Exam Focus
- Explain how detection thresholds, specific sensitivities, adaptation, fatigue, and experience produce legitimate judge variation.
- Use reference standards and shared vocabulary to calibrate without claiming that training erases biological differences.
- Separate a panelist's non-detection from proof that an attribute is absent from the sample.
Sources of legitimate variation
| Source | Lasting or temporary | Practical response |
|---|---|---|
| Detection threshold | Lasting, differs by person and compound | Learn your own thresholds for common faults |
| Specific anosmia | Lasting | Ask a panel partner to confirm or rule out the compound |
| Adaptation | Temporary, during continued exposure | Take short sniffs, pause, and come back |
| Palate fatigue | Temporary, late in a long or intense flight | Order the flight sensibly and take breaks |
| Illness or medication | Temporary | Tell the judge director if it affects your judging |
| Expectation bias | Situational | Keep judging blind and form your impression first |
| Experience | Develops over time | Improves naming and rating, not the underlying biology |
What calibration can and cannot do
Calibration aligns the language and the scale. Judges learn what "low" and "moderate" mean in practice and how a beer of a given quality scores. Useful tools include spiked samples of common faults, blind tasting of known references, and comparing completed scoresheets with experienced judges. A competition may also serve a calibration beer so that a panel can narrow its scoring range before starting. That beer is not a standard for the entries, which are still judged against the guidelines. Training helps a judge recognize a compound more reliably, but it does not give a judge who cannot perceive it the ability to do so.
Non-detection is not proof of absence
"I don't get any diacetyl" means only that this judge did not detect it in this sample at this moment. It might be below their threshold, they may have adapted to it, or it might not be there. A strong claim from one judge is also not automatically correct. Both deserve another look: state where and how strongly the character appears, re-taste, and check whether the beer has changed as it warmed. Judges may also be using different words for one perception, so compare descriptions before deciding that someone missed something.
Handling a split panel
One judge reports moderate DMS in a Munich Helles and the other finds none. Rank does not settle it, and neither does averaging. The first judge says where it appears: cooked-corn aroma that grows as the beer warms. The second re-smells after swirling and still finds nothing. The panel can then record the finding on one sheet, weigh how much it would cost at that intensity, and agree on a consensus score within range. Each scoresheet still records that judge's own perception, because a judge should not write down a character they did not perceive.
Why It Matters
- Treating sensitivity differences as incompetence can silence valid observations and distort consensus.
- Calibration makes intensity and descriptor use more comparable while preserving honest uncertainty.
Practical Sensory Exercises
Chart a threshold disagreement
Have two tasters rate the same aroma reference independently at three dilutions. Compare detection and intensity notes, then write a consensus statement that reports the difference without deciding one taster is wrong.
Key Terms
- Detection threshold
- The lowest concentration at which a person can tell that a compound is present.
- Specific anosmia
- A lasting inability to smell one compound or a narrow family of compounds.
- Calibration
- Aligning how judges use descriptors and score ranges.
- Reference standard
- A known sample, such as a spiked beer, used to train recognition of a character.
- Expectation bias
- Perceiving what one expects because of outside information.