1997 Volume 1 Issue 2 Pages 27-39
In recent years, beer has been distributed on a global scale. Accordingly, if the future of beer is considered, it is important to know how flavor of beer is perceived in various regions of the world.
Yet, as far as we could determine, there are no published results of a taste-preference studies uniformly conducted in Japan, the U.S., and Germany. Hence, we conducted the series of surveys described herein (Ishii et al., 1997a, 1997b). For this purpose, the evaluation method was newly standardized and the evaluation terms were selected in each language.
Expressions for taste in each country are greatly influenced by its foodstuff environment, dietary customs, and food culture. If they are translated in verbal meaning alone, they would be of limited value. Moreover, the specialized flavor terminology heretofore published is not suitable for taste-preference surveys of consumers. Accordingly, for this study the beer flavor elements were defined anew for the first time. Then the evaluation terms in each language were selected so that the flavor elements expressed by them would be surely understood by consumers.
In addition, it was foreseen that differences in national character may make it difficult to compare the survey results, depending on the evaluation method. Therefore the evaluation method was standardized for the three countries to ensure that taste evaluation results would be reliably obtained in each country.
These evaluation terms and this evaluation method enabled us to conduct a three-country comparative study incorporating elements related to dietary habits and food culture. In this report, the evaluation terminology and evaluation method are explained. The results of the study will be reported in detail in another report (Ishii, T. et al., 1997b in press).
The terms were selected to enable Japanese, American, and German beer consumers to understand the flavor characteristics denoted by each of the terms. The flavor elements relating to taste of beer were incorporated into the selection process without omission.
Objective terminology for specialists or professional technicians has been published for evaluation of flavor of beer, including flavor terminology prepared by EBC/ASBC joint committee (Meilgaard et al., 1979). As it is difficult for many consumers to understand, however, these technical terms are not suitable for taste-preference surveys of consumers, as in the case of this study. Hence, in selecting the terms for this study, specialized terminology difficult to comprehend was excluded. Also highly abstract and Impressive expressions were avoided which may be easily understood as words but not directly associated with flavor characteristics. We have primarily intended to select terms adequately enabling someone with a consumer's mentality to understand the flavor characteristics denoted by the terms. Any previously published reports were not found concerning uniformity of terms for evaluation of beer flavor by consumers in Japanese, English, and German. Our attempt to select such terms in these three languages would be a first.
2. Term-Selection ProcessThe Japanese, English, and German terms were selected in accordance with the following process.
1) Preparation of the Japanese version:
Based on the above selection policy, the terms were selected which appropriately and comprehensibly express each flavor element, referring to terms previously used in beer taste surveys. In this selection, flavor elements related to taste of beer were identified from the standpoints of "overall impression" (e.g. deliciousness, smoothness (drinkability)), "strength of flavor" (e.g. aroma, bitterness), various "flavor characteristics of beer" which express taste of beer, and "impressions concerning quality of aroma, bitterness, and sweetness". The participants in this selection process included expert beer-flavor judges and members of the Master Panel, a group of beer flavor evaluation specialists formed within Kirin Beer Co. The adequacy of the selected terms and expressions was examined to check whether anything that should be regarded as a beer flavor element had been omitted and whether the flavor characteristic denoted by each term was understandable.
2) Preparation of the English version:
The above-described Japanese version was translated into English by a native English-speaking, professional translator. This translation was used as a basis to prepare the English version of the terms. The English expressions were repeatedly refined and comprehensibility of each term was confirmed with American beer-aficionados who were on hand sampling beer. Also on hand were an expert beer flavor judge and a translator who specializes in Japanese-to-English translation. In addition, before the use of the English terms, a preliminary test was conducted in New York to check whether any problems existed with respect to their comprehensibility.
3) Preparation of the German version:
The German version was prepared, taking into account of a German translation by a professional translator based on the above English version. Before the German version was finalized, the adequacy of each expression and term was refined by a German beer technician, persons who had lived in Germany, and an expert beer flavor judge. In addition, before the use of the German terms, a preliminary test was conducted in Frankfurt to check whether any problems existed with respect to their comprehensibility.
