Threshold detection (ISO 13301:2018)

Goal

Determine a panelist's detection threshold for a specific stimulus, following the 3-AFC (3-Alternative Forced Choice) procedure defined by ISO 13301:2018 and ASTM E679-19.

This is the quantitative test that, with ≥3 measurements in the last 6 months, qualifies a panelist at the selected level in the ISO 8586:2023 classification. The expert level is not reached through thresholds alone: it also requires revalidating basic taste recognition within the last 6 months (see Panel training).

What 3-AFC is

At each concentration level:

• 3 samples are presented to the panelist.
• 2 are water (or a neutral reference).
• 1 contains the stimulus at the level's concentration.
• The panelist identifies the different one. The test detects the concentration at which they can no longer tell the stimulus apart from the control.

AreaCacao presents the series in ascending order and detects the threshold when the panelist hits the number of consecutive correct answers set by the criterion. The BET is computed as the geometric mean between the last failed level and the first consecutive hit (ASTM E679-19).

Concentration levels

AreaCacao prepares a series with increasing concentrations. You can configure between 4 and 8 levels (the minimum is 4; more than 8 causes sensory fatigue and is not allowed).

Available compounds

The stimulus can be any of these compounds, each with its standard substance:

The name you'll see in the compound picker is the one in the first column:

Presentation medium

You choose the base the stimulus is presented in, depending on how close you want to stay to the real product:

• Mineral water — the cleanest base, ideal for measuring pure sensitivity.
• Chocolate base (real matrix) — the stimulus in a chocolate matrix, for thresholds that are realistic for your product.
• Neutral base (cocoa butter + sugar) — a middle ground without cocoa's own notes.

Before you start

• Prepare the stimulus solutions per ISO 13301, at each level's concentration and on the chosen medium.
• Small cups, mineral water as a palate cleanser, an odor-free environment.

Steps

How to read the result

The leader sets a maximum BET to pass. The lower the BET, the more sensitive the panelist (they detect the stimulus at lower concentrations):

• BET at or below the maximum → the measurement counts as valid and adds toward selected. The panelist becomes selected once they accumulate ≥3 threshold measurements in the last 6 months; a single test is not enough.
• BET above the maximum → the measurement does not count as valid for this test. It may be due to personal sensitivity, noise in the tasting, or the test itself. Repeat it later.
• Threshold not reached → if the panelist doesn't string together the criterion's consecutive hits within the tested range, no BET is obtained. In that case it's best to extend the series toward higher concentrations and repeat.

Where you see the results

Results are only computed when you close the tasting. While it's active, you'll see the notice "Results are calculated automatically when you close the tasting."

Once the tasting is closed, the detail shows the Calibration results table with one row per panelist and these columns:

Why it matters

• It confirms that your panel has quantified sensitivity, not just "experience."
• It's the piece that qualifies the selected level and, together with basic taste revalidation, the expert level, which lets you sign dossiers backed by ISO 8586 methodology.