Search

The food industry is now global, so that hedonic comparisons for various products may need to be made between consumers from different countries. Using the 9-point hedonic scale, it has been shown that East Asians use a significantly smaller range of scores than Americans; they are essentially using different scales, rendering comparisons between the two sets of consumers problematic. There is a necessity for a culture-free method of hedonic numerical assessment, which is not affected by such ‘cultural range variation’. American and Chinese consumers assessed candies, toothbrushes and pens, using the regular verbal 9-point hedonic scale as well as a purely numerical version. As expected from previous studies, the Chinese consumers tended to have a significantly smaller range of responses than the American consumers, although the effect was not as strong for the purely numerical 9-point hedonic scale. This meant that Chinese consumers were often using a shorter scale than the Americans, making comparisons between the two groups problematic. However, a ranking protocol with an R-Index analysis was also used. With this protocol, the range of responses for the Chinese and Americans was the same, being equal to the number stimuli being ranked, circumventing the problem of ‘cultural range variation’. [ABSTRACT FROM AUTHOR]

Background Paired preference tests are one of several types of measurement of consumer acceptance. Yet, the test has had issues regarding its statistical analysis. Initially, test designs were ‘forced choice’, without a ‘No Preference’ option. Accordingly, the data were limited. Later, putatively identical stimuli were used as controls. Consumers reported preferences for these stimuli. It was deemed not logically possible to have genuine preferences for identical stimuli, therefore these responses were assumed to be responses elicited in the ‘no preference’ condition. This provided the control condition for statistical comparison. It also allowed responses of ‘no preference’. Subsequent experimental designs were refinements of this approach. Scope and approach Difficulties with earlier forced choice designs are described and how statistical analysis changed with the introduction of control groups to represent the ‘no preference’ condition. It describes how the measurements were refined by supplementing frequency measures with d′ measures from signal detection theory. The factors causing consumers to report preferences for putatively identical stimuli are discussed. How these have spawned alternative protocols for paired preference tests are described. Key findings and conclusions Forced choice preference tests were adopted so that simple binomial statistics could be used. However, they did not allow consumers to express ‘no preference’. The introduction of control groups, using putatively identical stimuli, solved this problem. The supplementation of frequency measures with d ′ measures allowed the use of signal detection protocols, which elicited more accurate measures of preference strength. This is still work in progress; further developments are expected. [ABSTRACT FROM AUTHOR]

The balanced-reference duo-trio test using the comparison of distances (COD) strategy is a widely used unspecified discrimination method. However, it suffers lower testing power. A dual reference duo-trio (DRDT) method was proposed recently as an improved duo-trio in the sensory literature. This paper proposes a new cognitive decision strategy for the DRDT, involving the comparison of pairs of distances (COPD) and investigates theoretically the performance of the DRDT based on a simulation-derived psychometric function. The DRDT is compared with the conventional balanced-reference duo-trio (DT) and the triangle methods. Comparison results show that the DRDT assuming the COPD is more sensitive than the conventional DT assuming the COD, and is more powerful than the conventional DT and the triangle methods for both difference testing and similarity/equivalence testing. The averaged generalized power efficiency of the DRDT is about 34% over that of the conventional DT, and is about 13% over that of the triangle method. [ABSTRACT FROM AUTHOR]

• There are simple alternatives to the traditional ANOVA analysis for scaling. • Novel analyses avoid the assumption problems of parametric analysis. • Using ranked data avoids distortion from artifactual ties. • R-index analysis for the 9-point hedonic scale breaks no assumptions. • R-index analysis now possible for data from unstructured hedonic line scales. In two experiments, five types of potato chip were assessed by 185 subjects, using an unstructured hedonic line scale and the 9-point hedonic scale to measure preferences among the different flavors of chip. The goal was to extend the applicability of the R-index analysis of hedonic scales. All prior experimentation had been concerned with studying the R-index analysis for the 9-point hedonic scale. This study was to extend the R-index analysis to the unstructured hedonic line scale. The result was that the R-index analysis could be successfully applied to unstructured hedonic line scales. It also confirmed earlier findings that it could be applied to the 9-point hedonic scale. The paper also contains a critical review of the traditional protocols for measuring preferences using hedonic scales, their methods of analysis and the assumptions made for these methods. [ABSTRACT FROM AUTHOR]

Abstract: The objective of this research was twofold: first, the performance of the tetrad protocol was compared to that of the triangle test under conditions that could possibly lower its sensitivity, consequently resulting in the loss of its theoretical power advantage. Second, the same samples were compared with a preference test to investigate whether a no difference conclusion obtained with a discrimination test would consistently result in a non-significant preference (consumer relevance). The investigation involved sensory differences that could be deemed small (d′ values less than 1.0) as well as the comparison of resampling vs. no resampling conditions. 456 consumers performed tests using apple and orange juices for which slight sensory differences were created through dilution. In all conditions, the tetrad always exhibited a greater number of correct answers than the triangle, confirming its greater statistical power. Therefore, it was concluded that even for small sensory differences, and in conditions where sensory fatigue could play a greater role (resampling allowed), the tetrad test sill appears like a good alternative to the triangle. Also, the theoretical increase in performance predicted when allowing sample resampling was confirmed. For the preference study, the same stimuli were evaluated by 208 subjects. Consumer relevance was defined as a significant result between two products in a preference test (assuming no population segmentation). Such significant preferences were found for three out of the four conditions, including the one with the smallest difference for which a significant result had not been found with either the tetrad or triangle. The non-significant preference in the fourth condition was attributed to segmentation in the population. Therefore, this investigation confirmed further that the tetrad test is a viable alternative to the triangle test, as it exhibits a greater statistical power even in conditions that could potentially affect it negatively. Also, it was shown that a non-significant sensory difference can still result in a significant preference test, outlining the necessity to go beyond the simple use of a ‘more powerful’ discrimination test when making decisions and to define the actual consumer relevance of an underlying sensory difference. [Copyright &y& Elsevier]

