Sensory And Instrumental Evaluation (Heymann & Ebeler, 2017) — Methods, Corrections And Blending Lab Applications

Source: No access

Source And Review Scope

Hildegarde Heymann and Susan E. Ebeler, Academic Press, 2017. Supplied PDF: 282 sheets; main printed page = PDF sheet minus 15. Entire text extracted; contents and all six chapters/ten case-study topics screened, with selected chapter/case openings and focused close reading. This is complete processing under the agreed approach, not every-page visual or statistical verification.

Findings And Precise Anchors

• Chapter 2, pp. 25–30 / PDF 40–45: analytical difference and consumer acceptance are distinct questions. Chemical/sensory measurements can be related without establishing causation.
• Chapter 3, pp. 35–40 / PDF 50–55: distinguish test chance models and balance presentation. Repeated answers from one participant are not automatically independent panel members. Non-detection does not establish equivalence.
• Chapter 3, pp. 45–55 / PDF 60–70: descriptive analysis needs agreed meanings, references and demonstrated performance; personal free-choice language is valuable but not automatically comparable across people. Preserve raw ratings, training and replication. PCA axes describe variation under a particular scaling choice, not quality or a universal flavor space.
• Chapter 3, pp. 56–60 / PDF 71–75: time–intensity and temporal dominance are distinct. Dominance is attentional selection, not intensity. Simple averaging can distort temporal curves; a personal finish journal should remain a journal.
• Chapter 3, pp. 61–65 / PDF 76–80: liking and preference differ, scale properties matter, sample sizes depend on the study objective, and averages may hide segments. Preserve ties/no-preference and missingness rather than silently reallocating them. Do not impose a fixed sample number as scientific validation.
• Chapter 5, pp. 89–90 / PDF 104–105: pH and titratable acidity are different measurements; report method, calibration, endpoint and units. This reinforces the existing Brown pH note.
• Chapter 5, pp. 100–102 / PDF 115–117: choose a fit-for-purpose method, matrix calibration, controls and a justified application domain. A fitted calibration model is not evidence of future prediction performance.
• Chapter 6, pp. 117–120 / PDF 132–135: compound identity and amount require appropriate evidence; sample preparation and internal standards affect what is measured. SPME depends on partition among liquid, vapor and fiber, as well as time and temperature. Peak area is not automatically concentration or a human smell measurement.
• Chapter 6, pp. 124–126 / PDF 139–141: GC-O and OAVs screen candidates; matrix-appropriate recombination/omission tests strengthen causal claims. Thresholds depend on matrix and sensory method. Statistical correlation is not proof that a molecule causes an attribute.
• Case 8, p. 217 / PDF 232: gin projective mapping demonstrates aroma-based, replicated similarity assessment; its choices are not a whiskey serving protocol.
• Case 9, pp. 225 and 233 / PDF 240 and 248: trained descriptions and consumer liking are separate datasets; average liking conceals groups. The wine study does not validate a whiskey recommendation engine.
• Case 10, p. 241 / PDF 256: closure and storage-temperature effects were studied in wine over six months. Record those variables, but do not transplant outcomes or timescales to whiskey in glass.

Critical Corrections

Printed p. 38 / PDF 53 was visually checked. The tetrad example's 12 correct responses out of 25 is nonsignificant at 0.05; the text incorrectly interprets this as at least 95% probability of no perceptible difference. Correct interpretation: this design did not detect a difference at the chosen significance level. It does not establish equivalence or a posterior probability of sameness. Under independent binomial responses with chance 1/3, the exact upper-tail probability is approximately 0.091793, calculated separately; the book's normal approximation is about 0.0901. Its description of the z approximation as exact is misleading. Do not copy those interpretation statements into learner materials. Page 40's warning about power is more appropriate.

Training-hour examples, recommendations to force choices, covariance-only PCA advice and historical regulatory/brand statements are contextual author choices, not universal Lab requirements. Some case studies use unpublished data; examples are not independent replications of their original datasets. No raw-data reproduction of the ten studies is claimed.

Whole-Book Relevance Map

Chapter 1 perception: conceptual context; 2 uses/good practices: core; 3 sensory techniques: core; 4 chemical testing purposes: method selection; 5 rapid methods: calibration and units; 6 instrumental analysis: core. Case 1 alcohol reduction: discrimination; 2 Chardonnay and 3 mead: trained description; 4 sparkling wines: carbonation and timing; 5 multiple sips: time-intensity and saliva; 6 three wines: TDS; 7 cream liqueurs: sorting; 8 gin: projective mapping; 9 Cabernet: preference mapping; 10 closure/storage: chemical–sensory linkage. These studies support methods learning, not universal whiskey thresholds or outcomes.

Connections And Teaching Use

No access — chemistry explains mechanisms and model assumptions; these sensory methods test what people perceive. Sensory Evaluation Techniques — Methods, Critical Corrections And Academy Applications — extend the existing sensory-methods synthesis rather than replacing it. Whisky dilution changes both headspace chemistry and discrimination — Ashmore et al. — connect dilution and headspace evidence while retaining the study's sample and method limits. No access — proposed teaching and product requirements; not implemented software.

Teach one question, a small coded comparison and a clear conclusion in the Short Version. Use the Full Story for the mechanism, controls, source locations and uncertainty.

Integrated Recommendations

No access The synthesis connects all five books and the revised Lab requirements. Original source text remains in the local Library.