- Email: [email protected]
The Effect of Flavor Principles on Willingness to Taste Novel Foods
Appetite, 1999, 33, 209–221 Article No. appe.1999.0263, available online at http://www.idealibrary.com on
The Effect of Flavor Principles on Willingness to Taste Novel Foods
C. STALLBERG-WHITE and P. PLINER Erindale College and Ontario Institute for Studies in Education, University of Toronto
E. Rozin and P. Rozin have suggested that one of the functions of “flavor principles” (the distinctive seasoning combinations which characterize many cuisines) is to facilitate the introduction of novel staple foods into a culture by adding sufficient familiarity to decrease the neophobia ordinarily produced by a new food. We tested this idea experimentally, predicting that the addition of a familiar flavor principle to a novel food would increase individuals’ willingness to taste it, in comparison to their willingness to taste the same food in the absence of the flavor principle. Since people have little reluctance to approach familiar foods, addition of a familiar flavor principle to a familiar food should have little effect on willingness to taste it. In a pilot study, subjects selected from a list, a sauce which was high in familiarity and liking for them. They then rated their willingness to taste one novel and one familiar food with the sauce and one of each with no sauce. Subjects did not actually see any foods—the familiar and novel foods were simply described—and they were aware that they would not actually be tasting any foods. In the study proper, subjects rated their willingness to try each of the four food/sauce combinations described above. In this study, they actually saw the foods they were rating and were under the impression their ratings would determine what they would taste later in the study. In both studies, the addition of a familiar sauce to a novel food increased subjects’ willingness to taste it (in comparison to the same food with no sauce) while the addition of a sauce to a familiar food either had no effect or decreased subjects’ willingness to taste it. In the pilot study, the “flavor principle” effect interacted with subjects’ levels of food adventurousness. 1999 Academic Press
I The reluctance to approach novel foods exhibited by omnivorous animals (including humans) is assumed to protect them from the consumption of toxic materials (P. Rozin, 1984; Wong, 1995). It is a conservative force, operating “to keep the organism’s taste preferences from straying” from familiar foods and to keep its feeding behavior “locked in on a safe track” (Schulze & Watson, 1995; p. 230). Although acceptability is not perfectly correlated with familiarity and there are many examples of familiar foods which are not liked, much research on the “mere exposure” This research was supported by a grant from the Social Services and Humanities Research Council of Canada. Address correspondence to: Patricia Pliner, Erindale College, University of Toronto, 3359 Mississauga Rd., Mississauga, Ontario, Canada, L5L 1C6. 0195–6663/99/050209+13 $30.00/0
1999 Academic Press
210
C. STALLBERG-WHITE AND P. PLINER
phenomenon indicates that both liking and acceptability do vary directly with exposure (e.g. Zajonc, 1968). At the same time, in order to capitalize on the advantages of omnivory, the omnivore must be willing to sample unfamiliar foods. Thus, the omnivore’s dilemma (E. Rozin & P. Rozin, 1981) is that it has good reason both to approach and to avoid novel foods. Whether the conflict in any particular instance is resolved in favor of approach or avoidance will depend on factors both internal and external to the organism. A variety of factors have been shown to increase, at least temporarily, the likelihood that humans will approach, rather than avoid, novel foods. Decreased arousal levels (Pliner et al., 1995; Pliner & Melo, 1997), information about novel foods (Martins et al., 1997; McFarlane & Pliner, 1997; Pelchat & Pliner, 1995), observation of a neophilic model (Harper & Sanders, 1975; Hobden & Pliner, 1995), and increased familiarity with a particular novel food or with novel foods in general (Birch & Marlin, 1982; Birch et al., 1987; Pliner et al., 1993; Tuorila et al., 1994) have all been shown to decrease neophobic responses. Another factor known to influence level of food neophobia in humans is the amount of familiarity in the general situation. Pelchat (unpubl.) found that the more familiarity contained in an eating situation (eating food offered by a vendor in the subways vs. a psychologist in the laboratory vs. a friend in the friend’s home), the more willing subjects were to eat an unfamiliar food. Increases in familiarity can also reduce children’s neophobia. In a study conducted by Harper and Sanders (1975), children, even in their own homes, were more likely to accept a novel food from their mothers than they were from the unfamiliar experimenter. Although the mother represents more than simply familiarity, one interpretation of these results is that adding familiarity to the situation appears to reduce food neophobia. People within a variety of cultures appear to have adopted a version of this strategy to encourage the acceptance of novel foods into the culture. This version involves the ubiquitous usage of combinations of seasonings, referred to by E. Rozin (1983) as flavor principles and described as one of the factors that serve to transform foods into culturally distinctive cuisines. Examples of flavor principles would include the tomato–garlic–olive oil combination which characterizes Italian cuisine or the soy sauce–rice wine–ginger mixture which typifies Chinese cooking. In some instances, these general flavor principles are further differentiated into more specific regional principles. Thus, in Cantonese cooking, garlic and black beans are often added to the more general flavor principle, while in the cooking of Szechuan, a “hot” ingredient such as brown pepper or some form of chilli is added. To a great degree, the flavor principles “make” Italian food, Italian, or Chinese food, Chinese, or Cantonese food, Cantonese. Flavor principles are commonly employed in many different dishes of a cuisine from soups and salads to grains, meats and vegetables. E. Rozin and P. Rozin (1981) note that individuals have a deep attachment to their culture’s characteristic seasonings and find it difficult to imagine food prepared without them. Food simply doesn’t taste good without the appropriate flavor principle. “Traditional flavorings are high-priority culinary items and immigrant groups typically go to great lengths and expense to procure them in foreign settings” (p. 6). E. Rozin and P. Rozin (1981) hypothesize that one of the main functions of flavor principles is to add sufficient familiarity to novel foods to decrease neophobia, thereby facilitating their introduction into a culture. Anecdotal evidence illustrates this supposition. As a part of efforts in the early twentieth-century by home economists and other food specialists to increase variety into the Japanese diet by adopting
