The field of sensory science has grown exponentially since the publication of the first edition of Sensory Evaluation of Food. Fifteen years ago, the journal Food Quality and Preference was fairly new. Now it holds an eminent position as a venue for research on sensory test methods (among many other topics).  Knowledge of the intricate cellular processes in chemoreception, as well as their genetic basis has undergone nothing less than a revolution, culminating in the award of the Nobel Prize to Buck and Axel in 2004 for their discovery of the olfactory receptor gene super family. Advances in statistical methodology have accelerated as well. Sensometrics meetings are now vigorous and well-attended annual events. And yet, some things stay the same. Sensory testing will always involve human participants. But humans are tough measuring instruments to work with. They come with varying degrees of acumen, training, experiences, differing genetic equipment, sensory capabilities, and of course, different preferences. Human foibles and their associated error variance will continue to place a limitation on sensory tests and actionable results.

Although methods continue to evolve, appreciation of the core principles of the field is the key to effective application of sensory test methods. This book has been expanded to reflect the advances in methodologies, theory, and analysis that have transpired in the last 15 years. The chapters are now divided into numbered subsections. This may be of assistance to educators who may wish to assign only certain critical sections to beginning students. In some of the opening sections instructors will find suggestions about which sections are key for fundamental understanding of that topic or method. In many chapters we have gone out on a limb and specified a “recommended procedure.” In cases where there are multiple options for procedure or analysis, we usually chose a simple solution over one that is more complex.

This text attempts to be comprehensive, yet understandable to all students at the university level. All the major sensory test methods are illustrated and discussed, including discrimination, descriptive, and affective tests. Some chapters are devoted to special topics, such as thresholds, time-intensity methods, similarity testing, color, texture, sensory quality control, qualitative research methods, consumer test methods and questionnaires, shelf life testing, an introduction to multivariate statistical techniques, and strategic sensory research. The statistical appendix provides basic instruction in the common statistical analyses for sensory evaluation with worked examples.

 

TABLE OF CONTENTS :

 

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

    1.1 Introduction and Overview . . . . . . . . . . . . . . . . . . . 1
1.1.1 Definition . . . . . . . . . . . . . . . . . . . . . . . 1
1.1.2 Measurement . . . . . . . . . . . . . . . . . . . . . 3

    1.2 Historical Landmarks and the Three Classes
of Test Methods . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2.1 Difference Testing . . . . . . . . . . . . . . . . . . . 5
1.2.2 DescriptiveAnalyses . . . . . . . . . . . . . . . . . 6
1.2.3 AffectiveTesting . . . . . . . . . . . . . . . . . . . 7
1.2.4 The Central Dogma—Analytic Versus
Hedonic Tests . . . . . . . . . . . . . . . . . . . . . 8

    1.3 Applications: Why Collect Sensory Data? . . . . . . . . . . . 10
1.3.1 Differences from Marketing Research Methods . . . 13
1.3.2 Differences from Traditional Product
GradingSystems . . . . . . . . . . . . . . . . . . . 15

    1.4 Summary and Conclusions . . . . . . . . . . . . . . . . . . . 16
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

2 Physiological and Psychological Foundations of Sensory Function . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
    2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

    2.2 Classical Sensory Testing and Psychophysical Methods . . . . 20
2.2.1 Early Psychophysics . . . . . . . . . . . . . . . . . 20
2.2.2 The Classical Psychophysical Methods . . . . . . . . 21
2.2.3 Scaling andMagnitudeEstimation . . . . . . . . . . 23
2.2.4 Critiques of Stevens . . . . . . . . . . . . . . . . . . 25
2.2.5 Empirical Versus Theory-Driven Functions . . . . . 25
2.2.6 Parallels of Psychophysics and Sensory
Evaluation . . . . . . . . . . . . . . . . . . . . . . . 26

    2.3 Anatomy and Physiology and Functions of Taste . . . . . . . . 27
2.3.1 Anatomy andPhysiology . . . . . . . . . . . . . . . 27
2.3.2 Taste Perception: Qualities . . . . . . . . . . . . . . 30
2.3.3 Taste Perception: Adaptation and Mixture
Interactions . . . . . . . . . . . . . . . . . . . . . . 30
2.3.4 Individual Differences and Taste Genetics . . . . . . 33

    2.4 Anatomy and Physiology and Functions of Smell . . . . . . . . 34
2.4.1 Anatomy and Cellular Function . . . . . . . . . . . . 34
2.4.2 Retronasal Smell . . . . . . . . . . . . . . . . . . . 36
2.4.3 Olfactory Sensitivity and Specific Anosmia . . . . . 37
2.4.4 Odor Qualities: Practical Systems . . . . . . . . . . 38
2.4.5 Functional Properties: Adaptation, Mixture
Suppression, and Release . . . . . . . . . . . . . . . 39

    2.5 Chemesthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.5.1 Qualities of Chemesthetic Experience . . . . . . . . 41
2.5.2 Physiological Mechanisms of Chemesthesis . . . . . 42
2.5.3 Chemical “Heat” . . . . . . . . . . . . . . . . . . . 43
2.5.4 Other Irritative Sensations and Chemical
Cooling . . . . . . . . . . . . . . . . . . . . . . . . 44
2.5.5 Astringency . . . . . . . . . . . . . . . . . . . . . . 45
2.5.6 Metallic Taste . . . . . . . . . . . . . . . . . . . . . 46

    2.6 Multi-modal Sensory Interactions . . . . . . . . . . . . . . . . 47
2.6.1 Taste andOdor Interactions . . . . . . . . . . . . . . 47
2.6.2 IrritationandFlavor . . . . . . . . . . . . . . . . . . 49
2.6.3 Color–Flavor Interactions . . . . . . . . . . . . . . . 49
2.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

3 Principles of Good Practice . . . . . . . . . . . . . . . . . . . . . . . 57
    3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

    3.2 The Sensory Testing Environment . . . . . . . . . . . . . . . . 58
3.2.1 EvaluationArea . . . . . . . . . . . . . . . . . . . . 59
3.2.2 ClimateControl . . . . . . . . . . . . . . . . . . . . 62