Abstract: Purpose: This randomized, open, controlled, multicenter study (110886/NCT00578227) evaluated human papillomavirus (HPV)-16/18 AS04-adjuvanted vaccine (HPV-16/18 vaccine) coadministered with inactivated hepatitis A and B (HAB) vaccine. Coprimary objectives were to demonstrate noninferiority of hepatitis A, hepatitis B, and HPV-16/18 immune responses at month 7 when vaccines were coadministered, compared with the same vaccines administered alone. Methods: Healthy girls (9–15 years) were age-stratified (9, 10–12, and 13–15 years) and randomized to receive HPV (n = 270), HAB (n = 271), or HPV + HAB (n = 272). Vaccines were administered at months 0, 1, and 6. Immunogenicity was evaluated at months 0 and 7. Results: The hepatitis A immune response was noninferior for HPV + HAB, versus HAB, for seroconversion rates (100% in each group) and geometric mean antibody titers (GMTs) (95% CI) (4,504.2 [3,993.0–5,080.8] and 5,288.4 [4,713.3–5,933.7] mIU/mL, respectively). The hepatitis B immune response was noninferior for HPV + HAB, versus HAB, for anti-HBs seroprotection rates (98.3% and 100%); GMTs were 3,136.5 [2,436.0–4,038.4] and 5,646.5 [4,481.3–7,114.6] mIU/mL, respectively. The HPV-16/18 immune response was noninferior for HPV + HAB, versus HPV, for seroconversion rates (99.6% and 100% for both antigens) and GMTs (22,993.5 [20,093.4–26,312.0] and 26,981.9 [23,909.5–30,449.1] EL.U/mL for HPV-16; 8,671.2 [7,651.7–9,826.6] and 11,182.7 [9,924.8–12,600.1] EL.U/mL for HPV-18, respectively). No subject withdrew because of adverse events. No vaccine-related serious adverse events were reported. Immune responses and reactogenicity were similar in girls aged 9 years compared with the entire study population. Conclusions: Results support coadministration of HPV-16/18 vaccine with HAB vaccine in girls aged 9–15 years. The HPV-16/18 vaccine was immunogenic and generally well tolerated in 9-year-old girls. [Copyright &y& Elsevier]

Abstract: Paired preference tests of liking require consumers to specify which of two foods are preferred or whether there is no preference. For ‘Buying’ preference tests, consumers specify which of two foods they are more likely to buy or whether they might buy either or neither. For ‘Choosing’ preference tests, consumers specify which of two foods they are more likely to choose when offered both, or whether they might choose both or neither. The first test is designed to predict greater degrees of liking, the second to predict buying behavior and the third to predict choice behavior. Yet, test results may not necessarily predict ‘real life’ behavior once consumers have left the testing situation; the ‘test’ preferences may not be ‘operational’. To validate the predictive ability of such tests, consumer behavior should be monitored for several months. Another approach is to allow consumers to take away some of the foods used in the test and observe what they take away. Consumers of potato chips were required to taste chips whose flavor and appearance were obviously different. A first group (N =107) were given ‘Liking’, ‘Buying’ and ‘Choosing’ preference tests for two palatable types of chip. After the test, out of sight of the experimenter, consumers were presented with two rows of plain plastic ‘snack’ bags, filled with the chips used in the test. They could take away either two bags of the same type of chip (a ‘Take Away’ preference) or one of each type (no ‘Take Away’ preference) or take neither (rejection). They were also subtly questioned regarding whether they were going to consume the chips themselves or were going to share them with friends. A second group (N =103) repeated the experiment. However, for these consumers, one type of chip was unpalatable. In general, ‘Buying’ preferences corresponded more to ‘Take Away’ preferences while ‘Liking’ and ‘Choosing’ preferences showed little correspondence. For the palatable chips, the preferences were spread more evenly than for the case where one type of chip was undesirable. In this latter case, approximately 50% of consumers were consistent over all tests in choosing the palatable chip. This might suggest operational preferences. ‘Take Away’ preferences suggested low predictability for paired preference tests of liking, choosing and buying. ‘Take Away’ preferences are not a substitute for monitoring consumer behavior for several months but they are a step in the right direction. [Copyright &y& Elsevier]

Abstract: Paired preference tests of liking require consumers to specify which of two foods are preferred or whether there is no preference. For ‘Buying’ preference tests, consumers specify which of two foods they are more likely to buy or whether there is no operational difference in likelihood to buy. The former test is designed to predict choice behaviour, the latter, buying behaviour. To validate the predictive ability of such tests, consumer choice and buying behaviour should be observed for several months. Another approach is to allow consumers to take away some of the foods used in the test and observe what they take. Consumers of potato chips were required to taste chips whose flavour and appearance were obviously different. A first group (N =109) were given a traditional paired preference test of liking. A second group (N =108) were given a preference test of buying. A third group (N =101) were given both tests. After the test, out of sight of the experimenter, consumers were presented with two rows of plain plastic ‘snack’ bags, filled with the chips used in the test. They could take away either two bags of the same type of chip (a ‘Take Away’ preference) or one of each type (no ‘Take Away’ preference) or take neither (rejection). ‘Liking’ preferences predicted what was taken away for only 48% of consumers for the first group and 45% (third group). For ‘Buying’ preferences, the correspondence was 56% (second group) and 60% (third group). Using buying tests separately or together with liking produced some differences. More reliably, approximately 60% of consumers in all groups chose one bag of each. ‘Take Away’ preferences suggested low predictability for paired preference tests of liking and buying. ‘Take Away’ preferences are not a substitute for following a consumer for several months but they are a step in the right direction. [ABSTRACT FROM AUTHOR]