EFFECT OF FLAVOR PRINCIPLES
211
foods from abroad, cookery columns of women’s magazines encouraged middleclass housewives to try Western dishes. Among the strategies employed to facilitate this was the addition of Japanese seasonings to make the dishes more familiar (Cwiertka, 1998). More recently, E. Rozin and P. Rozin (1981) asked women in a Mexican village to imagine serving three novel foods (alfalfa, sardines and a common shrub). The women could do so only after picturing each food seasoned with their traditional flavor principle, consisting of chilli and lime. The study which follows was designed to examine experimentally the neophobiareducing function of flavor principles. Thus, an important task was to decide how exactly to operationalize the flavor principle construct. In order to do so, we first attempted to abstract from the notion of flavor principles, a cultural concept, their important elements so we could translate these into the concrete operations of the psychology laboratory. E. Rozin’s (1983) description suggests that a flavor principle is pervasive to the foods of a particular culture, is eaten by members of the culture frequently, typically many times a day, and is strongly liked by members of the culture. Thus, flavor principles are ubiquitous, familiar and palatable. In addition, because they are habitually eaten in association with particular foods, they must have some degree of “combinatorial” appropriateness with these foods. Of course, these characteristics of flavor principles are not independent. For example, familiarity and liking are known to be highly correlated (Birch & Marlin, 1982; Zajonc, 1968), and ubiquity and combinatorial appropriateness are presumably also related. That is, the more culinary situations in which a flavor principle appears, the more foods with which it is likely to be combined and, therefore, the greater its range of appropriateness. However, there must be limits on the extent of a flavor principle’s combinatorial appropriateness. A general limit can be seen in the fact that flavor principles are not often used in beverages or sweet dishes (E. Rozin, 1983). For the purposes of the present study, we decided that two of these characteristics of flavor principles, familiarity and palatability, were most important. Our goal was to test the hypothesis that the presence of a flavor principle could provide enough dietary familiarity to induce individuals to try an unfamiliary food; thus, familiarity appeared to be a characteristic crucial for inclusion. Because familiarity and palatability are highly correlated and because palatability is such an important aspect of food choice, we also chose to incorporate it into our operational definition of the flavor principle construct. We made the decision not to include ubiquity and appropriateness in our operational definition of flavor principles. Although ubiquity is a characteristic of flavor principles in general, this seemed less important to our specific goal of using them to decrease neophobia toward particular novel foods. Appropriateness also seemed less important, since with novel foods, individuals would presumably not have well-defined ideas of the appropriateness of the combination. With these considerations in mind, we decided on an approach in which participants, unselected for their membership in any particular cultural group, would chose for themselves “flavor principles” which were individually appropriate. We put the term in quotation marks at this point to indicate that we are now referring to our experimental analogues of flavor principles. In the experiment to be described, these consisted of common and readily available sauces and condiments for which participants indicated high degrees of familiarity and liking. Thus, the “flavor principles” used in our study are analogous to what E. Rozin (1983) describes in that they incorporate the characteristics of familiarity and palatability. However, even on these dimensions, they are obviously far from perfect replicas of flavor
212
C. STALLBERG-WHITE AND P. PLINER
principles; they must be much “weaker” than real ones since it is unlikely that our subjects eat them as frequently or feel as positively toward them as would be the case with traditional seasonings. The basic prediction of the study is that the addition of a familiar and liked “flavor principle” to an unfamiliar food will increase individuals’ willingness to taste it, in comparison to their willingness to taste the same novel food in the absence of the flavor principle. Since individuals have little reluctance to approach familiar foods, addition of a familiar flavor principle to a familiar food should have little effect; thus, we expect an interaction between the novelty of the food and the presence of a flavor principle. We expected that the addition of the flavor principles would increase individuals’ willingness to try novel foods. We were not sure, however, how the addition of the flavor principles would change the novel foods or individuals’ perceptions of them to produce increased willingness. In other words, we wondered how flavor principles might mediate changes in willingness. At least two different mediators seemed possible. We introduced the flavor principle notion originally in the context of research which implies that increases in familiarity are likely to be the mediator; that is, the addition of an extremely familiar sauce could increase the familiarity of the novel food, causing its anticipated taste or appearance to be more familiar. This increased familiarity would lead to greater willingness. However, it also seemed possible that hedonic changes could be important. That is, the addition of a hedonically positive sauce could increase the hedonic aspects of the novel food, causing it to (be expected to) look or taste better. Some of our dependent measures attempted to clarify this issue. Finally, we assessed participants’ degree of adventurousness with respect to food in two different ways. Pliner and Hobden (1992) have demonstrated that there are stable individual differences in the extent to which individuals report themselves willing to approach novel foods and in the extent of experience with exotic foods. A previous study testing the effectiveness of a modeling manipulation to reduce food neophobia found that it was differentially effective, depending on subjects’ levels of food neophobia (Hobden & Pliner, 1995). In the present study, we examined the extent to which the flavor principle manipulation interacted with food adventurousness, as assessed by the Food Neophobia Scale (FNS) and by a measure of experience with various ethnic cuisines, shown in a previous study to be negatively related to the FNS (Pliner & Hobden, 1992).