    3.3 Test Protocol Considerations . . . . . . . . . . . . . . . . . . 63
3.3.1 Sample Serving Procedures . . . . . . . . . . . . . . 63
3.3.2 SampleSize . . . . . . . . . . . . . . . . . . . . . . 63
3.3.3 SampleServingTemperatures . . . . . . . . . . . . 64
3.3.4 ServingContainers . . . . . . . . . . . . . . . . . . 64
3.3.5 Carriers . . . . . . . . . . . . . . . . . . . . . . . . 65
3.3.6 Palate Cleansing . . . . . . . . . . . . . . . . . . . . 65
3.3.7 Swallowing and Expectoration . . . . . . . . . . . . 66
3.3.8 Instructions toPanelists . . . . . . . . . . . . . . . . 66
3.3.9 Randomization and Blind Labeling . . . . . . . . . . 66

    3.4 Experimental Design . . . . . . . . . . . . . . . . . . . . . . . 66
3.4.1 Designing aStudy . . . . . . . . . . . . . . . . . . . 66
3.4.2 Design andTreatmentStructures . . . . . . . . . . . 69

    3.5 PanelistConsiderations . . . . . . . . . . . . . . . . . . . . . 72
3.5.1 Incentives . . . . . . . . . . . . . . . . . . . . . . . 72
3.5.2 Use of Human Subjects . . . . . . . . . . . . . . . . 73
3.5.3 PanelistRecruitment . . . . . . . . . . . . . . . . . 74
3.5.4 Panelist Selection and Screening . . . . . . . . . . . 74
3.5.5 Training ofPanelists . . . . . . . . . . . . . . . . . 75
3.5.6 Panelist Performance Assessment . . . . . . . . . . . 75

    3.6 Tabulation andAnalysis . . . . . . . . . . . . . . . . . . . . . 75
3.6.1 DataEntrySystems . . . . . . . . . . . . . . . . . . 75

    3.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

4 Discrimination Testing . . . . . . . . . . . . . . . . . . . . . . . . . 79
    4.1 DiscriminationTesting . . . . . . . . . . . . . . . . . . . . . . 79

    4.2 Types of Discrimination Tests . . . . . . . . . . . . . . . . . . 80
4.2.1 PairedComparisonTests . . . . . . . . . . . . . . . 80
4.2.2 Triangle Tests . . . . . . . . . . . . . . . . . . . . . 83
4.2.3 Duo–Trio Tests . . . . . . . . . . . . . . . . . . . . 84
4.2.4 n-Alternative Forced Choice (n-AFC) Methods . . . 85
4.2.5 A-Not-Atests . . . . . . . . . . . . . . . . . . . . . 85
4.2.6 Sorting Methods . . . . . . . . . . . . . . . . . . . . 87
4.2.7 TheABXDiscriminationTask . . . . . . . . . . . . 88
4.2.8 Dual-Standard Test . . . . . . . . . . . . . . . . . . 88

    4.3 Reputed Strengths and Weaknesses of Discrimination
Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

    4.4 DataAnalyses . . . . . . . . . . . . . . . . . . . . . . . . . . 89
4.4.1 BinomialDistributions andTables . . . . . . . . . . 89
4.4.2 The Adjusted Chi-Square (χ2)Test . . . . . . . . . . 90
4.4.3 The Normal Distribution and the Z-Test
on Proportion . . . . . . . . . . . . . . . . . . . . . 90

    4.5 Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
4.5.1 ThePower of theStatisticalTest . . . . . . . . . . . 92
4.5.2 Replications . . . . . . . . . . . . . . . . . . . . . . 94
4.5.3 Warm-UpEffects . . . . . . . . . . . . . . . . . . . 97
4.5.4 Common Mistakes Made in the Interpretation
ofDiscriminationTests . . . . . . . . . . . . . . . . 97

Appendix: A Simple Approach to Handling the A, Not-A,
andSame/DifferentTests . . . . . . . . . . . . . . . . . . . . 98

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

5 Similarity, Equivalence Testing, and Discrimination Theory . . . . 101
    5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

    5.2 Common Sense Approaches to Equivalence . . . . . . . . . . 103
    5.3 EstimationofSampleSize andTestPower . . . . . . . . . . . 104
    5.4 How Big of a Difference Is Important?
Discriminator Theory . . . . . . . . . . . . . . . . . . . . . . 105
    5.5 Tests for Significant Similarity . . . . . . . . . . . . . . . . . 108
    5.6 The Two One-Sided Test Approach (TOST)
and IntervalTesting . . . . . . . . . . . . . . . . . . . . . . . 110
    5.7 Claim Substantiation . . . . . . . . . . . . . . . . . . . . . . . 111
    5.8 Models for Discrimination: Signal Detection Theory . . . . . . 111
5.8.1 TheProblem. . . . . . . . . . . . . . . . . . . . . . 112
5.8.2 Experimental Setup . . . . . . . . . . . . . . . . . . 112
5.8.3 Assumptions and Theory . . . . . . . . . . . . . . . 113
5.8.4 AnExample . . . . . . . . . . . . . . . . . . . . . . 114
5.8.5 A Connection to Paired Comparisons Results
Through the ROC Curve . . . . . . . . . . . . . . . 116
    5.9 ThurstonianScaling . . . . . . . . . . . . . . . . . . . . . . . 116
5.9.1 The Theory and Formulae . . . . . . . . . . . . . . . 116
5.9.2 Extending Thurstone’s Model to Other
ChoiceTests . . . . . . . . . . . . . . . . . . . . . . 118

    5.10 Extensions of the Thurstonian Methods, R-Index . . . . . . . . 119
5.10.1 Short Cut Signal Detection Methods . . . . . . . . . 119
5.10.2 AnExample . . . . . . . . . . . . . . . . . . . . . . 120
5.11 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . 120

Appendix: Non-Central t-Test for Equivalence of Scaled Data . . . . . 122
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

6 Measurement of Sensory Thresholds . . . . . . . . . . . . . . . . . 125

    6.1 Introduction: The Threshold Concept . . . . . . . . . . . . . . 125

    6.2 Types of Thresholds: Definitions . . . . . . . . . . . . . . . . 127

    6.3 Practical Methods: Ascending Forced Choice . . . . . . . . . . 128

    6.4 Suggested Method for Taste/Odor/Flavor DetectionThresholds . . . . . . . 129
6.4.1 Ascending Forced-Choice Method of Limits . . . . . 129
6.4.2 Purpose of theTest . . . . . . . . . . . . . . . . . . 129
6.4.3 PreliminarySteps . . . . . . . . . . . . . . . . . . . 130
6.4.4 Procedure . . . . . . . . . . . . . . . . . . . . . . . 131
6.4.5 DataAnalysis . . . . . . . . . . . . . . . . . . . . . 131
6.4.6 Alternative Graphical Solution . . . . . . . . . . . . 131
6.4.7 Procedural Choices . . . . . . . . . . . . . . . . . . 133