3. The Beer Taste SurveyThe beer taste surveys were conducted from September to October 1992 in three cities: Tokyo, Japan; New York, U.S.A.; and Frankfurt, Germany.
The sample comprised 600 (200 per city) male, avid drinkers ranging the ages between 20 and 59. Each subject sampled and evaluated three kinds of beer.
In all three countries, the same eight kinds of beer, representative brands from Japan, the U.S., Canada, Germany, Holland, and Czech were used.
The trilingual terms thus selected for evaluating taste of beer are shown in Table 1.
2. Evaluation-Method Standardization Results(1) For each of the five terms for "overall impression", a three-level evaluation scale was adopted.
(Ex.) the initial flavor (deliciousness at first sip)
1. unpalatable 2. not unpalatable but not delicious 3. delicious
(2) For the terms for "evaluation of strength of beer's aroma/flavor", the responses "too strong", "agreeable", and "too weak" were divided into two levels each to enable to more precisely and directly investigate receptivity to strength of aroma/flavor. That is, the following six-level evaluation scale was adopted.
1. Too weak. Not agreeable to my tastes.
2. Somewhat too weak. Not entirely disagreeable but not agreeable either.
3. Agreeable but I wouldn't mind if it was a little stronger.
4. Agreeable but I wouldn't mind if it was a little weaker.
5. Somewhat too strong. Not entirely disagreeable but not agreeable either.
6. Too strong. Not agreeable to my tastes.
(3) For each term for "evaluation of flavor characteristics of beer", a "multiple choice format" was adopted whereby the subjects select whichever choices apply to the beer concerned. For "characteristics of aroma, bitterness, and sweetness", the same type of multiple choice format was adopted.
Tems for Evaluation of Beer's Taste in Japanease, English and German
Below is the summary of the main characteristics of the evaluation terms and evaluation method selected for this study.
1) Overall Impression
"Deliciousness" was selected as the term for overall impression of taste of beer. The overall impression was divided into impressions of "smoothness (drinkability)", "aroma", and "taste" for further survey. Regarding "deliciousness", impression of "initial flavor of beer (deliciousness at the first sip)" cannot be said to be the same as impression of its "flavor upon continued drinking". One's impression may differ between the first and subsequent sips. Yet prior taste surveys have been conducted without particular regard for this point.
In our survey, the impression of initial flavor of beer felt at the first sip was designated as the "deliciousness at the first sip", and the impression felt at the conclusion of the taste survey, after a number of sips of the beer had been taken, as "the flavor upon continued drinking", by which both evaluation were distinguished each other.
2) Evaluation of the Strength of Aroma and Flavor of Beer
How one perceives the strength of aroma and flavor of beer is a key determinant of one's impression of "deliciousness" of the beer concerned. At first the strength of "aroma" and "taste" of beer was investigated, and then the strength of specific aspects of the flavor. For these aspects, three of the four basic tastes were used: "bitterness", "sourness", and "sweetness" to which "aftertaste" were added (excepting "saltiness", the fourth basic taste, considered to be of little importance in the case of beer).
We also aimed to more precisely and directly investigate consumers' taste-preference reactions to the intensity of flavor and aroma. That is, not a semi-objective evaluation in terms of "strong", "regular", and "weak" but the expressions "too strong", "agreeable", and "too weak" were used for our survey to evaluate consumers' (subjects') subjective receptivity. The intensity evaluation was conducted with a six-level evaluation scale. To find the point of optimal receptivity of intensity, "agreeable" was divided in two: "agreeable but I wouldn't mind if it was a little weaker", and" agreeable but wouldn't mind if it was a little stronger".
3) Evaluation of Flavor Characteristics of Beer For thorough inclusion of all flavor elements related to taste of beer, included were elements such as the stimulative sensation of carbonation, perception of alcohol content, and textural sensations such as "a clean, water-like taste on the palate" and "a creamy, smooth taste on the palate".