M Pilot Study Before conducting the study proper, we conducted a questionnaire-based pilot study which had two main purposes. The first was to present participants with a fairly extensive list of sauces so we could determine which were rated as most familiar and highest in palatability; these would comprise the flavor principles to be made available for choice in the study proper. The second purpose of the pilot study was to obtain some preliminary data to test our flavor principle hypothesis. The study was conducted by students in an upper-level undergraduate Psychology course who administered questionnaires to 98 participants (primarily students) with a mean age
EFFECT OF FLAVOR PRINCIPLES
213
of 26·8. The first page of the questionnaire listed 19 sauces (e.g. Italian tomato sauce, teriyaki sauce, beef gravy) and requested ratings (on 9-point scales) of familiarity and liking. Once a participant had completed this page, the experimenter briefly examined it, and used the ratings to select as a flavor principle the sauce with the highest familiarity rating (and, if there were more than one, the one of these with the highest liking rating). After selecting this flavor principle, the experimenter wrote its name in the appropriate place on the second page of the questionnaire, as described below. This sauce is referred to subsequently as the participant’s “chosen” flavor principle. The second page of the questionnaire listed and described four foods, two novel and two familiar. Following the name and description of each of the first two foods (one familiar, one novel) was the question (answered on 9-point bipolar rating scales) “how willing would you be to taste a small amount of (food)?” For the third and fourth foods, the willingness question read “how willing would you be to taste a small amount of (food) with ——?” Written into the blank was the name of the particular participant’s chosen flavor principle. Subjects also completed two measures of their adventurousness with respect to food. One requested ratings of the frequency of eating foods from 10 relatively exotic (e.g. Korean, Vietnamese) and not-so-exotic (e.g. Chinese, Italian) cuisines, and the other was the Food Neophobia Scale (Pliner & Hobden, 1992). The results indicated that our list of sauces was sufficiently diverse to accommodate nearly all subjects. For 90%, the chosen flavor principle had been rated as “9” on the familiarity dimension and for 95%, it had been rated as “8” or “9” (M=8·84). Furthermore, subjects liked their chosen flavor principles (M=8·81). More important, the results of the pretest indicated that by having available a relatively small number of sauces in the study proper, it would be possible to provide a sauce high in familiarity and palatability for nearly all participants. The five sauces most frequently chosen as participants’ flavor principles according to the criteria above represented those chosen by more than 60% of participants. For the remaining participants, whose chosen flavor principle was not one of the top five, at least one of the top five received high liking and familiarity ratings. Thus, by having available only the top five sauces in terms of their rated familiarity, it would be possible to provide a familiar and liked flavor principle for participants in the study proper. To test the flavor principle hypothesis with our pilot data, we examined the participants’ reported willingness to taste the four foods (two novel, two familiar; one of each with, and one without, the chosen flavor principle) described in their questionnaires. We anticipated that subjects would be generally less willing to taste novel (vs. familiar) foods; the addition of a familiar flavor principle to the novel, but not the familiar, food should increase willingness to taste it. Thus, results in accordance with the prediction would appear in the form of an interaction between the two (within subjects) independent variables, novelty of food and presence of sauce. As expected, there was a highly significant main effect of novelty of food, F(1,99)=67·05, p<0·001; participants were more willing to try the familiar than the novel food (means: 7·8 and 5·9, respectively). However, the predicted interaction did not approach significance. In a second analysis, we characterized participants as relatively more or less experienced with respect to exotic foods, by dividing them at the median in terms of their frequency of eating the five least commonly eaten of the 10 cuisines desribed above. We then added status (above or below the median) on this new variable as a between subjects variable in a 2×2×2 analysis which also included novelty of food and presence of sauce as above. In addition to a novelty
214
C. STALLBERG-WHITE AND P. PLINER
