12.2 Sensory evaluation methods

Sensory evaluation methods are how food scientists measure the way people perceive food. These methods fall into three main categories: discriminative tests (detecting differences), descriptive analysis (characterizing attributes), and affective tests (measuring consumer preference). Together, they provide the data behind product development, quality control, and reformulation decisions.

Discriminative Tests

Discriminative tests answer a simple question: can people tell the difference? They don't ask panelists to describe what they taste or whether they like it. They just determine whether a perceptible sensory difference exists.

Triangle Test and Paired Comparison

Triangle test: A panelist receives three samples. Two are identical, one is different. The panelist's job is to identify the odd sample. This test is widely used because it doesn't require the panelist to describe how the samples differ, only that they differ. For example, a juice company reformulating its recipe might use a triangle test to see if consumers can detect the change at all.

The probability of guessing correctly by chance is 1 in 3 (about 33%), so statistical analysis accounts for that when determining whether results are significant.

Paired comparison test: A panelist receives two samples and is asked which one has more of a specific attribute, like sweetness or crunchiness. Unlike the triangle test, this one targets a particular sensory characteristic. For instance, comparing two brands of potato chips to determine which one consumers perceive as crunchier.

Ranking Test

Ranking tests present multiple samples at once, and panelists order them by the intensity of a specific attribute or by overall preference. This is useful when you need to compare several products simultaneously rather than just two. A classic example: ranking five salsa brands from mildest to spiciest.

• Panelists assign each sample a rank (1st, 2nd, 3rd, etc.)
• Results show relative differences between products, not absolute intensity values
• Works well for quick screening of multiple formulations or competing products

Descriptive Analysis Methods

Descriptive analysis goes deeper than discriminative testing. Instead of asking "is there a difference?" it asks "what exactly does this product taste, smell, feel, and look like, and how intense is each characteristic?" These methods require trained panelists.

Quantitative Descriptive Analysis (QDA)

QDA is one of the most detailed sensory methods available. A trained panel first develops a shared vocabulary of descriptive attributes covering appearance, aroma, flavor, and texture. Then they rate the intensity of each attribute on a structured scale (typically 0 to 15).

Here's how a QDA study typically works:

1. Orientation: Panelists sample a range of products and discuss their sensory characteristics.
2. Attribute development: The panel agrees on a set of terms (e.g., "buttery aroma," "graininess," "astringency") and defines each one with reference standards.
3. Training: Panelists practice rating intensity using the agreed-upon scale until their scores are consistent and reproducible.
4. Evaluation: Panelists rate each product independently, scoring every attribute.
5. Analysis: Results are compiled into a quantitative sensory profile, often displayed as a spider (radar) chart showing intensity ratings across all attributes.

QDA is commonly used for products like wine, cheese, or coffee where subtle differences in flavor profile matter.

Time-Intensity Method and Free Choice Profiling

Time-Intensity (TI) method: Most sensory ratings capture a single snapshot, but perception changes over time. TI tracks how the intensity of a specific attribute rises, peaks, and fades. A panelist continuously records intensity from the moment they first perceive the attribute until it disappears.

This is especially useful for attributes with a dynamic profile. Think about biting into dark chocolate: the bitterness builds, peaks as the chocolate melts, and then gradually fades. A TI curve captures that entire arc, giving food scientists data on onset time, peak intensity, and how long the sensation lingers.

Free Choice Profiling (FCP): In standard descriptive analysis, the panel agrees on shared terminology. FCP takes the opposite approach. Each panelist creates their own set of descriptive terms. This avoids the time-consuming process of building group consensus and can be useful for complex or unfamiliar products where a common vocabulary is hard to establish (perfumes, novel food products).

The tradeoff is that data analysis becomes more complex. Since each panelist uses different terms, a statistical technique called Generalized Procrustes Analysis (GPA) is used to align the individual data sets and find common patterns across panelists.

Affective Tests

Affective tests shift the focus from trained panels to consumers. The goal isn't to describe a product objectively but to measure how much people like it. These tests use untrained participants, often in large numbers, to get results that reflect real consumer preferences.

Hedonic Scale

The hedonic scale is the most widely used affective test in food science. Panelists rate how much they like or dislike a product on a balanced scale, most commonly a 9-point version:

• 9 = Like extremely
• 5 = Neither like nor dislike
• 1 = Dislike extremely

The scale is symmetrical, with equal numbers of "like" and "dislike" categories on either side of the neutral midpoint. This structure lets researchers calculate meaningful averages and compare products statistically.

Hedonic testing is typically done with large consumer panels (often 50 to 100+ people) to ensure the results represent a broader population. A company developing a new snack bar, for example, might test three flavor formulations with 100 consumers and use the hedonic scores to decide which one to bring to market.