"Freshness" and "dryness" were included, as well as "pungency", the distinctive stimulative sensation peculiar to alcoholic beverages, which was denoted with the term "a sharp and piquant flavor".
Moreover, for the English translation of "koku/ajiwai ga aru", "good body" was used because "koku/ajiwai" is not simply richness of flavor; rather, it is an agreeable quality which is favorably received.
Though difficult to translate into the other languages, "karakuchi" was included in the Japanese version, which seems to be already generally recognized quality of beer.
4) Evaluation of Characteristics of Aroma, Bitterness, and Sweetness
For example for "evaluation of bitterness characteristics", several terms which express "bitterness characteristics" were selected in order to investigate with greater precision how consumers perceive bitterness. And adoption of a format which requires a separate evaluation response for "aroma", "bitterness", and "sweetness", precluded the possibility of not obtaining evaluations for a basic flavor to which consumers in a given country are largely indifferent.
Based on the evaluation results obtained in our beer taste study, the adequacy of the terms and evaluation method selected for this study are examined hereinbelow.
1. Overall ImpressionEach of the five terms for "overall impression" was surveyed respectively at a threelevel evaluation scale. Accordingly, in order to compare the level of response to the terms, the evaluations for each term were classified into positive evaluations and negative evaluations to calculate "term-selection rates" .
That is, the term-selection rates were calculated by regarding the positive and negative responses for each term as two separate evaluation terms. For example, in the case of the term "the initial flavor", responses of "1. Unpalatable" were defined as the term "the initial flavor is unpalatable" and responses of "3. Delicious" as "the initial flavor is delicious".
Table 2 shows by country average values of the term-selection rates and standard deviations for the eight kinds of beer used in our study.
In Germany and the U.S., the selection rate for positive expressions was overwhelmingly high for all the terms. By contrast, in Japan, only in the case of "the initial flavor" and "smoothness (drinkability)", the selection rate for the positive expression ("the initial flavor is delicious" and "smooth", respectively) was high. For all the other terms, the selection rate for the negative expression was higher than that for the positive expression.
The terms for "overall impression" are intended to investigate taste preferences with respect to the eight kinds of sample beers. For nearly all of these terms, the overall standard deviation among the three countries is lower than the standard deviation for each country. This indicates that taste preferences with respect to the eight sample beers differ among Japan, the U.S., and Germany.
Multiple regression analysis for each country revealed that the evaluation terms for "flavor characteristics of beer" which were correlated with "the flavor upon continued drinking" differed markedly among the three countries--particularly, between the U.S. and Germany. For example, "strong hops flavor" had a positive correlation to "the flavor upon continued drinking" in Germany but a negative correlation in the U.S., while "a clean, water-like taste on the palate" and "mild fizziness" had a positive correlation in the U.S. but a negative one in Germany. In short, this indicates that the flavor characteristics perceived as "delicious" differ considerably among Japan, the U.S., and Germany.
Additionally, "deliciousness" was evaluated this time by distinguishing between "the initial flavor" and "the flavor upon continued drinking". In order to investigate the differences be tween these two terms, the simple correlation between the terms and the selection rate for each of the evaluation terms for "flavor characteristics of beer" was analyzed by country. The differences between the two terms were not pronounced; rather, they differed in only a small number of instances. For example, in the case of Japan, for the two terms "fresh flavor", and "exquisite aromatic flavor", a positive correlation with "the flavor upon continued drinking" was admitted while any correlation with "the initial flavor" was not admitted.
Proportions of Choice of Terms for “Overall Impression”: Mean(%)and Standard Deviations(%)for the Entire Sample
For the six terms for evaluation of the strength of aroma and flavor of beer, the average values of the evaluation results for the eight beers used in this study are shown in Fig 1. by country.
It can be seen that more Americans than Japanese and Germans perceived "aroma", "taste", "bitterness", "sourness", and "aftertaste" as being "too strong". "Sweetness" was the only characteristic that relatively many Americans found to be "too weak".