main effect, there was a significant interaction involving all three variables, F(1,98)= 7·06, p<0·01. Examination of the means revealed the expected pattern for the participants below the median on our food experience measure: adding a sauce to a novel food increased willingness to taste it but had no effect on willingness to taste a familiar food (means: novel–no sauce=5·34, novel–sauce=6·11; familiar–no sauce=7·53; familiar–sauce=7·66). The pattern for the participants above the median on food experience was very different, showing no greater willingness to try a novel food with than without a sauce (means: novel–no sauce=6·28; novel–sauce=6·00; familiar–no sauce=8·02; familiar–sauce=8·25). When the data for the two groups of participants were analysed in separate analyses, there was a novelty by sauce interaction for those below the median in experience, F(1,44)=4·63, p<0·05, while for those above the median, the interaction did not approach significance (F<1). Contrasts on the means revealed that the relatively unexperienced individuals were significantly more willing to try a novel food with than without a sauce, t(98)=3·26, p<0·01. For the same participants, adding a sauce to a familiar food had no effect on willingness, and for relatively highly experienced participants, addition of a sauce had no effect on willingness to try familiar or novel foods. Thus, the flavor principle hypothesis received some empirical confirmation, albeit only for participants who were below the median on one of our food adventurousness measures. When we did a similar analysis, dividing participants at the median on the FNS, we did not obtain the same interaction. However, based on the results with the experience variable, we decided to proceed to the experiment proper. Participants Participants were 44 female and 20 male undergraduate students enrolled in an Introductory Psychology course, who participated for credit. They ranged in age from 18 to 33 with a mean of 20·3 years (SD=2·6). They were unselected for cultural group membership, and we did not collect information relevant to this characteristic. Overview Participants were led to believe at the time of recruitment that they would be taking part in an experiment designed to determine how various food combinations influence flavor-intensity perception. The experimental session consisted of three phases: (1) administration of questionnaires (questionnaire phase); (2) tasting and rating for flavor intensity either one or two novel sauces which were combined with a series of familiar foods (tasting phase); and (3) assessment of willingness to try novel and familiar foods alone or in combination with a flavor principle selected by the participant (rating phase). As in the pilot study, the main independent variables were novelty of food and presence of sauce and the dependent variable was willingness to taste. Procedure Questionnaire phase The purposes of the questionnaire phase were to select flavor principles for participants and to measure their degree of food adventurousness. They completed
EFFECT OF FLAVOR PRINCIPLES
215
a questionnaire on the pretext that “. . . eating habits, especially experience with certain cuisines, are known to influence flavor intensity perception . . .”. Participants rated the 19 sauces used in the pilot study for liking and familiarity, using 9-point bipolar scales (endpoints “not at all” and “extremely”) and rank ordered their top five sauces in terms of liking. The experimenter used this information to select a highly familiar and liked flavor principle for each participant for use during the willingness phase of the experiment. As noted earlier, the pilot data enabled us to determine the most familiar and preferred sauces for this population, and we kept supplies of the top-rated sauces on hand in the laboratory. Flavor principles were then selected according to the following criteria: (1) if the participant’s top-ranked sauce was among those on hand, choose that; (2) if not, choose the next ranked sauce. Using these criteria, we were able to accommodate all participants with the sauces on hand. On another section of the questionnaire, participants rated (9-point scales) their frequency of eating foods from 10 cuisines, as described in the pilot study; this provided the measure of experience with exotic cuisines. Finally, participants completed the Food Neophobia Scale (Pliner & Hobden, 1992). Tasting phase The purposes of the tasting phase were primarily to make it clear to participants that the study involved the actual tasting of foods and to provide a context in which we could subsequently ask them about their willingness to taste various foods both in combination with a sauce and alone. In the tasting phase, participants ingested small samples of familiar foods (such as rice, potato, celery and corn chip) both alone and accompanied by one or two commercially available cooking sauces (Creole Sauce or Ginger Tamari Sauce, both Master’s Choice brand), described as “flavor mediums”. In keeping with the cover story, participants made ratings of flavor intensity (9-point bipolar scales with endpoints labeled “extremely bland” and “extremely intense”) after they tasted the foods, both alone and in combination with the sauces. Willingness phase The main purpose of the willingness phase was to measure the dependent variable: willingness to taste novel and familiar foods both with and without flavor principles. To accomplish this, we led participants to believe that the experiment was over and that we were now pretesting foods for future use. They were told: “. . . that’s all the food I need you to taste for the experimental part of the session. What I’d like to do now is to get your opinion regarding a variety of foods and flavor mediums that we are thinking of using in another part of this study. This time, though, you won’t have to try any of the samples I’m about to show you, unless you are willing to do so.”
