 70-80 foot-candles suits most evaluations. When colour differences between samples need to be masked – for example, when comparing products made with different cooking methods – red or green filtered lights can be used, though these should be applied cautiously with consumer panels.

Temperature and humidity: Room temperature should be maintained at 20-22°C with relative humidity of 45-55%. These conditions keep panellists comfortable and prevent samples from drying out or changing temperature during evaluation.

Ventilation: Maintaining slight positive air pressure in the testing room prevents cooking odours from the preparation area from entering. Professional labs use air exchange rates of up to six times per hour.

Noise: The testing room should be quiet and free from distracting sounds. If the facility cannot be located away from production areas, sound insulation is necessary.

Step 7: Conduct the test and collect data 🔗

With panellists trained, samples prepared, and the environment controlled, the actual test can begin. Panellists should receive clear written instructions explaining the procedure, the evaluation form, and what is expected of them.

Palate cleansers – typically room-temperature water and plain unsalted crackers – should be available between samples. Panellists should also be instructed to avoid eating, smoking, or chewing gum for at least one hour before the test to ensure their palates are neutral.

Each panellist evaluates samples independently. Communication between panellists during the evaluation is not permitted, as it can introduce bias. All responses are recorded on standardised evaluation forms or digital data collection systems. The American Meat Science Association’s research guidelines emphasise that trained panellists should never be asked for their personal preferences – their role is analytical, not affective.

Step 8: Analyse and interpret results 🔗

The data collected must be analysed using appropriate statistical methods. For discrimination tests, significance is typically determined using binomial or chi-square statistics – essentially calculating whether panellists identified the correct sample more often than chance alone would predict.

For descriptive analysis, tools like analysis of variance (ANOVA) and principal component analysis (PCA) help identify which attributes differ significantly between products and how products cluster based on their sensory profiles.

For affective tests, mean liking scores and frequency distributions reveal how well a product was received. Statistical comparisons between products determine whether observed preference differences are meaningful.

The key is matching the statistical method to the test design. Poorly analysed data can lead to incorrect conclusions – either missing a real difference or finding one where none exists.

Best practices to keep in mind 🔗

Beyond following the step-by-step process, several overarching best practices help ensure your sensory tests produce dependable results:

Always match the test method to the objective. A triangle test is not suitable for measuring consumer preference, and a hedonic test cannot tell you whether two formulations are detectably different.

Never mix trained and consumer panel roles. Trained panellists develop analytical skills that may not reflect typical consumer behaviour. Their preferences should not be treated as representative of the market.

Limit sample numbers per session. Sensory fatigue reduces the reliability of responses. For discrimination tests, a maximum of six sample sets per session is a widely accepted guideline.

Regularly calibrate trained panels. Panel performance should be monitored over time using statistical measures of consistency and discrimination ability. Periodic recalibration sessions help maintain data quality.

Document everything. From sample preparation procedures to environmental conditions to panellist attendance – thorough documentation supports reproducibility and helps troubleshoot unexpected results.

What do you think? How important is the testing environment in your experience with sensory evaluation – and have you seen cases where poor environmental control led to unreliable results?

References

1. https://www.labmanager.com/sensory-evaluation-methods-in-food-and-beverage-research-34289
2. https://www.flavoractiv.com/sensory-evaluation-of-food-quick-start-guide/
3. https://flavorsum.com/sensory-analysis-guidelines-food-and-beverage/
4. https://www.foodresearchlab.com/blog/rte-rtc/sensory-evaluation-methods-for-food-and-beverage-products/
5. https://blog.ansi.org/ansi/iso-8586-2023-training-of-sensory-assessors/
6. https://auctoresonline.org/article/selection-and-performance-of-sensory-panelists-a-comprehensive-review-of-factors-influencing-sensory-evaluation-outcomes
7. https://www.dlg.org/en/mediacenter/dlg-expert-reports/food-sensory-technology/dlg-expert-report-07-2017-practice-guide-for-sensory-panel-training-part-1
8. https://foodsafety.institute/food-fundamentals-chemistry/ideal-sensory-evaluation-environment/
9. https://www.puratos.com/blog/how-to-conduct-a-sensory-analysis-test-the-golden-rules
10. https://meatscience.org/docs/default-source/publications-resources/research-guide/2015-amsa-sensory-guidelines-1-0.pdf?sfvrsn=42d380b3_6