    6.5 Case Study/Worked Example . . . . . . . . . . . . . . . . . . 133

    6.6 Other Forced Choice Methods . . . . . . . . . . . . . . . . . . 134

    6.7 ProbitAnalysis . . . . . . . . . . . . . . . . . . . . . . . . . . 136

    6.8 Sensory Adaptation, Sequential Effects, and Variability . . . . 136

    6.9 Alternative Methods: Rated Difference, Adaptive Procedures, Scaling . . 137
6.9.1 Rated Difference from Control . . . . . . . . . . . . 137
6.9.2 Adaptive Procedures . . . . . . . . . . . . . . . . . 138
6.9.3 Scaling as an Alternative Measure of Sensitivity . . . 140

6.10 DilutiontoThresholdMeasures . . . . . . . . . . . . . . . . . 140
6.10.1 Odor Units and Gas-Chromatography
Olfactometry (GCO) . . . . . . . . . . . . . . . . . 140
6.10.2 Scoville Units . . . . . . . . . . . . . . . . . . . . . 142

6.11 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
Appendix: MTBE Threshold Data for Worked Example . . . . . . . . 143
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145

7 Scaling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149   

    7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 149

    7.2 Some Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . 151

    7.3 Common Methods of Scaling . . . . . . . . . . . . . . . . . . 152
7.3.1 Category Scales . . . . . . . . . . . . . . . . . . . . 152
7.3.2 LineScaling . . . . . . . . . . . . . . . . . . . . . . 155
7.3.3 MagnitudeEstimation . . . . . . . . . . . . . . . . . 156

    7.4 Recommended Practice and Practical Guidelines . . . . . . . . 158
7.4.1 Rule 1:ProvideSufficientAlternatives . . . . . . . . 159
7.4.2 Rule 2: The Attribute Must Be Understood . . . . . . 159
7.4.3 Rule 3: The Anchor Words Should Make Sense . . . 159
7.4.4 ToCalibrate orNot toCalibrate . . . . . . . . . . . . 159
7.4.5 A Warning: Grading and Scoring are Not Scaling . . 160

    7.5 Variations—Other Scaling Techniques . . . . . . . . . . . . . 160
7.5.1 Cross-Modal Matches and Variations on  MagnitudeEstimation . . . . . . . . . . . . 160
7.5.2 Category–Ratio (Labeled Magnitude) Scales . . . . . 162
7.5.3 Adjustable Rating Techniques: Relative Scaling . . . 164
7.5.4 Ranking . . . . . . . . . . . . . . . . . . . . . . . . 165
7.5.5 Indirect Scales . . . . . . . . . . . . . . . . . . . . . 166

    7.6 Comparing Methods: What is a Good Scale? . . . . . . . . . . 167   

    7.7 Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
7.7.1 “Do People Make Relative Judgments”
Should They See Their Previous Ratings? . . . . . . 168
7.7.2 Should Category Rating Scales Be Assigned
Integer Numbers in Data Tabulation? Are
They Interval Scales? . . . . . . . . . . . . . . . . . 169
7.7.3 Is Magnitude Estimation a Ratio Scale or
Simply a Scale with Ratio Instructions? . . . . . . . 169
7.7.4 What is a “Valid” Scale? . . . . . . . . . . . . . . . 169

    7.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . 170

Appendix 1: Derivation of Thurstonian-Scale Values
for the 9-Point Scale . . . . . . . . . . . . . . . . . . . . . . . 171

Appendix 2: Construction of Labeled Magnitude Scales . . . . . . . . 172
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174

8 Time–Intensity Methods . . . . . . . . . . . . . . . . . . . . . . . . 179

    8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 179

    8.2 ABriefHistory . . . . . . . . . . . . . . . . . . . . . . . . . 180

    8.3 Variations ontheMethod . . . . . . . . . . . . . . . . . . . . 182
8.3.1 Discrete or Discontinuous Sampling . . . . . . . . . 182
8.3.2 “Continuous” Tracking . . . . . . . . . . . . . . . . 183
8.3.3 Temporal Dominance Techniques . . . . . . . . . . . 184   

    8.4 Recommended Procedures . . . . . . . . . . . . . . . . . . . . 185
8.4.1 Steps in Conducting a Time–intensity Study . . . . . 185
8.4.2 Procedures . . . . . . . . . . . . . . . . . . . . . . . 186
8.4.3 Recommended Analysis . . . . . . . . . . . . . . . . 186

    8.5 DataAnalysisOptions . . . . . . . . . . . . . . . . . . . . . . 187
8.5.1 General Approaches . . . . . . . . . . . . . . . . . . 187
8.5.2 Methods to Construct or Describe Average Curves . . 188
8.5.3 Case Study: Simple Geometric Description . . . . . 189
8.5.4 Analysis by Principal Components . . . . . . . . . . 192

    8.6 Examples andApplications . . . . . . . . . . . . . . . . . . . 193
8.6.1 Taste and Flavor Sensation Tracking . . . . . . . . . 193
8.6.2 Trigeminal and Chemical/Tactile Sensations . . . . . 194
8.6.3 Taste and Odor Adaptation . . . . . . . . . . . . . . 194
8.6.4 Texture and Phase Change . . . . . . . . . . . . . . 195
8.6.5 FlavorRelease . . . . . . . . . . . . . . . . . . . . . 195
8.6.6 Temporal Aspects of Hedonics . . . . . . . . . . . . 196

    8.7 Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
    8.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198

9 Context Effects and Biases in Sensory Judgment . . . . . . . . . . . 203

    9.1 Introduction: The Relative Nature of Human Judgment . . . . . 203

    9.2 SimpleContrastEffects . . . . . . . . . . . . . . . . . . . . . 206
9.2.1 A Little Theory: Adaptation Level . . . . . . . . . . 206
9.2.2 IntensityShifts . . . . . . . . . . . . . . . . . . . . 207
9.2.3 Quality Shifts . . . . . . . . . . . . . . . . . . . . . 207
9.2.4 Hedonic Shifts . . . . . . . . . . . . . . . . . . . . . 208
9.2.5 Explanations forContrast . . . . . . . . . . . . . . . 209
    9.3 Range and Frequency Effects . . . . . . . . . . . . . . . . . . 210
9.3.1 A Little More Theory: Parducci’s Range
and Frequency Principles . . . . . . . . . . . . . . . 210
9.3.2 Range Effects . . . . . . . . . . . . . . . . . . . . . 210
9.3.3 Frequency Effects . . . . . . . . . . . . . . . . . . . 211