Abstract: Using a Rank-Rating protocol and a traditional protocol whereby stimuli could not be re-tasted nor the scores be reviewed and modified, judges’ performance on category and line scales of intensity, in terms of the number of ‘different-stimulus’ scaling errors, was compared. A different stimulus scaling error was defined as giving a perceptually more intense stimulus an equal or lower score than a perceptually less intense stimulus. The scales compared were 9- and 20-point category scales and 10 and 20cm line scales. The stimuli were beef flavored soups of varying saltiness and orange flavored gelatin cubes of varying sweetness. There was a tendency for longer scales (20-point category, 20cm line) to yield fewer ‘different-stimulus’ scaling errors than the shorter scales (9-point category, 10cm line). [Copyright &y& Elsevier]

Abstract: After experimenting with questionnaires, Odesky had recommendations for eliminating ‘No Preference’ responses in a paired preference test. He concluded that, judges who had chosen the ‘No Preference’ option would, if forced to choose, express preferences in the same pattern as those who had actually expressed a preference. Using questionnaire, food and personal product stimuli, this conclusion was found to be commonly violated. Odesky’s recommendation, to ignore or distribute ‘No Preference’ responses in the same proportion as expressed preferences was sometimes seen to distort data, sometimes not, depending on relative response frequencies and extent of violation of his conclusion regarding forced preferences. When the number of ‘No Preference’ responses is not insignificantly small, to eliminate them, runs the risk of distortion of the overall pattern of preferences. [Copyright &y& Elsevier]

The present article provides an overview of our current knowledge on the topic of discrimination testing. First, the various goals of discrimination testing are outlined in terms of the objective of the investigation: psychophysics (understanding how the human senses work), Sensory Evaluation I (using the human senses as instruments to evaluate food characteristics) and Sensory Evaluation II (investigating the consumer''s ability to discriminate between foods). Then, theories are described allowing the selection of the most appropriate protocol based on the aim of the study. These theories include the Thurstonian approach to product measurements, taking into account the central processing of information in the brain. The effect of experimental factors such as memory, sensory ‘fatigue’, sample retasting and practice are also considered. The consideration of all these variables will allow the selection of the most suitable protocol for investigations involving discrimination testing. [Copyright &y& Elsevier]

• Psychometric function based on a normal distribution function. • Psychometric function based on a bivariate normal distribution function. • Psychometric function based on a non-central F -distribution function. • Four-interval oddity test is more powerful than the Triangle and Duo-trio tests. • Tables and R codes for data analysis using the four-interval oddity task. The four-interval oddity task is an extension of the Triangle method, which is the three-interval oddity task and is one of the most broadly used conventional sensory discrimination methods. It is a special case of the m - interval oddity task or the m - alternative odd-man-out task where m = 4. It is also a special situation of the unspecified 'M + N' with M = 3 and N = 1. Some new forms of the analytical psychometric function for the four-interval oddity task based on a normal distribution function, a bivariate normal distribution function, and a non-central F -distribution function are provided originally in this paper. The performance of the four-interval oddity task in both difference testing and similarity/equivalence testing is explored and compared with some unspecified forced-choice sensory discrimination methods including the Triangle. Tables and R codes are presented and provided for estimations of the probability of correct response, P c , Thurstonian discriminal distance δ or d´ , and the B value for estimating the variance of d´ for the four-interval oddity task. [ABSTRACT FROM AUTHOR]

• Stimuli in same category of 9-point hedonic scale are not liked equally. • Assigning numbers to the categories introduces artifactual ties. • Artifactual ties reduce test power. • R-Index analyses produce more significant differences than ANOVA/LSD analysis. • R-index analysis to the 9-point hedonic scale for measurements of preference is fit for purpose. The traditional protocol for using the 9-point hedonic scale for measuring multiple preferences among a set of products, involves assigning numbers (1–9) to the scale's verbal categories, which are erroneously treated as interval data. Means of these numbers for each product are treated by parametric statistics to yield an ordinal series of means, representing the degrees of liking as the effect sizes for the products in the set. Accordingly, for preference, products with a higher liking mean scores are deemed as being preferred to products with lower mean scores. This gives only the direction of the preferences among the products in the set but not their strengths. These have to be surmised from the means that were significantly different. Recently, an alternative protocol for the 9-point hedonic scale was developed, using the ranks of the scores on the 9-point scale. It used an R-index analysis of these ranks directly to yield preference probabilities for comparisons among products as its effect sizes. Besides appropriate effect sizes, attention was drawn to the absence of questionable statistical assumptions as well as superior ergonomics. Answering a call for further research and confirmation, this paper challenged these early results and was more demanding than the original. Using smaller numbers of products, the results of the R-index analysis shadowed those for the traditional analysis as expected. For these conditions, the only advantage for the R-index analysis was its ergonomics. As the number of products under assessment were increased, another advantage emerged; it was the elimination of artifactual ties. This was confirmed by the number of cases that were significant for the R-index analysis but not for the traditional analysis. All previous experiments were confirmed as were a set of conclusions, which were illustrated by feasible case studies. Together they indicated that the R-index analysis is definitely fit for purpose. [ABSTRACT FROM AUTHOR]

• An analytical psychometric function for the dual reference duo-trio (DRDT). • Two forms of the analytical psychometric function for the DRDT. • Tables and R codes for P c , d ' , and variance of d ' (B value) based on the function. The dual reference duo-trio (DRDT) is an improved balanced-reference duo-trio method, in which two same references are presented in a test but the reference is changed across the tests. The performance of the DRDT has been investigated theoretically based on a simulation-derived psychometric function in a previous paper published in this journal. The DRDT method is more powerful than the conventional duo-trio and the triangle methods. This paper derives an analytical psychometric function for the DRDT. Another form of the psychometric function for the DRDT is also derived based on the Torgerson's method of triads. Tables and R codes are provided for presentations and calculations of P c , δ (or d '), and variance of d ' (B -value) based on the analytical psychometric function for the method. [ABSTRACT FROM AUTHOR]

