1 Purpose
This document sets out the assessment method for Core Taste Assessment Method for Good Taste Awards.
2 Scope
This document applies to the requirements, decisions, records, controls, and outputs expressly identified under Core Taste Assessment Method. Award scope and exclusions are defined by GTA-STD-100; this document shall not create an additional award dimension. Operational prerequisites identified here affect assessment validity and never constitute product certification.
3 Authority and interpretation
The Method Authority owns GTA-METHOD-001; the Standards Authority approves it. formal verbs, defined terms, stable requirement identities, policy constants, and generated representations shall be interpreted under GTA-GOV-002, GTA-GOV-005, and the defined registers.
3.1 Registered relationships
| Relationship | Direction | Related document |
|---|---|---|
records_original_basis_for | inbound | GTA-STD-DEV-002 |
implemented_by | outbound | GTA-STD-200 |
verifies_method_of | inbound | GTA-VALID-001 |
4 Requirements
GTA-METHOD-001-R001 — Taste-only construct
The Method Authority shall evaluate sensory pleasure and no excluded non-sensory construct.
GTA-METHOD-001-R002 — Target population
The Method Authority shall define the intended potential-consumer group before assessor assignment and limit interpretation to that group.
GTA-METHOD-001-R003 — Overall Liking
The Method Authority shall use the locked defined Overall Liking response as the sole primary candidate input.
GTA-METHOD-001-R004 — Explanatory evidence
The Method Authority shall use aroma, flavour, texture, balance, integration, finish, strengths, limitations, and defect observations only to explain and confirm evidence.
GTA-METHOD-001-R005 — Blind independence
The Method Authority shall exclude brand, package, price, result expectation, and group opinion before response lock.
GTA-METHOD-001-R006 — Staged panel
The Method Authority shall begin at the defined initial panel and add assessors only through confirmed progression routes.
GTA-METHOD-001-R007 — Panel composition
The Method Authority shall maintain a strict consumer-side majority and the defined professional minimum at every decision point.
GTA-METHOD-001-R008 — Deterministic candidate
The Method Authority shall calculate the candidate from locked valid responses and defined constants without discretionary weighting.
GTA-METHOD-001-R009 — Bounded confirmation
The Method Authority shall require professional confirmation within defined evidence grounds and prohibit promotion above the calculated candidate.
GTA-METHOD-001-R010 — Confirmed-level preservation
The Method Authority shall preserve a confirmed lower star while a product is assessed only for a higher star unless a material validity ground is established.
GTA-METHOD-001-R011 — Comment entitlement
The Method Authority shall provide a written product comment after every completed first assessment regardless of star outcome.
GTA-METHOD-001-R012 — Claims boundary
The Method Authority shall classify the ordinary result as individual product non-comparative recognition and prohibit population inference, cross-product preference inference, ranking, or universal-superiority interpretation.
5 Method architecture
The assessed construct is integrated sensory pleasure. Each assessor shall provide an independent Overall Liking response before discussion. Aroma, flavour, texture, balance, integration, finish, strengths, limitations, and suspected defects may explain the response but shall not be weighted into a second score.
The intended audience shall be defined before panel assignment. The panel is a defined small-panel assessment and shall not be described as a population-representative survey. The consumer-side membership shall remain a strict majority at every decision point. Professional assessors contribute individual liking responses on the same basis as other assessors and separately perform only the bounded confirmation duties assigned by the method.
| Decision stage | Cumulative valid responses | Minimum consumer-side assessors | Minimum professional assessors |
|---|---|---|---|
| Initial | 8 | 5 | 2 |
| First progressed | 12 | 7 | 3 |
| Extended | 16 | 9 | 4 |
| Final | 24 | 13 | 6 |
Assessment shall be blind, anonymous in presentation, randomized, and ordered through a balanced design whenever the product count exceeds the capacity of one complete block. Each product shall be judged through a sequential independent monadic response even where multiple products are assigned to one session. Brand, price, identity-bearing packaging, commercial reputation, service, and venue shall be unavailable as assessment inputs.
Consumer-side orientation shall be limited to task comprehension, scale use, confidentiality, blind discipline, response independence, and procedural conduct. It shall not teach a preferred flavour profile, calibrate personal liking to a reference answer, or convert potential consumers and food enthusiasts into a conformity-trained descriptive panel.
6 Candidate, confirmation, and authorization
The decision engine shall calculate a candidate state only from locked valid Overall Liking responses and the policy constants. Professional confirmation shall not recalculate the candidate, negotiate assessor scores, or substitute professional preference for consumer-side evidence.
Confirmation may confirm the candidate, confirm a fully supported lower level for a specified evidence-based reason, declare the affected assessment invalid, or authorize procedural reassessment. Every professional assessor assigned confirmation authority at the decision point shall record one individual disposition after response lock and supporting records discussion. Where no disposition has a strict majority, including a two-professional split, the state shall enter independent professional review; it shall not be treated as a failed candidate and shall not cause statistical panel expansion.
Once a lower star has been confirmed, adding assessors solely to test a higher star shall not remove that confirmed level. Removal is permitted only for an invalid assessment, product-identity failure, non-comparable sample batch, or material procedure failure.
7 Public meaning
Every product completing a completed first assessment is entitled to a written product comment. The publication index is a explanatory display transformation and shall never affect a candidate, progression, confirmation, review, or result. Open public feedback may supplement public interpretation but shall never enter the star decision.
The ordinary award result is individual product non-comparative recognition. A ranking exercise, paired-preference exercise, difference test, similarity test, descriptive profile, or cross-product ordering may be used only in a separately confirmed design for its stated purpose. None may determine, promote, lower, overturn, or retrospectively reinterpret a star candidate. A difference between star levels or publication indexes shall not be treated as a measured preference difference between products.
8 Conformity and records
Conformity with Core Taste Assessment Method requires evidence comprising the source evidence, defined decision, accountable role, and independent or automated verification. The record for each applicable GTA-METHOD-001 requirement shall permit reconstruction without relying on an unrecorded explanation.
Evidence created under GTA-METHOD-001 shall use stable identifiers, immutable event sequence numbers, and machine event timestamps. A correction shall be append-only. A nonconformity shall place the affected object in a non-releasable state until its consequence and confirmed disposition are recorded. Role combination remains subject to the mandatory separation rules.