    9.4 Biases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
9.4.1 Idiosyncratic Scale Usage and Number Bias . . . . . 212
9.4.2 Poulton’sClassifications . . . . . . . . . . . . . . . 213
9.4.3 Response Range Effects . . . . . . . . . . . . . . . . 214
9.4.4 TheCenteringBias . . . . . . . . . . . . . . . . . . 215

    9.5 Response Correlation and Response Restriction . . . . . . . . 216
9.5.1 Response Correlation . . . . . . . . . . . . . . . . . 216
9.5.2 “Dumping” Effects: Inflation Due
to Response Restriction in Profiling . . . . . . . . . 217
9.5.3 Over-Partitioning . . . . . . . . . . . . . . . . . . . 218

    9.6 Classical Psychological Errors and Other Biases . . . . . . . . 218
9.6.1 Errors in Structured Sequences: Anticipation
andHabituation . . . . . . . . . . . . . . . . . . . . 218
9.6.2 TheStimulusError . . . . . . . . . . . . . . . . . . 219
9.6.3 Positional or Order Bias . . . . . . . . . . . . . . . . 219

    9.7 Antidotes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
9.7.1 Avoid orMinimize . . . . . . . . . . . . . . . . . . 219
9.7.2 Randomization and Counterbalancing . . . . . . . . 220
9.7.3 Stabilization andCalibration . . . . . . . . . . . . . 221
9.7.4 Interpretation . . . . . . . . . . . . . . . . . . . . . 222

    9.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223

10 Descriptive Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 227

    10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 227   

    10.2 UsesofDescriptiveAnalyses . . . . . . . . . . . . . . . . . . 228

    10.3 Language and Descriptive Analysis . . . . . . . . . . . . . . . 228

    10.4 Descriptive Analysis Techniques . . . . . . . . . . . . . . . . 231
10.4.1 Flavor Profile R . . . . . . . . . . . . . . . . . . . . 231
10.4.2 Quantitative Descriptive Analysis R . . . . . . . . . 234
10.4.3 Texture Profile R . . . . . . . . . . . . . . . . . . . 237
10.4.4 Sensory Spectrum R . . . . . . . . . . . . . . . . . . 238

    10.5 Generic Descriptive Analysis . . . . . . . . . . . . . . . . . . 240
10.5.1 How to Do Descriptive Analysis in Three Easy Steps . . . . . . . . . . . . . . . . . . . . 240
10.5.2 Studies Comparing Different Conventional
Descriptive Analysis Techniques . . . . . . . . . . . 246

    10.6 Variations ontheTheme . . . . . . . . . . . . . . . . . . . . . 247
10.6.1 Using Attribute Citation Frequencies Instead
of Attribute Intensities . . . . . . . . . . . . . . . . 247
10.6.2 Deviation from Reference Method . . . . . . . . . . 248
10.6.3 IntensityVariationDescriptiveMethod . . . . . . . . 249
10.6.4 Combination of Descriptive Analysis
and Time-Related Intensity Methods . . . . . . . . . 249
10.6.5 FreeChoiceProfiling . . . . . . . . . . . . . . . . . 249
10.6.6 FlashProfiling . . . . . . . . . . . . . . . . . . . . . 252
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253

11 Texture Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . 259

    11.1 Texture Defined . . . . . . . . . . . . . . . . . . . . . . . . . 259

    11.2 Visual, Auditory, and Tactile Texture . . . . . . . . . . . . . . 262
11.2.1 VisualTexture . . . . . . . . . . . . . . . . . . . . . 262
11.2.2 AuditoryTexture . . . . . . . . . . . . . . . . . . . 262
11.2.3 Tactile Texture . . . . . . . . . . . . . . . . . . . . . 264
11.2.4 Tactile Hand Feel . . . . . . . . . . . . . . . . . . . 268

    11.3 Sensory Texture Measurements . . . . . . . . . . . . . . . . . 270
11.3.1 TextureProfileMethod . . . . . . . . . . . . . . . . 270
11.3.2 Other Sensory Texture Evaluation Techniques . . . . 272
11.3.3 Instrumental Texture Measurements
and Sensory Correlations . . . . . . . . . . . . . . . 274
    11.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . 276
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276

12 Color and Appearance . . . . . . . . . . . . . . . . . . . . . . . . . 283

    12.1 Color and Appearance . . . . . . . . . . . . . . . . . . . . . . 283

    12.2 What IsColor? . . . . . . . . . . . . . . . . . . . . . . . . . . 284

    12.3 Vision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
12.3.1 NormalHumanColorVisionVariations . . . . . . . 286
12.3.2 Human Color Blindness . . . . . . . . . . . . . . . . 286

   12.4 Measurement of Appearance and Color Attributes . . . . . . . 286
12.4.1 Appearance . . . . . . . . . . . . . . . . . . . . . . 286
12.4.2 VisualColorMeasurement . . . . . . . . . . . . . . 289

    12.5 Instrumental Color Measurement . . . . . . . . . . . . . . . . 293
12.5.1 MunsellColorSolid . . . . . . . . . . . . . . . . . . 293
12.5.2 MathematicalColorSystems . . . . . . . . . . . . . 294

    12.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . 299
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299

13 Preference Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303   

    13.1 Introduction—Consumer Sensory Evaluation . . . . . . . . . . 303

    13.2 Preference Tests: Overview . . . . . . . . . . . . . . . . . . . 305
13.2.1 TheBasicComparison . . . . . . . . . . . . . . . . 305
13.2.2 Variations . . . . . . . . . . . . . . . . . . . . . . . 305
13.2.3 SomeCautions . . . . . . . . . . . . . . . . . . . . 306

    13.3 Simple Paired Preference Testing . . . . . . . . . . . . . . . . 306
13.3.1 Recommended Procedure . . . . . . . . . . . . . . . 306
13.3.2 StatisticalBasis . . . . . . . . . . . . . . . . . . . . 307
13.3.3 WorkedExample . . . . . . . . . . . . . . . . . . . 308
13.3.4 UsefulStatisticalApproximations . . . . . . . . . . 309
13.3.5 The Special Case of Equivalence Testing . . . . . . . 310