Abstract: Judges performed over 2000 same–different tests followed by over 2000 A Not-A tests, to distinguish between ‘threshold’ sodium chloride solutions and purified water. Fitting ROC curves to the data indicated that three out of four judges used a τ-strategy for the same–different test. The hypothesis that judges might, with experience, start to categorize the stimuli and switch to a β-strategy was generally not confirmed. On changing to the A Not-A method, half the judges appeared to continue initially to use the τ-strategy. The hypothesis that judges might make comparisons with the preceding stimulus rather than the standard stimulus, because of prior extensive experience with the same–different test, was generally not confirmed. Yet, in all cases there were exceptions. [ABSTRACT FROM AUTHOR]

Abstract: The original 9-point scale, developed by the U.S. army for menu planning for their canteens, consisted of a series of nine verbal categories representing degrees of liking from ‘dislike extremely’ to ‘like extremely’. For subsequent quantitative and statistical analysis, the verbal categories are generally converted to numerical values: ‘like extremely’ as ‘9’, ‘dislike extremely’ as ‘1’. Yet, sometimes what is termed a 9-point hedonic scale is an unstructured numerical scale, labeled at the ends with ‘dislike extremely’ and ‘like extremely’. The former scale requires consumers to categorize foods according to how much they are liked or not; the latter requires the consumers to differentiate numerically between the foods in terms of the relative degree of liking for each. Foods that were placed in the same verbal category for the former scale might be given different numerical scores on the second scale. To illustrate this, consumers rated five chocolates, in a series of experiments, on these two types of 9-point scale (verbal categories only vs numbers only) and the proportion responding differently to the two scales ranged from 100% to 79%. This indicated that numerical data obtained from both types of 9-point scale were not interchangeable. It also suggested that consumers were using different cognitive strategies for verbal categories and numbers. To check that the difference was not caused by the fact that the verbal categories were bipolar and the numbers unipolar, the experiment was repeated using a bipolar number scale (–4 through 0 to +4). The same results were obtained. For comparison, a 9-point hedonic scale including both verbal categories and numbers together, was also used. The results for this scale showed a greater similarity to the version of the 9-point scale consisting only of verbal categories than the unstructured numerical version. Stimulus equalizing bias was used as a tool to make a preliminary investigation into the cognitive strategies involved for the two versions of the scale. The hypothesized relative strategy was confirmed for the unstructured numerical scale but the hypothesized absolute strategy was not confirmed for the scale using only verbal categories; the strategy appeared to have relative elements. Regardless of the precise nature of the cognitive strategies used for two versions of the scale, they do not give the same results and data obtained from each version should be compared with caution. [Copyright &y& Elsevier]

Abstract: Background: The dissolution profiles of generic oral bisphosphonate alendronate (ALN) sodium for the treatment of postmenopausal osteoporosis differ by formulation, suggesting potential differences in the risk for upper gastrointestinal (GI) irritation. Objective: This study compared the tolerability profile of ALN monohydrate with that of placebo, with a focus on upper GI irritation, in postmenopausal women with osteoporosis. Methods: This multicenter, double-blind, placebo-controlled estimation study enrolled postmenopausal women with osteoporosis. Patients were randomized in a 2:1 ratio to receive ALN monohydrate 10 mg or placebo once daily for 12 weeks. Tolerability was monitored throughout the study and up to 14 days after administration of the final dose. Primary end points were the proportions of patients with upper GI adverse events (AEs); upper GI AEs that were rated as serious or study drug related or that led to study discontinuation; and esophageal AEs. Between-treatment differences and associated 95% CIs were assessed using the Wilson score method. Results: Of 438 patients who were randomized, 367 (mean age, 65.5 years; history of osteoporotic fracture, 6.8%; ALN monohydrate, 237; placebo, 130) completed the study. The proportion of patients with a history of upper GI disorders at baseline was numerically greater in the ALN monohydrate group than in the placebo group (117 [40.2%] and 45 [30.6%], respectively). The proportions of patients with active baseline upper GI disease were 83 (28.5%) and 30 (20.4%) in the ALN monohydrate and placebo groups, respectively. The proportions of patients who experienced an upper GI AE during the study period were 66 (22.7%) and 30 (20.4%) (95% CI, −6.2 to 10.0). The proportions of patients with upper GI AEs that were rated as serious or study drug related or that led to study discontinuation were 20.3% and 12.9% (95% CI, −0.3% to 14.1%). Three serious AEs in the active-treatment group (breast cancer, 2; wrist fracture, 1) were not considered related to the study drug, nor was the 1 serious AE in the placebo group (wrist fracture). One patient (ALN monohydrate) had an esophageal AE (nonserious spasm). Approximately 8% of patients who received ALN monohydrate reported dyspepsia, compared with none who received placebo. Within each treatment group, the rates of upper GI AEs were numerically higher in patients with a history of upper GI disease. Conclusions: In these postmenopausal women who received ALN monohydrate or placebo, upper GI AEs were common (20.4%–22.7%). The proportion of patients who experienced upper GI AEs considered drug related or that led to discontinuation was appar- ently greater with ALN monohydrate compared with placebo. [Copyright &y& Elsevier]

Abstract: Trained judges were required to perform a descriptive analysis of a lemon flavored beverage, using a serial monadic protocol and an attribute-by-attribute protocol. They had received sufficient training to establish three intensity exemplars for each attribute in the beverage. They were tested until they reached criterion performance for each protocol. Criterion performance required judges to rate all attributes according to the rank order of their physical strengths. Judges attained ‘criterion’ performance more rapidly using the serial monadic protocol. The results contrast with those of another study where untrained judges reached criterion performance more rapidly with the attribute-by-attribute protocol. [Copyright &y& Elsevier]

