Search

Abstract: The 2-AC protocol is a 2-AFC protocol with a “no-difference” option and is technically identical to the paired preference test with a “no-preference” option. The Thurstonian model for the 2-AC protocol is parameterized by δ and a decision parameter τ, the estimates of which can be obtained by fairly simple well-known methods. In this paper we describe how standard errors of the parameters can be obtained and how exact power computations can be performed. We also show how the Thurstonian model for the 2-AC protocol is closely related to a statistical model known as a cumulative probit model. This relationship makes it possible to extract estimates and standard errors of δ and τ from general statistical software, and furthermore, it makes it possible to combine standard regression modelling with the Thurstonian model for the 2-AC protocol. A model for replicated 2-AC data is proposed using cumulative link mixed models. Examples will be given throughout the paper and the methodology is implemented in the authors’ free R-packages sensR and ordinal. [Copyright &y& Elsevier]

Besides responses to the target pair of stimuli being tested in a paired preference test, comparison with responses to a putatively identical "placebo" pair can be used for a test of significance, to determine whether the response frequencies elicited by the target pair are different from those that would be obtained had there been no preference. Yet, preference responses elicited by the placebo pair can also be used as a way of identifying consumers who bias avoided the "No preference" option and thus generated biased response. Such consumers can then be eliminated. Yet, using this approach greatly reduces the sample size. However, using some response options associated with purchase intent rather than liking can reduce the number of consumers that are eliminated. This was confirmed for potato chip, milk and orange juice stimuli. From psychology, the concepts of operational preference and types of response bias appropriate to preference testing were introduced and discussed. PRACTICAL APPLICATIONS: The measurement of preference and acceptance of foods is important for product development and decisions regarding the launching of new products into the market. The paired preference test has some issues associated with its design and analysis, and these must be considered if preference tests are to give valid and reliable data. One of the problems is that the majority of consumers tend to report preferences even when the stimuli are putatively identical; here, a "No preference" response would seem more appropriate. Such consumers would seem to be reacting to the extraneous conditions of the preference test rather than the sensory characteristics of the putatively identical stimuli under consideration. Such responses could be termed "biased response." Yet, the consumers' responses to the target (different) stimuli under consideration could be trusted more if the consumers had shown that they did not have this tendency to give "biased responses." They would show this by selecting a "No preference" response to putatively identical stimuli. This paper shows that the tendency to give "biased response" can be reduced by manipulating the wording of the preference question. This does not furnish a complete solution to the problem but it indicates one possible step toward solution and adds valuable information to the body of knowledge that is being built with a view avoiding the difficulties associated with preference testing. [ABSTRACT FROM AUTHOR]

Preference tests were performed for varieties of potato chips, orange juices and chocolate chip cookies using three response protocols: the traditional paired preference test with the “no preference” option, a 9-point hedonic scale and a 6-point hybrid hedonic/purchase intent scale. The different stimuli to be assessed were presented in pairs, but putatively identical stimuli were also presented as a “placebo” pair. Performance on the placebo pair with identical stimuli provided a measure of the hidden demand characteristics of the test protocol. The presentation of the different pairs provided a measure of preference accompanied by such hidden demand effects. Comparison between the two allowed a better measure of preference per se. The order of presentation of the identical and different pairs did show occasional slight evidence of contrast effects. For the placebo “identical” pairs, a majority of consumers reported false preferences. Liking questions with the hedonic and hybrid scales elicited fewer false preferences than preference questions with the paired preference protocol. Yet, the effects tended to be slight. The 6-point hedonic/purchase intent scale exhibited the fewest false preferences in the placebo condition, and this was because of its fewer categories rather than any cognitive strategy change elicited by its different labels. PRACTICAL APPLICATIONS A problem with paired preference testing is the tendency of consumers to give false preferences, which produces the seriously misleadingoverestimation of the proportion of consumers who have preferences for one or other of the products being assessed. The placebo condition is an important control for alleviating this problem. The statistical analysis can be improved by finding a protocol that maximizes the proportion of “no preference” responses in the placebo condition. The key finding here is that using hedonic or purchase intent questions rather than preference questions may possibly provide a way of achieving this aim. [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]