    13.4 Non-forced Preference . . . . . . . . . . . . . . . . . . . . . . 311

    13.5 Replicated Preference Tests . . . . . . . . . . . . . . . . . . . 313

    13.6 Replicated Non-forced Preference . . . . . . . . . . . . . . . . 313

    13.7 Other Related Methods . . . . . . . . . . . . . . . . . . . . . 315
13.7.1 Ranking . . . . . . . . . . . . . . . . . . . . . . . . 315
13.7.2 Analysis ofRankedData . . . . . . . . . . . . . . . 316
13.7.3 Best–WorstScaling . . . . . . . . . . . . . . . . . . 317
13.7.4 Rated Degree of Preference and Other Options . . . . 318

    13.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . 320
Appendix 1: Worked Example of the Ferris k-Visit Repeated
Preference Test Including the No-Preference Option . . . . . . 320
Appendix 2: The “Placebo” Preference Test . . . . . . . . . . . . . . . 321
Appendix 3: Worked Example of Multinomial Approach
to Analyzing Data with the No-Preference Option . . . . . . . 322
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323

14 Acceptance Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . 325

    14.1 Introduction: Scaled Liking Versus Choice . . . . . . . . . . . 325

    14.2 Hedonic Scaling: Quantification of Acceptability . . . . . . . . 326   

    14.3 Recommended Procedure . . . . . . . . . . . . . . . . . . . . 327
14.3.1 Steps . . . . . . . . . . . . . . . . . . . . . . . . . . 327
14.3.2 Analysis . . . . . . . . . . . . . . . . . . . . . . . . 328
14.3.3 Replication . . . . . . . . . . . . . . . . . . . . . . 328

    14.4 Other Acceptance Scales . . . . . . . . . . . . . . . . . . . . . 328
14.4.1 Line Scales . . . . . . . . . . . . . . . . . . . . . . 328
14.4.2 MagnitudeEstimation . . . . . . . . . . . . . . . . . 330
14.4.3 Labeled Magnitude Scales . . . . . . . . . . . . . . 331
14.4.4 Pictorial Scales and Testing with Children . . . . . . 332
14.4.5 Adjustable Scales . . . . . . . . . . . . . . . . . . . 333
   

    14.5 Just-About-Right Scales . . . . . . . . . . . . . . . . . . . . . 334
14.5.1 Description . . . . . . . . . . . . . . . . . . . . . . 334
14.5.2 Limitations . . . . . . . . . . . . . . . . . . . . . . 335
14.5.3 Variations on Relative-to-Ideal Scaling . . . . . . . . 336
14.5.4 Analysis of JARData . . . . . . . . . . . . . . . . . 336
14.5.5 Penalty Analysis or “Mean Drop” . . . . . . . . . . 339
14.5.6 Other Problems and Issues with JAR Scales . . . . . 340

14.6 Behavioral and Context-Related Approaches . . . . . . . . . . 340
14.6.1 Food Action Rating Scale (FACT) . . . . . . . . . . 341
14.6.2 Appropriateness Scales . . . . . . . . . . . . . . . . 341
14.6.3 Acceptor Set Size . . . . . . . . . . . . . . . . . . . 342
14.6.4 Barter Scales . . . . . . . . . . . . . . . . . . . . . 343

14.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344

15 Consumer Field Tests and Questionnaire Design . . . . . . . . . . . 349

15.1 Sensory Testing Versus Concept Testing . . . . . . . . . . . . 349

15.2 Testing Scenarios: Central Location, Home Use . . . . . . . . 351
15.2.1 Purpose of theTests . . . . . . . . . . . . . . . . . . 351
15.2.2 Consumer Models . . . . . . . . . . . . . . . . . . . 352
15.2.3 CentralLocationTests . . . . . . . . . . . . . . . . 353
15.2.4 HomeUseTests (HUT) . . . . . . . . . . . . . . . . 354

15.3 Practical Matters in Conducting Consumer Field Tests . . . . . 355
15.3.1 Tasks andTestDesign . . . . . . . . . . . . . . . . . 355
15.3.2 SampleSize andStratification . . . . . . . . . . . . 355
15.3.3 TestDesigns . . . . . . . . . . . . . . . . . . . . . . 356

15.4 InteractingwithFieldServices . . . . . . . . . . . . . . . . . 358
15.4.1 Choosing Agencies, Communication,
and Test Specifications . . . . . . . . . . . . . . . . 358
15.4.2 Incidence, Cost, and Recruitment . . . . . . . . . . . 359
15.4.3 SomeTips:Do’s andDon’ts . . . . . . . . . . . . . 360
15.4.4 Steps in Testing with Research Suppliers . . . . . . . 360

15.5 Questionnaire Design . . . . . . . . . . . . . . . . . . . . . . 362
15.5.1 Types of Interviews . . . . . . . . . . . . . . . . . . 362
15.5.2 Questionnaire Flow: Order of Questions . . . . . . . 362
15.5.3 Interviewing . . . . . . . . . . . . . . . . . . . . . . 363

15.6 Rules of Thumb for Constructing Questions . . . . . . . . . . 364
15.6.1 General Principles . . . . . . . . . . . . . . . . . . . 364
15.6.2 Brevity . . . . . . . . . . . . . . . . . . . . . . . . . 364
15.6.3 Use Plain Language . . . . . . . . . . . . . . . . . . 364
15.6.4 Accessibility of the Information . . . . . . . . . . . 365
15.6.5 Avoid Vague Questions . . . . . . . . . . . . . . . . 365
15.6.6 Check for Overlap and Completeness . . . . . . . . . 365
15.6.7 Do Not Lead the Respondent . . . . . . . . . . . . . 365
15.6.8 Avoid Ambiguity and Double Questions . . . . . . . 366
15.6.9 Be Careful in Wording: Present Both Alternatives . . 366
15.6.10 Beware ofHalos andHorns . . . . . . . . . . . . . . 366
15.6.11 Pre-test . . . . . . . . . . . . . . . . . . . . . . . . 366

    15.7 Other Useful Questions: Satisfaction, Agreement and Open-Ended Questions . . . . . . . . . . . . . . . . . . . 367
15.7.1 Satisfaction . . . . . . . . . . . . . . . . . . . . . . 367
15.7.2 Likert (Agree–Disagree) Scales . . . . . . . . . . . . 367
15.7.3 Open-Ended Questions . . . . . . . . . . . . . . . . 367

   15.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . 368

Appendix 1: Sample Test Specification Sheet . . . . . . . . . . . . . . 370
Appendix 2: Sample Screening Questionnaire . . . . . . . . . . . . . . 371
Appendix 3: Sample Product Questionnaire . . . . . . . . . . . . . . . 374
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378