We have demonstrated that bronchoalveolar lavage fluid from chronic obstructive pulmonary disease patients contains higher levels of interferon-y compared with controls. Interferon-γ is a potent inducer of various cathepsins and matrix metalloproteases. Therefore, we postulated that interferon-y could induce protease expression by macrophages in acute and chronic lung disease. Chronic obstructive pulmonary disease patients had greater levels of cathepsin S and matrix metalloprotease-12 in their bronchoalveolar lavage fluid. Macrophages incubated with chronic obstructive pulmonary disease bronchoalveolar lavage fluid exhibited increased expression of cathepsin S and matrix metalloprotease-12, which was inhibited by the addition of interferon-γ-neutralizing immunoglobulin. Human secretory leukocyte protease inhibitor is an 11.7-kDa cationic non-glycosylated antiprotease synthesized and secreted by cells at the site of inflammation. We have demonstrated that secretory leukocyte protease inhibitor can inhibit interferon-γ-induced cathepsin S production by macrophages. Pretreatment of macrophages with secretory leukocyte protease inhibitor inhibited interferon-γ-induced inhibitor κB β degradation and activation of nuclear factor κB. Secretory leukocyte protease inhibitor may prove to be therapeutically important as a potential inhibitor of protease expression in chronic obstructive pulmonary disease. [ABSTRACT FROM AUTHOR]

Abstract: Two experiments were performed in which untrained judges were required to perform a descriptive analysis, using a serial monadic protocol and an attribute-by-attribute protocol. They reached ‘criterion’ performance more rapidly using the attribute-by-attribute protocol, indicating that they were using a relative scaling strategy. Criterion performance required judges to rate all attributes according to the rank order of their physical strengths. Later, they reached criterion performance with the serial monadic protocol, indicating that during experimentation, they had learned to establish sufficient intensity exemplars for an absolute scaling strategy. [Copyright &y& Elsevier]

Abstract: Paired preference tests were performed using plain salted and lemon flavored potato chips. In some cases, identical chips were tested to determine the effects of the experimental situation per se (placebo condition). In other cases, different chips were tested and this allowed measurement of preference, accompanied by the effects of the experimental situation (test condition). Comparison between the two, allowed a measure of preference. A total of 1100 consumers were tested in a variety of experimental conditions. In one condition, a single pair of identical chips was tested while in a different condition, two pairs of chips, one identical, one different, were tested; this allowed both related samples and independent samples designs. All possible combinations of the two orders of pairs (same vs different) and the four orders of tasting within the pairs were tested. Whether the placebo condition was tested first or second appeared to have little effect. Chi-squared was used to determine whether the preference results were significantly different from a ‘no preference’ situation and d′ was used to give a measure of the degree of preference for one product or another. Within-subjects and between-subjects analyses gave the same results. The placebo condition allowed the detection in the test condition of significant preferences as well as the presence of two equally balanced preference groups. The placebo condition also had the potential to be used as a screening tool. [Copyright &y& Elsevier]

Abstract: For sensory difference tests, one way, but not the only way, of dealing with the problem of overdispersion is to use a beta-binomial analysis. Commonly, binomial statistical analyses are used for these methods and they assume that the sensitivity of the judges is uniform. However, judge sensitivity varies and this adds a problematical extra variance to the distribution. This is termed overdispersion and renders simple binomial analysis prone to Type I error. The distribution of sensitivity of the judges is described by a beta-distribution. The analysis, combining beta and binomial distributions, gives an index, gamma. This ranges from zero, for no overdispersion, to unity, for total overdispersion. A compact beta-distribution clustered around the mean of the binomial distribution, would add little extra variance and elicit minimum distortion of the binomial distribution, yielding a zero or near zero gamma value. A more scattered or even bimodal beta-distribution would have a substantial effect and yield a significant gamma value. One question that has been posed is whether some test methods are more prone to overdispersion than others. Yet, a consideration of the reasons for overdispersion would suggest that significant gamma values were more a result of obtaining a heterogenous sample of sensitive and insensitive judges by chance. To confirm this, ‘less sensitive’ and ‘more sensitive’ samples of judges performed 2-AFC and 3-AFC tests with resulting zero gamma values, indicating no overdispersion. However, when the less and more sensitive groups were combined, significant gamma values were obtained, indicating the presence of overdispersion. However, in a further experiment using 2-AFC tests, when the ‘less sensitive’ group had its sensitivity increased by a ‘warm-up’ procedure, combination with the ‘more sensitive’ group did not result in overdispersion. [Copyright &y& Elsevier]

• A semantic model of user activities as a timeline of activity objects. • A generic activity recommendation framework for multiple recommendation scenarios. • Recommending the next activity, the next sequence, & context for the next activity. • Recommendation-based and context-aware post-filtering. • Experiments on the lifelogging, urban computing and tourism (LBSN) domains. Since the last decade, data collection is becoming more pervasive, passive and easier to perform. This is resulting in the rise of data wherein a user performs some activities in a sequence, such as locations visited, physical activities performed, and modes of transport taken. In such cases, activities are often performed in a particular order, and each activity in turn may influence the subsequent activities to be performed. Moreover, such activities may be associated with multiple features or contexts, such as location, time, weather, etc. The order encoded in such data, along with the context, capture important information when it comes to modelling the preferences and personal habits of users. Traditional recommender systems, however, typically do not consider the order in which users perform activities and there is little work which considers both sequence and context simultaneously. In this work, a generic recommendation framework is proposed which leverages both sequences and context in user activity data for activity recommendation. To model user activities, a semantic view of the user's past activities as a timeline of activity objects is presented. An essential step in the recommendation process is finding patterns in past activities performed which are closely aligned to the recent activities undertaken by the user. To calculate the distance between timelines, a novel two-level distance metric is presented which calculates distance with respect to the order of the activities as well as the context features associated with each activity occurrence. The efficacy of the proposed activity recommendation framework in various recommendation scenarios, is demonstrated using real-world datasets from multiple domains. [ABSTRACT FROM AUTHOR]