The set of foods used in the willingness phase consisted of four novel and four familiar foods. The novel foods were parval (an Indian vegetable), gathiya (an Indian snack chip), tapioca stick (an Asian noodle made from tapioca) and hominy (hulled corn kernels with the germ removed); the familiar foods were raw carrot, taco chips, cooked spaghetti and corn niblets. Participants were presented with one of the novel and one of the familiar foods with their chosen sauce and different novel and familiar foods with no sauce. Different participants received different foods, and the same food was presented to equal numbers of participants with and without a sauce. They
216
C. STALLBERG-WHITE AND P. PLINER
T 1 Willingness to try foods as a function of novelty of food and sauce condition in the study proper Food condition Sauce condition Plain Sauced
Familiar
Novel
8·19 (1·38) 64 7·06 (2·27) 64
6·03 (2·48) 64 6·48 (2·17) 64
rated their willingness to taste each of the four foods presented, with the understanding that their ratings would determine what they actually did taste. They also rated anticipated liking and anticipated familiarity for each. The food samples were displayed in small, clear plastic containers and were covered with clear, plastic lids, so that they could be seen but not smelled. The sauces were displayed in small plastic cups, also with lids. The samples remained behind a partition until they were shown. The names of the foods and sauces were announced as they were presented, and their names were also written on the lids of their containers. So, for example, when foods were presented in the “flavor principle” (or with sauce) condition, the sauce container was placed on top of the food container, and the two were shown together and described as “—— (food) with —— (sauce)”. At the beginning of the willingness phase, participants were told that they would be given fresh samples of the foods they were willing to taste, “. . . not these old pretest samples”. Just before food presentation began, participants were reminded that, when they completed the ratings, their responses would be analysed to see which of the foods and/or food and food–sauce combinations the experimenter would prepare for them to try.
R Willingness Measure Preliminary analyses revealed no effects of sex of participant even approaching significance; accordingly, data for males and females were combined for analysis. The main analysis was a 2×2 ANOVA on participants’ willingness ratings, with novelty of food and presence of sauce as within subjects independent variables. The results of the analysis revealed a significant effect of food novelty, F(1,60)=35·67, p<0·001; participants were more willing to try the familiar foods than the novel ones (means: 7·6 and 6·3, respectively). The only other significant effect was a significant novelty of food by presence of sauce interaction, F(1,60)=14·61, p<0·001. The means are presented in Table 1, where it can be seen that the addition of participants’ flavor principles made them more likely to taste a novel food, t(60)=1·60, p=0·06,
217
EFFECT OF FLAVOR PRINCIPLES
T 2 Expected liking and expected familiarity as a function of novelty of food and sauce condition in the study proper Food condition Measure Liking Familiarity
Sauce condition
Familiar
Novel
Plain Sauced Plain Sauced
7·7 6·3 8·8 8·6
5·0 5·7 1·8 2·8
and less likely to taste a familiar food, t(60)=3·87, p<0·01, in comparison to the unsauced versions. Our second analysis was a 2×2×2 ANOVA in which participants’ status on the food experience variable (above or below the median) was added to the familiarity and sauce variables. In this analysis, the effects of novelty and novelty by sauce obtained in the original analysis remained significant; however, the food experience variable did not interact with the familiarity and sauce variables. A similar analysis, with FNS scores added as a between subjects variable, also yielded significant effects of novelty and a significant novelty by sauce interaction but did not yield a significant second order interaction. Thus, in this study, unlike the pilot study, a flavor-principle effect was obtained for all subjects, not just those low in experience with exotic cuisines. Anticipated Liking and Familiarity We performed 2 (novelty of food) ×2 (presence of sauce) within subjects ANOVAs on participants’ reports of anticipated liking for and anticipated familiarity with the taste of the foods/food–sauce combinations. The analysis on liking yielded a main effect for novelty of food, F(1,59)=65·74, p<0·001. Participants had higher anticipated liking for familiar foods than for novel foods (means: 7·0 and 5·3, respectively). There was also a significant novelty by sauce interaction F(1,59)=26·54, p<0·001. Means for this interaction reveal that the effect of the addition of the sauce on anticipated liking depended on whether it was offered with a novel or a familiar food. Participants anticipated greater liking for a novel food when it was served with their own sauce than when it was served plain (means: 5·7 and 5·0, respectively), t(59)=2·59, p<0·05; the reverse occurred for a familiar food (means: 6·3 vs. 7·7), t(59)=4·67, p<0·01. With anticipated familiarity as the dependent variable, the same two effects emerged; participants rated familiar foods as significantly more familiar than novel foods, F(1,60)=790·78, p<0·001 (means: 8·7 and 2·3, respectively). This finding provides a check on our novelty manipulation; the foods we have been referring to as novel or familiar were perceived accordingly by our participants. There was also an interaction between novelty and presence of sauce, F(1,60)= 10·64, p<0·01; novel foods presented in combination with a familiar sauce seemed more familiar than when they were offered plain (means: 2·8 and 1·8, respectively), while for familiar foods there was very little difference (means: 8·6 vs. 8·8, respectively).
218
C. STALLBERG-WHITE AND P. PLINER
In order to examine the extent to which the flavor principle effect is mediated by changes in expected familiarity and/or palatability produced by the addition of the flavor principle, we did the following analysis. We computed scores for each participant consisting of the differences (for both the hedonic and familiarity ratings) between the sauced and non-sauced versions of the novel foods. These scores presumably represent how much more (or less) familiar/palatable the sauced version is than the unsauced version. We then regressed these scores on participants’ willingness to taste the sauced version of the novel food, using a stepwise procedure. Thus, in a sense, we are examining the rated “consequences” of adding flavor principles to novel foods as predictors of willingness to taste them. The analysis revealed that the hedonic difference score was the only significant predictor of participants’ willingness to taste the novel food, t=4·69, p<0·001, accounting for 28% of the variance (multiple R=0·52). Thus, the more that a participant expected the sauced version of the novel food to be more palatable than the unsauced version, the more willing he/she was to try the novel sauced food.