16 Qualitative Consumer Research Methods . . . . . . . . . . . . . . . 379

    16.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 380
16.1.1 Resources,Definitions, andObjectives . . . . . . . . 380
16.1.2 Styles of Qualitative Research . . . . . . . . . . . . 380
16.1.3 Other Qualitative Techniques . . . . . . . . . . . . . 382

     16.2 Characteristics of Focus Groups . . . . . . . . . . . . . . . . . 383
16.2.1 Advantages . . . . . . . . . . . . . . . . . . . . . . 383
16.2.2 Key Requirements . . . . . . . . . . . . . . . . . . . 384
16.2.3 Reliability and Validity . . . . . . . . . . . . . . . . 384

    16.3 Using Focus Groups in Sensory Evaluation . . . . . . . . . . . 385

    16.4 Examples,CaseStudies . . . . . . . . . . . . . . . . . . . . . 386
16.4.1 Case Study 1: Qualitative Research Before
Conjoint Measurement in New Product Development 387
16.4.2 Case Study 2: Nutritional and Health Beliefs About Salt 387

    16.5 Conducting Focus Group Studies . . . . . . . . . . . . . . . . 388
16.5.1 AQuickOverview . . . . . . . . . . . . . . . . . . 388
16.5.2 A Key Requirement: Developing Good Questions . . 389
16.5.3 The Discussion Guide and Phases
of theGroup Interview . . . . . . . . . . . . . . . . 390
16.5.4 Participant Requirements, Timing, Recording . . . . 391

    16.6 Issues in Moderating . . . . . . . . . . . . . . . . . . . . . . . 392
16.6.1 Moderating Skills . . . . . . . . . . . . . . . . . . . 392
16.6.2 Basic Principles: Nondirection, Full
Participation, andCoverage of Issues . . . . . . . . . 393
16.6.3 Assistant Moderators and Co-moderators . . . . . . . 394
16.6.4 Debriefing: Avoiding Selective Listening
and Premature Conclusions . . . . . . . . . . . . . . 395

    16.7 Analysis and Reporting . . . . . . . . . . . . . . . . . . . . . 395
16.7.1 General Principles . . . . . . . . . . . . . . . . . . . 395
16.7.2 Suggested Method (“Sorting/Clustering
Approach”), also Called Classical Transcript
Analysis . . . . . . . . . . . . . . . . . . . . . . . . 396
16.7.3 Report Format . . . . . . . . . . . . . . . . . . . . . 397

    16.8 Alternative Procedures and Variations of the Group Interview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 398
16.8.1 Groups of Children, Telephone Interviews,
Internet-Based Groups . . . . . . . . . . . . . . . . 398
16.8.2 Alternatives to Traditional Questioning . . . . . . . . 399

    16.9 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . 400
Appendix: Sample Report Group Report . . . . . . . . . . . . . . . . . 402
Boil-in-bag Pasta Project Followup Groups . . . . . . . . . . . . . . . 402
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 404

17 Quality Control and Shelf-Life (Stability) Testing . . . . . . . . . . 407

    17.1 Introduction: Objectives and Challenges . . . . . . . . . . . . 408

    17.2 A Quick Look at Traditional Quality Control . . . . . . . . . . 409

    17.3 Methods for Sensory QC . . . . . . . . . . . . . . . . . . . . 409
17.3.1 Cuttings:ABadExample . . . . . . . . . . . . . . . 409
17.3.2 In–Out (Pass/Fail)System . . . . . . . . . . . . . . 410
17.3.3 Difference from Control Ratings . . . . . . . . . . . 411
17.3.4 Quality Ratings with Diagnostics . . . . . . . . . . . 412
17.3.5 DescriptiveAnalysis . . . . . . . . . . . . . . . . . 413
17.3.6 A Hybrid Approach: Quality Ratings
with Diagnostics . . . . . . . . . . . . . . . . . . . . 414
17.3.7 The Multiple Standards Difference Test . . . . . . . 414

    17.4 Recommended Procedure: Difference Scoring with
Key Attribute Scales . . . . . . . . . . . . . . . . . . . . . . . 415

    17.5 The Importance of Good Practice . . . . . . . . . . . . . . . . 417

    17.6 Historical Footnote: Expert Judges and Quality Scoring . . . . 419
17.6.1 Standardized Commodities . . . . . . . . . . . . . . 419
17.6.2 Example 1: Dairy Product Judging . . . . . . . . . . 419
17.6.3 Example 2: Wine Scoring . . . . . . . . . . . . . . . 420

    17.7 Program Requirements and Program Development . . . . . . . 422
17.7.1 Desired Features of a Sensory QC System . . . . . . 422
17.7.2 Program Development and Management Issues . . . 423
17.7.3 The Problem of Low Incidence . . . . . . . . . . . . 424

    17.8 Shelf-Life Testing . . . . . . . . . . . . . . . . . . . . . . . . 424
17.8.1 BasicConsiderations . . . . . . . . . . . . . . . . . 424
17.8.2 CutoffPoint . . . . . . . . . . . . . . . . . . . . . . 426
17.8.3 TestDesigns . . . . . . . . . . . . . . . . . . . . . . 426
17.8.4 Survival Analysis and Hazard Functions . . . . . . . 427
17.8.5 Accelerated Storage . . . . . . . . . . . . . . . . . . 428

    17.9 Summary and Conclusions . . . . . . . . . . . . . . . . . . . 428
Appendix 1: Sample Screening Tests for Sensory Quality Judges . . . . 429
Appendix 2: Survival/Failure Estimates from a Series
of Batches with Known Failure Times . . . . . . . . . . . . . . 429
Appendix 3: Arrhenius Equation and Q10Modeling . . . . . . . . . . . 430
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 431

18 Data Relationships and Multivariate Applications . . . . . . . . . . 433

18.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 433

18.2 Overview of Multivariate Statistical Techniques . . . . . . . . 434
18.2.1 Principal Component Analysis . . . . . . . . . . . . 434
18.2.2 Multivariate Analysis of Variance . . . . . . . . . . . 437
18.2.3 Discriminant Analysis (Also Known
as Canonical Variate Analysis) . . . . . . . . . . . . 438
18.2.4 Generalized Procrustes Analysis . . . . . . . . . . . 439

    18.3 Relating Consumer and Descriptive Data Through
Preference Mapping . . . . . . . . . . . . . . . . . . . . . . . 440
18.3.1 Internal Preference Mapping . . . . . . . . . . . . . 442
18.3.2 External Preference Mapping . . . . . . . . . . . . . 442

    18.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . 445
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 446