• Relative model is better fit for unstructured intensity scales than absolute. • Memory variance from serial monadic protocols is major source of discrimination errors. • Rank-rating produces fewer discrimination errors than serial monadic. • Rank rating with sufficient space elicits fewest discrimination errors. A sample of 62 untrained subjects were assessed on their ability to use unstructured numerical 9-point and 30-point category scales along with an unstructured line scale, using both rank-rating and serial monadic protocols. Visual stimuli were used for convenience, the task being to rate the heights of 12 easily discriminable columns of mung beans held in transparent vertical cylinders. Such stimuli had no perceptual variance, which would otherwise have added uncontrolled variance to the subjects' performance. Two measures of performance were used for each of the 6 experimental conditions. First, mean number of 'scaling errors' made in each of the six experimental conditions was computed. In this experiment, a scaling error was defined as giving a taller column a score equal to or less than a shorter column. The lower the error count, the better the subjects' performance. The second measure was to match the subjects' rating scale pattern of scores to a 'true' pattern of scores, derived from the physical measurements of the 12 columns. For this, a 'dissimilarity score' was developed. This compared the sum of the Euclidean distances between standardized true scale ratings for each of the column's 12 true heights, with those obtained from each subject. This gave a measure of the mismatch between the subject's set of scores and the true set of scores. Both the scaling error counts and the dissimilarity measures, indicated that subjects performed significantly better using the rank-rating protocol than the serial monadic. This was because of the effects of forgetting the exact intensities of stimuli once they had been removed, removal of stimuli being a necessary part of the serial monadic protocol. Subjects were penalized when using the 9-point scales, because there were too few categories to represent the different heights of all 12 columns. This introduced the concept of 'sufficient space'. Using the rank-rating protocol, the 30-point and line scales, with no memory problems and sufficient space elicited the best performances; they were not significantly different. [ABSTRACT FROM AUTHOR]

• Placebo pairs assessed after target pairs elicit more 'no preference'. • This poses a potential for a more powerful statistical analysis. • High products variance may contribute to sensory input for a placebo pair. • Triadic preference tests boost the number of 'no preference' in the placebo pair. One type of paired preference test uses the 'target' pair of stimuli under consideration to record the measured preferences (prefer A, Prefer B, 'no preference') and a second putatively identical control pair, the 'placebo' pair (AA or BB) to indicate 'false' preferences, unrelated to the target pair, elicited by the effects of the testing situation. From the literature there is disagreement regarding whether it is important to place a placebo pair before or after its corresponding target pair, to elicit a greater proportion of 'no preference' responses. This is important, because the higher the frequency of 'no preference' responses in the placebo pair, the more powerful will be the chi-squared style analysis, which determines whether the target pair displays a significant preference or not. It has been hypothesized that placing the placebo pair after the target pair would elicit more 'no preference' responses in the placebo, because the consumers would have had a chance to experience the difference in sensory characteristics of the two stimuli in the target pair. Using a related samples design, the hypothesis was confirmed although the difference between placebos placed before and after the target was not always significant. It was hypothesized that this lack of significance was caused by greater variance among the items being assessed, making it more difficult to decide whether a putatively placebo pair, in the context of the experiment, was really 'identical'. Psychological adjustments in the consumers were discussed in terms of difference and preference tau criteria. A boost in the proportion of 'no preferences' was observed for the placebo pair built into the triadic preference test, this was explained by the experimenter gaining some control over the consumers' preference tau criteria. [ABSTRACT FROM AUTHOR]

• Actionable preference tests have some predictive success for real-life choice. • Consumers as a whole maintained real-life constant preference over 3 months. • Largest group of individual consumers also remained constant over 3 months. • Consumers who changed preference >3 times comprised the smallest group. • Evidence that context change triggered preference change. In Part 1, 115 consumers were required to perform actionable 'choosing' preference tests which included 'no preference' options, between original and barbecue flavored cassava chips, in laboratory conditions. Double-DTM preference tests, 'Strict' paired preference tests (only choose A, never choose B), and 'Lenient' paired preference tests (choose A more than B) were used. In Part 2, consumers were required to indicate which of the two chips they would choose if presented with both, in real life conditions, on nine testing weeks during the following three months. Real life conditions present different extraneous variables. Considering the sample of consumers as a whole, a slight preference for the original flavored chips noted in all three tests in Part 1 was continued for all nine testing weeks in Part 2. Considering individual consumers, the largest group continued to choose the chips they had chosen in Part 1 for all nine testing weeks. The second largest group changed their preference in the first week and then stayed constant for the following eight weeks. This could suggest that a change in extraneous variables might be a trigger for preference change. Considering only consumers who changed their preferences, the majority changed between 1 and 3 times. The tendency for change was low. Consumers who had declared 'no preference' changed to exhibit preference behavior. A fourth preference test method was used in Part 1. This was a numerical bipolar rating scale for likelihood to choose. However, it lacked sufficient precision in its predictions for any usefulness. [ABSTRACT FROM AUTHOR]

This paper proposes a new type of sensory discrimination methods, i.e., paired versions of various sensory discrimination forced-choice methods, which can be regarded as modified versions of the same-different method. The new type of methods possesses the advantage of the same-different test for measuring overall sensory difference, but overcomes its weaknesses of large variations of response bias and lower testing power. An important finding is that this type of methods is closely related to the same-different area theorem and its extension. The same-different area theorem establishes an equivalence between the area under the receiver operating characteristic (ROC) curve of the same-different method and the proportion of correct responses ( P c ) in the dual-pair method. This paper re-demonstrates and extends the same-different area theorem. An analytical expression of a general psychometric function for the new type of methods is derived. Performances of the new type of methods are evaluated with testing power and sample size in both difference and similarity/equivalence tests. R codes are provided for the new type of methods for P c , Thurstonian discriminal distance δ, and variance of d′ , where d′ is an estimator of δ. [ABSTRACT FROM AUTHOR]