D We were able to provide evidence in favor of the flavor principle hypothesis; that is, the addition of a familiar and liked sauce to a novel food increased willingness to try it. This effect occurred in the pilot study, in which participants answered questions pertaining to their hypothetical willingness, only for the participants with less experience with the exotic ethnic cuisines, and in the study proper, in which participants were under the impression that their willingness ratings would determine what they would actually taste, for all participants. It is possible that when food choice is only hypothetical, individuals who are relatively adventurous/experienced with novel foods make neophilic choices, but when the choices become real, even these individuals need the added familiarity of a flavor principle to induce neophilic behavior. There is one respect in which the data from the study differed from expectation. We had anticipated that the addition of flavor principles to familiar foods would have no effect on willingness to taste them because individuals are not reluctant to approach familiar foods in the first place. If any effect could be expected, it would be that flavor principles might produce a slight increase willingness to try these foods to the extent that the flavor principles were perceived as more familiar and/or palatable than the foods themselves. What we found, however, was the reverse. The addition of flavor principles to familiar foods actually decreased participants’ willingness to try them. It is at this point that the issue of “combinatorial appropriateness”, mentioned in the Introduction as a characteristic of flavor principles, must be revisited. As mentioned earlier, we quite consciously decided not to include this appropriateness factor in our operational definition of flavor principles on the grounds that, for novel foods, individuals would probably not have well-defined ideas of appropriateness. In addition, the data from the pilot study showed no difference in willingness to try familiar foods with and without a sauce. Furthermore, in that study, as a rough measure of appropriateness, we asked participants to rate (on 9-point scales) the extent to which the addition of the sauce “. . . would clash with [the food’s] flavor”. For both the familiar and the novel foods the mean ratings were relatively low (3·9 and 4·5, respectively) and did not differ significantly. Thus,
EFFECT OF FLAVOR PRINCIPLES
219
it appears that, on the average, participants found the sauces to be appropriately combined with the foods, both novel and familiar. As a result, when we designed the study proper we did not concern ourselves with appropriateness. Given the unexpected results of the the study proper, we decided to use the “clash” ratings from the pilot study to explore further the notion of appropriateness. Although the overall means were low, there were sizeable individual differences in responses to these questions, produced at least in part by the fact that different subjects received different food–sauce combinations. There were two reasons for this: first, different subjects had chosen different flavor principles, and second, participants received, at random, one of eight different novel foods. If participants are divided at the median in terms of their responses to this question about “clashing”, those above the median (i.e. those for whom the pairing of food and sauce was perceived as relatively inappropriate) were less willing to try the familar food with the sauce than without it (means: 7·11 and 7·72, respectively). In contrast, those below the median (i.e. those for whom the pairing was relatively appropriate) were slightly more willing to try the familiar food with the sauce than without it (means: 8·22 and 7·76, respectively). In other words, for familiar foods, appropriate sauces increased willingness slightly while inappropriate ones decreased it. In the case of novel foods, however, the relatively less food adventurous participants (for whom the flavor principle effect occurred) were more willing to taste the novel food with sauce than without, regardless of their ratings of the sauce’s appropriateness (means for appropriate sauces=6·48 and 5·48, respectively; means for inappropriate sauces= 6·36 and 5·00, respectively). In other words, both relatively appropriate and inappropriate sauces increased willingness to try novel foods. We suggest that this occurred for one of two reasons. One possibility is that ideas about appropriateness are less firmly held when a food is novel—what is appropriate is less clear. A second, not incompatible, possibility is that despite inappropriateness, a familiar and liked flavor principle is able to overcome neophobia. That is, it may be precisely the aspects of flavor principles we included in our operational definition that are important for inducing individuals to try novel foods. Appropriateness, although an important characteristic of a flavor principle, may be less important than familiarity and palatability in the context of this particular function. However, there are clearly limits on the “unimportance” of appropriateness; Pliner