19 Strategic Research . . . . . . . . . . . . . . . . . . . . . . . . . . . 451

    19.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 451
19.1.1 Avenues for Strategic Research . . . . . . . . . . . . 451
19.1.2 ConsumerContact . . . . . . . . . . . . . . . . . . . 453

    19.2 Competitive Surveillance . . . . . . . . . . . . . . . . . . . . 453
19.2.1 TheCategoryReview . . . . . . . . . . . . . . . . . 453
19.2.2 Perceptual Mapping . . . . . . . . . . . . . . . . . . 455
19.2.3 Multivariate Methods: PCA . . . . . . . . . . . . . . 456
19.2.4 Multi-dimensional Scaling . . . . . . . . . . . . . . 458
19.2.5 Cost-Efficient Methods for Data Collection:
Sorting . . . . . . . . . . . . . . . . . . . . . . . . . 459
19.2.6 VectorProjection . . . . . . . . . . . . . . . . . . . 460
19.2.7 Cost-Efficient Methods for Data Collection:
Projective Mapping, aka Napping . . . . . . . . . . . 461

    19.3 Attribute Identification andClassification . . . . . . . . . . . . 462
19.3.1 Drivers ofLiking . . . . . . . . . . . . . . . . . . . 462
19.3.2 The Kano Model . . . . . . . . . . . . . . . . . . . 463

   19.4 Preference Mapping Revisited . . . . . . . . . . . . . . . . . . 464
19.4.1 Types of Preference Maps . . . . . . . . . . . . . . . 464
19.4.2 Preference Models: Vectors Versus Ideal Points . . . 464

    19.5 ConsumerSegmentation . . . . . . . . . . . . . . . . . . . . . 465

    19.6 Claim Substantiation Revisited . . . . . . . . . . . . . . . . . 467

    19.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . 468
19.7.1 Blind Testing, New Coke, and the Vienna
Philharmonic . . . . . . . . . . . . . . . . . . . . . 468
19.7.2 The Sensory Contribution . . . . . . . . . . . . . . . 469
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 469

Appendix A Basic Statistical Concepts for Sensory Evaluation . . . . . 473
A.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 473
A.2 Basic Statistical Concepts . . . . . . . . . . . . . . . . . . . . 474
A.2.1 DataDescription . . . . . . . . . . . . . . . . . . . 475
A.2.2 Population Statistics . . . . . . . . . . . . . . . . . . 476
A.3 Hypothesis Testing and Statistical Inference . . . . . . . . . . 478
A.3.1 The Confidence Interval . . . . . . . . . . . . . . . . 478
A.3.2 Hypothesis Testing . . . . . . . . . . . . . . . . . . 478
A.3.3 AWorkedExample . . . . . . . . . . . . . . . . . . 479
A.3.4 A Few More Important Concepts . . . . . . . . . . . 480
A.3.5 DecisionErrors . . . . . . . . . . . . . . . . . . . . 482
A.4 Variations of the t-Test . . . . . . . . . . . . . . . . . . . . . . 482
A.4.1 The Sensitivity of the Dependent t-Test for
Sensory Data . . . . . . . . . . . . . . . . . . . . . 484
A.5 Summary: Statistical Hypothesis Testing . . . . . . . . . . . . 485
A.6 Postscript: What p-Values Signify and What They Do Not . . . 485
A.7 StatisticalGlossary . . . . . . . . . . . . . . . . . . . . . . . . 486
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 487