It can be seen that the Japanese, in comparison with the Germans and Americans, have a broad range of receptivity to strength of aroma/flavor with respect to all six of the applicable evaluation terms.
In the evaluation of the strength of aroma and flavor of beer, it was possible to ascertain through the evaluation method employed this time that the degree of aroma/flavor perceived as "agreeable" varies greatly among the three countries of Japan, the U.S., and Germany.
Evaluation of strength of beer's aroma and flavor
Of the various terms for "flavor characteristics of beer", the ones judged as consistent with taste impressions are selected. Accordingly, it can be said that the more frequently selected terms (terms with high selection rate) are "highly reactive" terms, which are widely accepted as expressions of taste of beer.
It also can be said that the magnitude of the average value of selection rate of terms with respect to the eight kinds of sample beers indicates the degree of reactivity of terms as an expression of flavor of beer. In contrast, it can be said that the magnitude of the standard deviation expresses dispersion of the selection rate of the terms among the eight types of beer to indicates whether the term effectively identified taste differences among the eight sample beers used this time.
Figs. 2～4 show by country (Japan, the U.S., and Germany, respectively) the average values and standard deviations of the term-selection rates for the eight kinds of beer used in this study.
Terms with both high average selection rate and large standard deviation among the eight sample beers are widely understood as expressions of taste of beer, as well as useful for investigation of taste differences. Terms with low average selection rate but large standard deviation among the eight samples are useful for investigating taste differences because, despite low reactivity, these terms have gained common recognition as expressions of taste of beer. Conversely, terms with small standard deviation have not gained such common recognition, even if they have high average selection rate. Hence, it cannot be said that they are useful terms for investigation of taste differences.
The terms which proved to be comparatively useful for investigating taste differences in all three countries include "clean, water-like taste on the palate", "strong hop flavor", and "light and sprightly". In Japan, other useful terms were "mild and smooth on the palate", "refreshing and exhilarating", and "karakuchi". In the U.S., other useful terms included "a sharp and piquant flavor", "mild and Smooth on the palate", and "strong malt flavor".
"Dry" had both low average selection rate and small standard deviation in all three countries. It accordingly can be said that it is not useful as a beer-taste term.
Based on the results of simple-correlation analysis, "sharp piquant flavor" was found to have positive correlation to "sharp fizziness" and "strong hop flavor" in the U.S. and Germany, while it positively correlates to "karakuchi" in Japan.
"Karakuchi", which was included only in the Japanese version of the evaluation terms, had positive correlation to "strong hop flavor" and "strong malt flavor", and negative correlation to "exquisite aromatic flavor" and "mild and smooth on the palate".
Evaluative-term selection : Japan
Evaluative-term selection : U.S.A
Evaluative-term selection : Germany
With regard to the evaluation of characteristics of aroma, bitterness, and sweetness, the standard deviation of selection rates of the evaluation terms for the eight kinds of beer used in this survey are shown in Fig. 5 by country.
The way in which "aroma", "bitterness", and "sweetness" are perceived differs considerably among the three countries.
In the case of evaluation of aroma characteristics, differences in beer aroma were identified by means of the aroma characteristic "fruity" in Japan, "spicy" in Germany, and "yeasty" and "sulfur-like" in the U.S. In the evaluations of bitterness and sweetness characteristics as well, it was observed that the way in which differences in the characteristics are perceived differs by country.
Evaluation of characteristics of aroma, bittemess, and sweetness
Above we have discussed the adequacy of evaluation method and evaluation terms selection as indicated by the research results. Based on the results of the study, we can conclude that despite the problems were contained such as differences in understanding stemming from the selected taste-terms expressed in each language, the uniform evaluation terms and evaluation method selected this time were effective for finding differences in not only beer taste-preferences but also the beer taste elements considered important in each country, as well as the existence of differences in taste perception among the three countries.
Our sincere thanks go to those who cooperated with and supported us for this study --especially to Dr. Kiyoshi Yoshizawa, who enthusiastically provided guidance throughout the process.
