A NEW APPROACH TO CATEGORY SCALES OF INTENSITY I: TRADITIONAL VERSUS RANK-RATING

A traditional approach to intensity scaling, using ratings on a 9-point category scale was compared to a method that combined elements of ranking called Rank-Rating. The latter method forced judges to retaste stimuli whose tastes they had forgotten and accordingly reduced discrimination errors, thus increasing ability to discriminate.     

A COMPARISON OF CATEGORY AND LINE SCALES UNDER VARIOUS EXPERIMENTAL PROTOCOLS

Judges were required to rate the total intensity of NaCl solutions and the sweetness of orange flavored beverages using a variety of unstructured category and line scales. Discrimination errors were noted for each type of scale. No strong differences were observed in discrimination between category and line scales although a 20-point category scale incurred fewer discrimination errors than 9-point category and line scales. Allowing judges to retaste stimuli and review and modify their scores reduced discrimination errors.     

INDUCTION OF SCALING ERRORS

Judges were required to rate the total intensity of NaCl solutions using a variety of unstructured category and line scales under a 'rank-rating' protocol and a traditional protocol that did not allow retasting or the reviewing of scores. The various scales and protocols induced two types of scaling errors. The first type was named a different-stimulus error. This involved a judge rating a stronger stimulus as equal to or less than a weaker stimulus. The second type was named a same-stimulus error. This involved a judge giving different ratings to two stimuli of equal concentration. For all scales, judges made a higher proportion of same-stimulus errors than different-stimulus errors. 'Rank-rating' only reduced the proportion of different-stimulus errors. It was hypothesized that a category scale with fewer categories would induce a higher level of different-stimulus errors but lower level of same-stimulus errors. These trends were noted but not always significantly. For line scales, the equivalent hypotheses regarding line length were either weak or not supported. It would appear that increasing the length of a line scale is not always equivalent to increasing the number of categories in a category scale.     

THE DISRIMINATIVE EFFICIENCY OF COMMON SCALING METHODS

Four types of rating scales, nine-point category scales, line marking, magnitude estimation, and a hybrid of the category and line scale, were employed to assess visual, tactile and olfactory characteristics of products by consumers. The scales were compared for their ability to discriminate differences among products, for variability, for reliability, and for ease of use. All methods were able to generate highly significant differences among products. However, a modest advantage for category scales was observed in almost all comparisons.     

COMPARISON OF RATING SCALES: SENSITIVITY, REPLICATES AND RELATIVE MEASUREMENT

The relative sensitivities of four scaling methods were assessed in central location tests with untrained judges. The scales included category scales, line scales, magnitude estimation, and a hybrid of the line and category scales. Approximate parity was observed among category scales, line scales and the hybrid scale in their ability to differentiate small physical differences. Magnitude estimation was used as efficiently as the other methods by a college population, but less efficiently by a heterogeneous sample of consumers. Judges used the scales with greater accuracy as they became familiar with the range of products to be judged. In spite of relatively small physical differences, subjects used wide ranges of the scales, supporting the view that rating scales are relative, not absolute, measuring instruments.     

USE AND ABUSES OF CATEGORY SCALES IN SENSORY MEASUREMENT

Category ratings are a popular device for assessing preferences and for guiding efforts in the development of new products. They are frequently misinterpreted or misused, however, because their use is so closely analogized to the process of fundamental measurement. Unlike physical measures, psychological scales such as category ratings are strongly affected by situational or contextual factors which must be carefully controlled and fully understood for proper use of the scale. Category ratings are affected by stimulus range and frequency, number of stimuli and number of categories, as well as by stimulus sequence. Implications for appropriate use of the scale in sensory evaluation are addressed.     

EVALUATION OF THREE SCALING METHODS FOR HEDONICS

A collaborative study of twenty-three laboratories was conducted to compare the relative effectiveness of three scales: two forms of magnitude estimation scaling and one form of a category scale in the measurement of hedonic response to a controlled stimulus. Responses from 553 individual judges show that all scales yield hedonic measurements that are very similar in both direction and magnitude of difference between the stimuli. No scale showed any clear superiority in reliability, precision, or discrimination. Selection of a scale must be based on considerations other than the simple form of response.     