Appendix B Nonparametric and Binomial-Based Statistical
Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 489
B.1 Introduction to Nonparametric Tests . . . . . . . . . . . . . . 489
B.2 Binomial-Based Tests on Proportions . . . . . . . . . . . . . . 490
B.3 Chi-Square . . . . . . . . . . . . . . . . . . . . . . . . . . . . 493
B.3.1 A Measure of Relatedness of Two Variables . . . . . 493
B.3.2 Calculations . . . . . . . . . . . . . . . . . . . . . . 494
B.3.3 RelatedSamples:TheMcNemarTest . . . . . . . . . 494
B.3.4 TheStuart–MaxwellTest . . . . . . . . . . . . . . . 495
B.3.5 Beta-Binomial, Chance-Corrected
Beta-Binomial, and Dirichlet
MultinomialAnalyses . . . . . . . . . . . . . . . . . 496
Contents xxi
B.4 UsefulRankOrderTests . . . . . . . . . . . . . . . . . . . . . 499
B.4.1 TheSignTest . . . . . . . . . . . . . . . . . . . . . 499
B.4.2 The Mann–Whitney U-Test . . . . . . . . . . . . . . 500
B.4.3 Ranked Data with More Than Two Samples,
Friedman andKramerTests . . . . . . . . . . . . . . 501
B.4.4 RankOrderCorrelation . . . . . . . . . . . . . . . . 502
B.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . 503
B.6 Postscript . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 503
B.6.1 Proof showing equivalence of binomial
approximation Z-test and χ2 test for
difference of proportions . . . . . . . . . . . . . . . 503
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 504
Appendix C Analysis of Variance . . . . . . . . . . . . . . . . . . . . . . 507
C.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 507
C.1.1 Overview . . . . . . . . . . . . . . . . . . . . . . . 507
C.1.2 Basic Analysis of Variance . . . . . . . . . . . . . . 508
C.1.3 Rationale . . . . . . . . . . . . . . . . . . . . . . . 508
C.1.4 Calculations . . . . . . . . . . . . . . . . . . . . . . 509
C.1.5 AWorkedExample . . . . . . . . . . . . . . . . . . 509
C.2 Analysis of Variance from Complete Block Designs . . . . . . 510
C.2.1 Concepts and Partitioning Panelist Variance
fromError . . . . . . . . . . . . . . . . . . . . . . . 510
C.2.2 The Value of Using Panelists
AsTheirOwnControls . . . . . . . . . . . . . . . . 512
C.3 Planned Comparisons Between Means Following ANOVA . . . 513
C.4 Multiple Factor Analysis of Variance . . . . . . . . . . . . . . 514
C.4.1 AnExample . . . . . . . . . . . . . . . . . . . . . . 514
C.4.2 Concept: A Linear Model . . . . . . . . . . . . . . . 515
C.4.3 A Note About Interactions . . . . . . . . . . . . . . 516
C.5 Panelist by Product by Replicate Designs . . . . . . . . . . . . 516
C.6 Issues and Concerns . . . . . . . . . . . . . . . . . . . . . . . 519
C.6.1 Sensory Panelists: Fixed or Random Effects? . . . . 519
C.6.2 A Note on Blocking . . . . . . . . . . . . . . . . . . 520
C.6.3 Split-Plot or Between-Groups (Nested) Designs . . . 520
C.6.4 Statistical Assumptions and the Repeated
MeasuresANOVA . . . . . . . . . . . . . . . . . . 521
C.6.5 OtherOptions . . . . . . . . . . . . . . . . . . . . . 522
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 522
Appendix D Correlation, Regression, and Measures of Association . . . 525
D.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 525
D.2 Correlation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 527
D.2.1 Pearson’s Correlation Coefficient Example . . . . . . 528
D.2.2 Coefficient ofDetermination . . . . . . . . . . . . . 529
D.3 Linear Regression . . . . . . . . . . . . . . . . . . . . . . . . 529
D.3.1 Analysis of Variance . . . . . . . . . . . . . . . . . 530
D.3.2 Analysis of Variance for Linear Regression . . . . . 530
D.3.3 Prediction of theRegressionLine . . . . . . . . . . . 530
D.3.4 Linear Regression Example . . . . . . . . . . . . . . 531
xxii Contents
D.4 Multiple Linear Regression . . . . . . . . . . . . . . . . . . . 531
D.5 OtherMeasures ofAssociation . . . . . . . . . . . . . . . . . 531
D.5.1 Spearman Rank Correlation . . . . . . . . . . . . . . 531
D.5.2 Spearman Correlation Coefficient Example . . . . . 532
D.5.3 Cramér’s VMeasure . . . . . . . . . . . . . . . . . . 532
D.5.4 CramérCoefficientExample . . . . . . . . . . . . . 533
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 533
Appendix E Statistical Power and Test Sensitivity . . . . . . . . . . . . 535
E.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 535
E.2 FactorsAffecting thePower ofStatisticalTests . . . . . . . . . 537
E.2.1 SampleSize andAlphaLevel . . . . . . . . . . . . . 537
E.2.2 EffectSize . . . . . . . . . . . . . . . . . . . . . . . 538
E.2.3 How Alpha, Beta, Effect Size, and N Interact . . . . 539
E.3 WorkedExamples . . . . . . . . . . . . . . . . . . . . . . . . 541
E.3.1 The t-Test . . . . . . . . . . . . . . . . . . . . . . . 541
E.3.2 An Equivalence Issue with Scaled Data . . . . . . . 542
E.3.3 Sample Size for a Difference Test . . . . . . . . . . . 544
E.4 Power in Simple Difference and Preference Tests . . . . . . . . 545
E.5 Summary and Conclusions . . . . . . . . . . . . . . . . . . . 548
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 549
Appendix F Statistical Tables . . . . . . . . . . . . . . . . . . . . . . . . . 551
Table F.A Cumulative probabilities of the standard normal
distribution. Entry area 1–α under the standard
normal curve from −∞ to z(1–α) . . . . . . . . . . . . . 552
Table F.B Table of critical values for the t-distribution . . . . . . . . 553
Table F.C Table of critical values of the chi-square (χ2)
distribution . . . . . . . . . . . . . . . . . . . . . . . . . 554
Table F.D1 Critical values of the F-distribution at α =0.05 . . . . . . 555
Table F.D2 Critical values of the F-distribution at α =0.01 . . . . . . 556
Table F.E Critical values of U for a one-tailed alpha at 0.025
or a two-tailedalpha at 0.05 . . . . . . . . . . . . . . . . 556
Table F.F1 Table of critical values of ρ (Spearman Rank
correlation coefficient) . . . . . . . . . . . . . . . . . . . 557
Table F.F2 Table of critical values of r (Pearson’s correlation
coefficient) . . . . . . . . . . . . . . . . . . . . . . . . . 558
Table F.G Critical values for Duncan’s multiple range test
(p, df, α =0.05) . . . . . . . . . . . . . . . . . . . . . . . 559
Table F.H1 Critical values of the triangle test for similarity
(maximum number correct as a function of the
number of observations (N), beta, and proportion
discriminating) . . . . . . . . . . . . . . . . . . . . . . . 560
Table F.H2 Critical values of the duo–trio and paired
comparison tests for similarity (maximum number
correct as a function of the number of observations
(N), beta, and proportion discriminating) . . . . . . . . . . 561
Table F.I Table of probabilities for values as small as
observed values of x associated with the binomial
test (p=0.50) . . . . . . . . . . . . . . . . . . . . . . . . 562
Contents xxiii
Table F.J Critical values for the differences between rank
sums (α =0.05) . . . . . . . . . . . . . . . . . . . . . . . 563
TableF.K Criticalvalues of thebeta binomial distribution . . . . . . 564
Table F.L Minimum numbers of correct judgments
to establish significance at probability levels of 5
and 1% for paired difference and duo–trio
tests (one tailed, p = 1/2) and the triangle
test (one tailed, p = 1/3) . . . . . . . . . . . . . . . . . . 565
Table F.M Minimum numbers of correct judgments
to establish significance at probability levels of 5
and 1% for paired preference test (two tailed,
p = 1/2) . . . . . . . . . . . . . . . . . . . . . . . . . . . 566
Table F.N1 Minimum number of responses (n) and correct
responses (x) to obtain a level of Type I
and Type II risks in the triangle test. Pd is
the chance-adjusted percent correct or proportion
of discriminators . . . . . . . . . . . . . . . . . . . . . . 567
Table F.N2 Minimum number of responses (n) and correct
responses (x) to obtain a level of Type I
and Type II risks in the duo–trio test. Pc is the
chance-adjusted percent correct or proportion
of discriminators . . . . . . . . . . . . . . . . . . . . . . 567
Table F.O1 d and B (variance factor) values for the duo–trio
and 2-AFC (paired comparison) difference tests . . . . . . 568
Table F.O2 d and B (variance factor) values for the triangle
and 3-AFC difference tests . . . . . . . . . . . . . . . . . 569
Table F.P Random permutations of nine . . . . . . . . . . . . . . . 571
Table F.Q Random numbers . . . . . . . . . . . . . . . . . . . . . . 572
Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 573
Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 587

 

 

 

Harry T. Lawless is Professor of Food Science at Cornell University where he teaches sensory evaluation. He has 35 years of experience in chemosensory research and psychophysics. He spent five years in consumer testing in industry, and serves as a consultant to various food and consumer products companies on sensory test methods.

Hildegarde Heymann is Professor of Viticulture and Enology at the University of California at Davis where she teaches sensory evaluation of wine and sensometrics. She spent nearly 17 years at the University of Missouri as a professor of sensory science.