As reformulations and processing changes are increasingly needed in the food industry to produce healthier, more sustainable, and cost effective products while maintaining superior quality, reliable measurements of consumers' sensory perception and discrimination are becoming more critical. Consumer discrimination methods using a preferred-reference duo–trio test design have been shown to be effective in improving the discrimination performance by customizing sample presentation sequences. However, this design can add complexity to the discrimination task for some consumers, resulting in more errors in sensory discrimination. The objective of the present study was to investigate the effects of different types of test instructions using the preference-reference duo–trio test design where a paired-preference test is followed by 6 repeated preferred-reference duo–trio tests, in comparison to the analytical method using the balanced-reference duo–trio. Analyses of d' estimates (product-related measure) and probabilistic sensory discriminators in momentary numbers of subjects showing statistical significance (subject-related measure) revealed that only preferred-reference duo–trio test using affective reference-framing, either by providing no information about the reference or information on a previously preferred sample, improved the sensory discrimination more than the analytical method. No decrease in discrimination performance was observed with any type of instruction, confirming that consumers could handle the test methods. These results suggest that when repeated tests are feasible, using the affective discrimination method would be operationally more efficient as well as ecologically more reliable for measuring consumers' sensory discrimination ability. [ABSTRACT FROM AUTHOR]

To achieve more efficient decision making for many business objectives including reformulations and processing changes in the food industry, the important questions become 1) what is the best way to measure consumer discrimination and 2) how much discrimination is meaningful in consumer preference that can be used as a business action standard for food quality maintenance. The purpose of the present paper was to develop a consumer sensory discrimination methodology using affective reference framing to provide more accurate information on both questions. For such affective methods, two versions of repeated duo–trio tests using the preferred sample as reference were implemented: Affective group I—each of 8 repeated duo–trio tests was preceded by a paired-preference test; and Affective group II—a paired-preference test was performed before each of the two sub-sessions consisted of 6 repeated duo–trio tests and the preferred sample was used as a constant-reference during the sub-session. The discrimination performance of these tests was compared to the analytical balanced-reference duo–trio test. Results indicated that for affective group I only, it was possible to identify the ‘affective discriminators’ who showed genuine preference for one product over another and their sensory discrimination in the affective method was significantly higher than in the analytical method. It suggested that this affective method is a desirable consumer sensory methodology that can reveal insightful consumer-relevant sensory differences by comparing sensory discriminations between the so-called ‘affective discriminators’ and ‘affective non-discriminators’. [ABSTRACT FROM AUTHOR]

[Display omitted] • A novel SDT-based detection threshold method of oral/taste sensitivity to NEFA was proposed. • It used repeated descending-block DR A-Not A tests with a warm-up procedure. • Using this novel method, hypo-sensitive subjects also showed learning over repeated testing. • This resulted in lower variability in group median thresholds. • On the other hand, using the ascending triangle method, hypo-sensitive subjects did not show learning effects. Accumulating psychophysical evidence suggests substantial individual variability in oral/taste sensitivity to non-esterified, long-chain fatty acids (NEFA), which is commonly referred to as fat taste or oleogustus. Recent studies have sought to determine its associations with human factors such as body mass index (BMI) and food preferences, as it has been claimed that excessive fat consumption is related to several health conditions, including obesity. Yet, the findings are controversial. On the other hand, it has been noted that considerable variability also occurs based on the methodology used to measure the fatty acid taste. Specifically, learning effects have been observed over repeated measurements of the detection threshold of NEFA, yet there has been no methodology available to take into account these learning effects. Accordingly, in the present study, a novel methodology using a descending-block dual reminder A-Not A (DR A-Not A) method with a warm-up has been proposed to measure the NEFA detection threshold based on the signal detection theory and considering NEFA taste learning effects over repeated sessions. Homogeneous subjects (young adult Korean females within the normal BMI range, non-vegetarians) were randomized to either the novel descending-block DR A-Not A method or ascending triangle method that is commonly used for fat perception studies. Pure oleic acid emulsions were used as fat taste stimuli to be discriminated from pure mineral water. Each subject completed 14 repeated visits. For the ascending triangle method, 14 thresholds were determined using a stopping rule, while for the novel method, 7 thresholds were determined each per two consecutive days, using a criterion of a lower limit of 50% confidence interval of d' = 0.5, considering the practical aspects of taste studies in food sensory science. Based on the group median results of the last two visits, the variability of the detection thresholds was reduced using the novel descending-block DR A-Not A method due to better learning effects over repeated sessions. This shows the potential of the descending-block DR A-Not A threshold method for further studies on oral/taste sensitivity to fat. [ABSTRACT FROM AUTHOR]