and Stallberg-White (1998) found that an extremely palatable and familiar, yet highly inappropriate, flavor principle did not increase children’s willingness to taste an unfamiliar food. Clearly, the sauces employed in our studies only approximated flavor principles as they were conceptualized by E. Rozin (1983). And, certainly, flavor principles may have functions other than the neophobia-reducing one tested in our studies. Other possibilities suggested by E. Rozin and P. Rozin (1981) include adding gustatory and/or visual zest to otherwise bland-tasting and/or -looking diets and providing a cultural “marker” in the sense of identifying individuals as members of a cultural group and separating that group from others. And here, aspects of flavor principles not incorporated into our operational definition might be relevant. For example, a particular means of seasoning food could be a more potent cultural marker, the more ubiquitously it appeared in the culinary repertoire of a particular culture. Or if a particular set of seasonings serves to reduce the blandness of a group’s diet, then the seasonings might be most appropriately combined with the bland elements of the diet and less with those with more distinctive flavors of their own. Our point here is to acknowledge that we are aware of the fact that our study utilized what
220
C. STALLBERG-WHITE AND P. PLINER
might be considered to be an incomplete or limited instantiation of the flavor principle construct—what might be considered to be a far cry from the real thing. However, our intention was also limited—namely, to study in the laboratory the neophobia-reducing effects of flavor principles. And for that purpose we chose the aspects of flavor principles which appeared to be most applicable: familiarity and palatability. Thus, although our laboratory experiment did not capture the flavor principle construct in its entirety, it may have captured the elements of it which are most relevant to the neophobia-reducing function of flavor principles. Although humans’ omnivorous heritage inclines them to be neophobic, it is clearly in their interest to be flexible on this issue. Flavor principles afford a means of “certifying” a new food as familiar and safe to eat, thereby providing a cultural solution to the omnivore’s dilemma. This is yet another example of a cultural solution to nutritional problems produced by incompatibility between the available foods and the biological characteristics of humans. Other examples would include techniques for detoxification of toxic but otherwise nutritious foods such as cassava and soybeans (Katz, 1982) and the treatment of foods to free otherwise unusable nutrients (Katz et al., 1974). Thus, flavor principles could be thought of as one of many means by which humans negotiate between their biology and their environment. R Birch, L. L. & Marlin, D. W. (1982). I don’t like it; I never tried it: Effects of exposure on two-year old children’s food preferences. Appetite, 3, 353–360. Birch, L. L., McPhee, L., Shoba, B. C., Pirok, E. & Steinberg, L. (1987). What kind of exposure reduces children’s food neophobia? Looking vs. tasting. Appetite, 9, 171–178. Cwiertka, K. (1998). A note on the making of culinary tradition—an example of modern Japan. Appetite, 30, 117–128. Harper, L. V. & Sanders, K. M. (1975). The effect of adults’ eating on young children’s acceptance of unfamiliar foods. Journal of Experimental Child Psychology, 2, 206–214. Hobden, K. & Pliner, E. (1995). Effects of a model on food neophobia in humans. Appetite, 25, 101–114. Katz, S. H. (1982). Food, behavior and biocultural evolution. In L. M. Barker (Ed.), The psychobiology of human food selection. Westport, CT: Avi Publishing Company, Inc. Katz, S. H., Hediger, M. L. & Valleroy, L. A. (1974). Traditional maize processing techniques in the New World. Science, 184, 767–773. Martins, Y., Pelchat, M. L. & Pliner, P. (1997). “Try it; it’s good and it’s good for you”: Effects of taste and nutrition information on willingness to try novel foods. Appetite, 28, 89–102. MacFarlane, T. & Pliner, P. (1997). Increasing willingness to taste novel foods: Effects of nutrition and taste information. Appetite, 28, 227–238. Pelchat, M. L. (1985). The effect of context on neophobia in college students. Unpublished manuscript. Pelchat, M. L. & Pliner, P. (1995). “Try it. You’ll like it”: Effects of information on willingness to try novel foods. Appetite, 24, 153–165. Pliner, P., Eng, A. & Krishnan, K. (1995). The effects of fear and hunger on food neophobia in humans. Appetite, 19, 77–87. Pliner, P. & Hobden, K. (1992). Development of a scale to measure the trait of food neophobia in humans. Appetite, 19, 105–120. Pliner, P. & Melo, N. (1997). Food neophobia in humans: Effects of manipulated arousal and individual differences in sensation seeking. Physiology and Behavior, 61, 331–335. Pliner, P., Pelchat, M. & Grabski, M. (1993). Reduction of neophobia in humans by exposure to novel foods. Appetite, 20, 111–123. Pliner, P. & Stallberg-White, C. (1998). “Pass the ketchup, please”: Flavor principles increase children’s willingness to taste novel foods. Unpublished manuscript.