• Placebo pair preferences explained by cognitive systems analysis. • Increasing power by transfer to Central processing. • Using triadic protocol to induce central processing. • Assessing doubling to avoid odd sample bias. The development of the triadic preference test and how it differed cognitively from other paired preference tests involving placebo pairs, is outlined. The test elicited a far greater proportion of 'no preference' responses from placebo pairs than expected, thus increasing its power. Although the reasons for this were not fully understood at the time the test was first introduced, a necessary understanding of the mechanism later evolved with the help of a recently published systems analysis and some results from social psychology. Later, to this was added a warning that the test was prone to odd sample bias. The bias could be avoided by adjustment of the counterbalancing controls, using a method similar to that for the replicated duo-trio difference-preference test. Unfortunately, there had been some misunderstanding about the instructions in the warning, so an experiment was performed to re-affirm the warning using the correct instructions. The experiment confirmed that the test was prone to the odd sample bias, so that the counterbalancing changes were adopted. Yet, the bias was not present for all products being tested. This suggested that there are still some questions to be answered for what is still an under-researched area. [ABSTRACT FROM AUTHOR]

A triadic preference test was developed as an alternative to the paired preference test. The three stimuli in the test comprised a putatively identical placebo pair and a different stimulus. This was in contrast to the regular paired preference test that utilizes a placebo pair. Such a test requires the presentation of two pairs of stimuli: a putatively identical placebo pair and a test pair. The triadic preference test only requires one triad. With the regular test, the majority of consumers respond to the placebo pair with a preference response. It is generally assumed that these consumers are responding to extraneous factors: those factors that elicit a preference response that are different from the sensory attributes of the food under assessment. As an attempt to minimize the possibility of responses to extraneous factors when assessing the test pair, it has been suggested to only use those consumers who chose the ‘No Preference’ option for the placebo pair. However, this form of ‘screening’ is not viable because the resulting ‘screened’ sample size is greatly reduced to approximately one third. However, in the present study, with the triadic preference test, the resulting ‘screened’ sample size ranged 76.5–94% of the total. Thus, this form of ‘screening’ against consumers who demonstrated response to extraneous factors for the placebo pair, was now feasible. [ABSTRACT FROM AUTHOR]

Abstract: The effects of sensory and non-sensory factors on the liking of Korean style salad dressings and beverages among US subjects in two locations (California and Minnesota) and Korean subjects were investigated. Four types of dressing and five types of beverage samples were evaluated. Approximately, half of the subjects evaluated the samples under blind-labeled conditions while the other half evaluated the samples labeled with their corresponding names and flavor descriptions. The liking of each sample was rated and the reasons for liking and disliking each sample were surveyed. Various food attitudes were measured on the food neophobic scale, VARSEEK scale, and flavor attitude scales. Soy sauce & vinegar dressing was the most preferred sample among the US subjects, whereas sesame seed dressing was preferred as much as the soy sauce & vinegar dressing among Koreans. Cinnamon-ginger flavored beverage was preferred the most among the US subjects, whereas rice punch was preferred the most among Koreans. Sample labeling effect was relatively small compared to other factors. VARSEEK and food neophobia scale showed stronger effect on beverage than salad dressing. Preference attitudes for hot & spicy and roasted carbohydrate flavors affected the acceptance of salad dressing among consumers in California whereas preference attitude for garlic flavor affected the acceptance of dressing among consumers in Minnesota. [Copyright &y& Elsevier]

Abstract: The effects of the application of foliar fertilisation and pesticide on nutritional quality of mandarin orange juices were evaluated using 1H NMR metabolomics. Significant differences between the use of fertiliser and pesticides during fruit formation were observed, and included changes in sugar, amino acid and organic acid composition. To determine whether the difference in sugar concentration was enough for the consumer to detect, a sensory experiment was performed in which two orange juice samples were prepared to resemble the sweet/sour taste balance of juice from mandarin oranges in which foliar fertilisation was either applied or not. In a test using non-trained individuals, 68% could correctly identify which juice had a sourer, or less sweet, taste. The implications of this study could impact citrus growers, and ultimately aid in development of fruit with superior sensory quality. [Copyright &y& Elsevier]

Abstract: The use of a Pump-Suppressed Four Wave Mixing PS-FWM wavelength converter in a wavelength conversion and selection switching scheme is examined and demonstrated. The scheme requires a single 2×1 Mach Zenhder Interferometer (MZI) and offers substantial pump suppression and efficiency improvement in terms of optical signal-to-noise ratio. Performance measurements on the combination of PS-FWM wavelength converter with an Arrayed Waveguide Grating (AWG) for passive routing, shows the potentially successful use of the scheme in an ultra-fast wavelength routed optical packet switching scheme where the use of optical filters for the separation of the converted signal from the pump is avoided. [Copyright &y& Elsevier]

Two experiments were conducted to investigate the suitability of the Thurstonian models for the 2-AFC (2-alternative forced choice) and 2-AC protocols (2-AFC test with a “no difference” option). The Thurstonian approach predicts that if the same subjects and products are used, the degree of difference measured between the products (d′) should be the same. Experiment I tested this prediction using solutions with different salt concentration. The average d′ values for the 2-AFC and 2-AC were found to be 1.66 and 1.54, respectively, and were not significantly different (t-test, p=0.46). This indicated the suitability of the models for a taste stimulus. Experiment II used a trigeminal stimulus: sparkling water with different CO2 levels as a food system. Three levels of carbonation were used and all three possible pairs of samples were compared. The d′ values measured for the 2-AFC and 2-AC for each pair of products were: pair 1––1.7 and 1.7; pair 2––0.8 and 0.7; pair 3––1.2 and 1.0. For each pair of products, no significant difference was observed between the d′ of each protocol (t-tests, p=0.65, 0.20, 0.30, respectively), giving further support to the Thurstonian models of these paradigms and indicating their robustness. [Copyright &y& Elsevier]