THURSTONIAN MODELS AND VARIANCE II: EXPERIMENTAL CONFIRMATION OF THE EFFECTS OF VARIANCE ON THURSTONIAN MODELS OF SCALING

Subjects rated taste intensities using category scales under a variety of experimental protocols, which induced differences in computed values of d′. These were explained in a Thurstonian/signal detection context, by variation in the effects of adaptation changing perceived intensities (perceptual variance), by the effects of forgetting (memory variance) and by differences in the idiosyncratic ways that subjects use category‐rating scales (boundary variance).     

STRUCTURED AND UNSTRUCTURED 9-POINT HEDONIC SCALES: A CROSS CULTURAL STUDY WITH AMERICAN, JAPANESE AND KOREAN CONSUMERS

The range of scores elicited by a structured, an unstructured and a 'labels-only'version of the 9-point hedonic scale were compared using consumers from USA, Japan and Korea. It was found that the unstructured scale elicited a wider range of scores for American and Japanese consumers. After correction for hedonic ranges, it was found that Japanese had smaller ranges of scores on all three scales, although the effect was less pronounced for the unstructured scale. The Korean consumers were the exception. Their ranges were less than Americans but their ranges on the unstructured scale did not increase. The results were discussed in terms of the effects of inhibition of use of categories by the scale labels, effects of translation from the English, psychophysical style and order effects.     

POWER AND SENSITIVITY OF THE SAME-DIFFERENT TEST: COMPARISON WITH TRIANGLE AND DUO-TRIO METHODS

Same-different, duo-trio and triangle discrimination methods were compared using vanilla flavored yogurt with and without added sugar as the medium. Two experiments were performed, one in controlled laboratory conditions and the other in conditions approximating more to consumer testing. A modification of the same-different test had greater power than the duo-trio of triangle tests. At higher sugar concentrations, d'values for the three methods were equivalent. Yet, at lower sugar concentrations, the same-different d'tended to be higher. The results are discussed in terms of Sequential Sensitivity Analysis, memory effects and changes between β and criteria.     

EVALUATION OF WINE COMPETITION JUDGE PERFORMANCE USING PRINCIPAL COMPONENT SIMILARITY ANALYSIS

Principal component similarity (PCS) analysis was used to evaluate judge performance from a wine competition. Data were analyzed for five international judges and seven wine makers, for 42 white, 30 red and 25 specialty wines, using a 20-point quality scoring system. Principal similarity plots were used to group judges according to judging 'style' and to identify outliers, for each wine category. Judge groupings were consistent when three different references were used; however, the most interpretable PCS plot was obtained when the overall mean-judge-score was used as the reference. Results from PCS were compared to principal component analysis (PCA). PCS analysis allowed the information from all significant principal components to be graphically represented in two dimensions and was more successful in classifying judges than plots based on the first three principal components. The technique of PCS is an important complement to existing methodologies, and can provide wine competition coordinators with an objective technique for judge evaluation and selection.     

PROP (6-n-Propylthiouracil) tasting and sensory responses to caffeine,sucrose, neohesperidin dihydrochalcone and chocolate

The genetically determined ability to taste 6-n-propylthiouracil (PROP) has been linked with lowered acceptance of some bitter foods. Fifty-four women, aged 18-30 years, tasted and rated PROP-impregnated filter paper and seven solutions of PROP. Summed bitterness intensity ratings for PROP solutions determined PROP taster status. Respondents also tasted five sucrose and seven caffeine solutions, as well as seven solutions each of caffeine and PROP that had been sweetened with 0.3 mmol/l neohesperidin dihydrochalcone (NHDC). Respondents also rated three kinds of chocolate using 9-point category scales. PROP tasters rated caffeine solutions as more bitter than did non-tasters and liked them less. PROP tasters did not rate either sucrose or NHDC as more sweet. The addition of NHDC to PROP and caffeine solutions suppressed bitterness intensity more effectively for tasters than for non-tasters and improved hedonic ratings among both groups. PROP tasters and non-tasters showed the same hedonic response to sweetened caffeine solutions and did not differ in their sensory responses to chocolate. Genetic taste markers may have only a minor impact on the consumption of such foods as sweetened coffee or chocolate.     

Comparison of Different Risk Perception Measures in Predicting Seasonal Influenza Vaccination among Healthy Chinese Adults in Hong Kong: A Prospective Longitudinal Study

Background

Risk perception is a reported predictor of vaccination uptake, but which measures of risk perception best predict influenza vaccination uptake remain unclear.

Methodology

During the main influenza seasons (between January and March) of 2009 (Wave 1) and 2010 (Wave 2),505 Chinese students and employees from a Hong Kong university completed an online survey. Multivariate logistic regression models were conducted to assess how well different risk perceptions measures in Wave 1 predicted vaccination uptake against seasonal influenza in Wave 2.

Principal Findings

The results of the multivariate logistic regression models showed that feeling at risk (β = 0.25, p = 0.021) was the better predictor compared with probability judgment while probability judgment (β = 0.25, p = 0.029 ) was better than beliefs about risk in predicting subsequent influenza vaccination uptake. Beliefs about risk and feeling at risk seemed to predict the same aspect of subsequent vaccination uptake because their associations with vaccination uptake became insignificant when paired into the logistic regression model. Similarly, to compare the four scales for assessing probability judgment in predicting vaccination uptake, the 7-point verbal scale remained a significant and stronger predictor for vaccination uptake when paired with other three scales; the 6-point verbal scale was a significant and stronger predictor when paired with the percentage scale or the 2-point verbal scale; and the percentage scale was a significant and stronger predictor only when paired with the 2-point verbal scale.

Conclusions/Significance

Beliefs about risk and feeling at risk are not well differentiated by Hong Kong Chinese people. Feeling at risk, an affective-cognitive dimension of risk perception predicts subsequent vaccination uptake better than do probability judgments. Among the four scales for assessing risk probability judgment, the 7-point verbal scale offered the best predictive power for subsequent vaccination uptake.     

VARIATION IN SENSORY CONCEPTUAL STRUCTURE: AN INVESTIGATION INVOLVING THE SORTING OF ODOR STIMULI

Judges sorted forty-three odor stimuli into groupings of conceptual similarity, repeating the task over several sessions. There was little between-judges consistency in conceptual structure inferred from the sorting patterns and little within-judges consistency over sessions. The latter suggests that 'one time' sorting experiments would give different results on repetition. In a second experiment, the effect of prior categorization experience on subsequent categorical learning was examined. Naive judges and judges from the first study performed further sortings to see whether they could be induced to sort into a chosen set of perfumery categories. Various cues were given: category names, exemplar stimuli, numbers of stimuli within each category and finally the 'correct' answers. Increasing the cues increased the tendency to sort according to the a priori plan but complete success was not achieved. Judges who had previously sorted the stimuli according to their own conceptual structure, found the a priori perfumery structure harder to learn than naive judges.     

RATIO SCALES AND CATEGORY SCALES OF ODOUR INTENSITY

The relation between a ratio scale obtained by magnitude estimationand a category scale of the odour intensity of 1-butanol wasstudied, together with individual variations in the ratio scale.Series of solutions of butanol in water in small bottles werepresented to a panel for judgement, half using the method ofmagnitude estimation, the other half a category scale. Plotswere made of the category scale against the ratio scale, andthe ratio scales of individual members of the panel were analysed.A power function exponent of 0.48 was found for the panel'sratio scale, with individual values ranging from 0.25 to 0.49.The category scale was curved relative to the ratio scale; variabilityof the magnitude estimates was approximately proportional tothe magnitude estimates; and a small time-order error was found.Odour intensity exhibits the three tested characteristics ofa prothetic continuum, and the variability of individual exponentswas not as great as sometimes suggested.     

EFFECT OF QUESTIONNAIRE DESIGN AND THE NUMBER OF SAMPLES TASTED ON HEDONIC RATINGS1

A total of 450 consumers participated in a test to determine whether questionnaire length, presence of key diagnostic questions or serving position affected their hedonic discrimination among yellow cakes. Consumers evaluated four yellow cakes representing a 2 × 2 factorial design of texture and flavor flaws. They used one of the following six questionnaires: only a 9-point hedonic scale, a 9-point hedonic scale with open end questions, and four others comprising a 2 × 2 factorial design with two levels of questionnaire length and two levels of questionnaire completeness. Neither the presence of key attribute questions nor the length of the questionnaire affected the value or the sensitivity of the judges’ overall liking scores. Samples tasted first received higher hedonic scores than those same samples tasted second throughfifth. Judges could discriminate among the samples on the basis of overall liking best when samples were tasted fourth or fifth.