This paper reviews neurophysiological and psychological studies of oral irritation elicited by chemicals in spicy foods and carbonated drinks. Oral irritant, thermal and textural sensations are conveyed to the brain by the trigeminal pathway, which is separate from the gustatory and olfactory systems. In humans, repetitive application of capsaicin, citric acid, or concentrated NaCl elicits oral irritation that grows in intensity across trials (“sensitization”). After a rest period, reapplication elicits less irritation (“self-desensitization”), but if given recurrently will eventually evoke a progressive rise in irritation (“stimulus-induced recovery”=SIR). In neurophysiological recordings from neurons in the trigeminal subnucleus caudalis (Vc), the first relay in the pathway for oral somatosensation, these irritants elicit a similar pattern of progressively increasing firing, followed after a rest by self-desensitization and SIR. In contrast, nicotine, menthol or mustard oil elicit irritation that decreases across trials (“desensitization”), a pattern also observed in Vc neuronal responses to these irritants. Carbonated water elicits an oral tingling sensation and excites Vc neurons largely through its conversion to carbonic acid. The good correspondence in temporal profiles for perception and neuronal activity supports a role for Vc neurons in the mediation of oral irritation. Finally, the development of preference for foods containing aversive chemicals is addressed. This may involve mere exposure, social reinforcement, the “thrill” of the strong sensation, or physiological reinforcement associated with satiety or release of endorphins by the painful stimulus. [Copyright &y& Elsevier]

• Rival cognitive strategies were involved in traditional 9-point hedonic scale. • Absolute and relative behavioral models may be involved with the verbal version of 9-pt hedonic scale. • Numerical scale was much more prone to Poulson's Stimulus Range Equalizing Bias than the verbal scale. Besides confirming earlier work, this study provided evidence for the theoretical position that consumers use cognitive strategies based on an absolute behavioral model, for the traditional verbal version of the 9-point hedonic scale. This is opposed to the relative model, which is appropriate for the less frequently used purely numerical and unstructured version of the 9-point hedonic scale. After establishing a constant rank order of liking for six chocolate stimuli, consumers rated the chocolates using the traditional verbal version and an unstructured numerical version of the 9-point hedonic scale. The numerical data produced by each scale were different and incompatible for significant majorities of consumers, after correction for their unimodal and bimodal designs. This occurred with both rank-rating and serial monadic protocols. This was confirmed using pattern change scores, a parametric measure of the discrepancy between 'words only' and 'numbers only' scaling responses. It also checked for 'scaling drift', inconsistency in scale use. The fact that numerical scales based on the relative model are prone to Poulson's Stimulus Range Equalizing Bias and that scales based on an absolute model are not, was used to investigate the model for the traditional verbal 9-point hedonic scale. The verbal scale was much less prone to Poulson's bias than the numerical scale but it was still partially prone. Poulson's bias had demonstrated that the model for the traditional verbal 9-point scale was different from the relative model for a purely numerical scale. Yet, it had failed to demonstrate that the different model was the absolute model. It had partial absolute and partial relative properties. Descriptions of the nature of the model are discussed. Compared with rank-rating, the effects of high memory load, which is systemic for the serial monadic protocol, was evidenced by higher error rates due to forgetting, higher amounts of scaling drift and discrepancies between the order of ranked and rated data. [ABSTRACT FROM AUTHOR]

• Placebo pairs after target pairs elicit more 'no preference' responses. • The differences are not always significant. • Stimuli variances reduce 'no preference' in placebo pairs placed after target. • The results generalize between cultures. From the literature on paired preference testing, there has been disagreement regarding whether a placebo pair would have a significantly higher frequency of 'no preference' responses if it were to be placed for assessment after its corresponding target pair rather than before. This can be important, because the higher the frequency of 'no preference' responses in the placebo pair, the more powerful will be any chi-squared related analysis, which determines whether the target pair indicates a significant preference or not. In the first paper in this series, it was shown, that indeed a placebo placed after the target pair induced a higher proportion of consumers to respond with a 'no preference'. However, the response was uneven. For some stimuli, the response was strong and significant, for others it was weaker and not significant. It was hypothesized that the weak response could be due to greater variance among the individual stimuli in the placebo sample to be tasted. The effect was confirmed using a priori chosen high and low variance stimuli. Further evidence was obtained from predictions for preference tau criterion levels and frequency of preference changes between two target pairs. All these indicated that the weaker response of some stimuli was due to a higher level of variance among the individual stimuli. [ABSTRACT FROM AUTHOR]

This study evaluated safety and efficacy of esterified estrogens and methyltestosterone administered alone or in combination for the treatment of hot flashes in menopausal women. The 0.30-mg esterified estrogens and 0.30-mg methyltestosterone combination was the lowest effective dose, and our results are consistent with the known safety profile of estrogen and androgen combination products. [Copyright &y& Elsevier]

• Effect sizes are necessary to complete the analysis for the 9-point hedonic scale. • R-Index Preference Measurement is a nonparametric measure of effect size. • ANOVA/LSD analysis for 9-point hedonic scale introduce artifactual ties. • Artifactual ties noticeably reduce the power when more than 8 stimuli were tested. • R-Index Preference Measurement eliminates artifactual ties. 214 consumers used the verbal 9-point hedonic scale to assess 4 types of flavor coated peanuts and 4 types of flavored teas. They used the traditional ANOVA/LSD analysis to provide mean values derived from the 9-point hedonic scale along with measures of significant difference. However, these data did not provide effect sizes. They did not give direct measures of the strength of preference between the various products, which was the main interest. Accordingly, effect sizes were computed. For this, each consumer had also ranked their preferences as they made their ratings on the 9-point hedonic scale. From these, R-Index values were computed to provide the percentages of consumers, who preferred each product to every other product. These direct measures of effect size completed the analysis begun by the ANOVA analysis of the set of mean scores. Also, the measures were nonparametric and avoided issues of the validity of a parametric statistical analysis. They also avoided the problem with the traditional analysis when products in the same scale category are attributed the same scores, when they are not equally liked. Experiment 2, using 207 consumers indicated that this problem was only serious enough to reduce the power of the traditional analysis, compared with the R-Index Preference Measurement, when the number of products being tested approached a dozen say, for product optimization. [ABSTRACT FROM AUTHOR]