EFFECT OF FLAVOR PRINCIPLES
221
Rozin, E. (1983). Ethnic cuisine: The flavor principle cookbook. Lexington, MA: The Stephen Greene Press. Rozin, E. & Rozin, P. (1981). Culinary themes and variations. Natural History, 90, 6–14. Rozin, P. (1984). The acquisition of food habits and preferences. In J. D. Matarazzo, S. M. Weiss, J. A. Herd, N. E. Miller & S. M. Weiss (Eds), Behavioral health: A handbook of health enhancement and disease prevention, pp. 590–607. New York, NY: John Wiley. Shulze, G. & Watson, N. V. (1995). Comments on “Flavor neophobia in selected rodent species”. In R. Wong (Ed.), Biological perspectives on motivated activities, pp. 229–230. Norwood, NJ: Ablex Publishing Corp. Tuorila, H., Meiselman, H., Bell, R., Cardello, A. V. & Johnson, W. (1994). Role of sensory and cognitive information in the enhancement of certainty and liking for novel and familiar foods. Appetite, 2, 231–246. Wong, R. (1995). Flavor neophobia in selected rodent species. In R. Wong (Ed.), Biological perspectives on motivated activities, pp. 229–264. Norwood, NJ: Ablex Publishing Corp. Zajonc, R. (1968). Attitudinal effects of mere exposure. Journal of Personality and Social Psychology Monograph Supplement, 9 (2, Part 2), 1–27. Received 20 January 1998, revision 28 September 1998, accepted in revised form 3 June 1999
The Effect of Flavor Principles on Willingness to Taste Novel Foods
C. STALLBERG-WHITE and P. PLINER Erindale College and Ontario Institute for Studies in Education, University of Toronto
E. Rozin and P. Rozin have suggested that one of the functions of “flavor principles” (the distinctive seasoning combinations which characterize many cuisines) is to facilitate the introduction of novel staple foods into a culture by adding sufficient familiarity to decrease the neophobia ordinarily produced by a new food. We tested this idea experimentally, predicting that the addition of a familiar flavor principle to a novel food would increase individuals’ willingness to taste it, in comparison to their willingness to taste the same food in the absence of the flavor principle. Since people have little reluctance to approach familiar foods, addition of a familiar flavor principle to a familiar food should have little effect on willingness to taste it. In a pilot study, subjects selected from a list, a sauce which was high in familiarity and liking for them. They then rated their willingness to taste one novel and one familiar food with the sauce and one of each with no sauce. Subjects did not actually see any foods—the familiar and novel foods were simply described—and they were aware that they would not actually be tasting any foods. In the study proper, subjects rated their willingness to try each of the four food/sauce combinations described above. In this study, they actually saw the foods they were rating and were under the impression their ratings would determine what they would taste later in the study. In both studies, the addition of a familiar sauce to a novel food increased subjects’ willingness to taste it (in comparison to the same food with no sauce) while the addition of a sauce to a familiar food either had no effect or decreased subjects’ willingness to taste it. In the pilot study, the “flavor principle” effect interacted with subjects’ levels of food adventurousness. 1999 Academic Press
I The reluctance to approach novel foods exhibited by omnivorous animals (including humans) is assumed to protect them from the consumption of toxic materials (P. Rozin, 1984; Wong, 1995). It is a conservative force, operating “to keep the organism’s taste preferences from straying” from familiar foods and to keep its feeding behavior “locked in on a safe track” (Schulze & Watson, 1995; p. 230). Although acceptability is not perfectly correlated with familiarity and there are many examples of familiar foods which are not liked, much research on the “mere exposure” This research was supported by a grant from the Social Services and Humanities Research Council of Canada. Address correspondence to: Patricia Pliner, Erindale College, University of Toronto, 3359 Mississauga Rd., Mississauga, Ontario, Canada, L5L 1C6. 0195–6663/99/050209+13 $30.00/0
1999 Academic Press
210
C. STALLBERG-WHITE AND P. PLINER
phenomenon indicates that both liking and acceptability do vary directly with exposure (e.g. Zajonc, 1968). At the same time, in order to capitalize on the advantages of omnivory, the omnivore must be willing to sample unfamiliar foods. Thus, the omnivore’s dilemma (E. Rozin & P. Rozin, 1981) is that it has good reason both to approach and to avoid novel foods. Whether the conflict in any particular instance is resolved in favor of approach or avoidance will depend on factors both internal and external to the organism. A variety of factors have been shown to increase, at least temporarily, the likelihood that humans will approach, rather than avoid, novel foods. Decreased arousal levels (Pliner et al., 1995; Pliner & Melo, 1997), information about novel foods (Martins et al., 1997; McFarlane & Pliner, 1997; Pelchat & Pliner, 1995), observation of a neophilic model (Harper & Sanders, 1975; Hobden & Pliner, 1995), and increased familiarity with a particular novel food or with novel foods in general (Birch & Marlin, 1982; Birch et al., 1987; Pliner et al., 1993; Tuorila et al., 1994) have all been shown to decrease neophobic responses. Another factor known to influence level of food neophobia in humans is the amount of familiarity in the general situation. Pelchat (unpubl.) found that the more familiarity contained in an eating situation (eating food offered by a vendor in the subways vs. a psychologist in the laboratory vs. a friend in the friend’s home), the more willing subjects were to eat an unfamiliar food. Increases in familiarity can also reduce children’s neophobia. In a study conducted by Harper and Sanders (1975), children, even in their own homes, were more likely to accept a novel food from their mothers than they were from the unfamiliar experimenter. Although the mother represents more than simply familiarity, one interpretation of these results is that adding familiarity to the situation appears to reduce food neophobia. People within a variety of cultures appear to have adopted a version of this strategy to encourage the acceptance of novel foods into the culture. This version involves the ubiquitous usage of combinations of seasonings, referred to by E. Rozin (1983) as flavor principles and described as one of the factors that serve to transform foods into culturally distinctive cuisines. Examples of flavor principles would include the tomato–garlic–olive oil combination which characterizes Italian cuisine or the soy sauce–rice wine–ginger mixture which typifies Chinese cooking. In some instances, these general flavor principles are further differentiated into more specific regional principles. Thus, in Cantonese cooking, garlic and black beans are often added to the more general flavor principle, while in the cooking of Szechuan, a “hot” ingredient such as brown pepper or some form of chilli is added. To a great degree, the flavor principles “make” Italian food, Italian, or Chinese food, Chinese, or Cantonese food, Cantonese. Flavor principles are commonly employed in many different dishes of a cuisine from soups and salads to grains, meats and vegetables. E. Rozin and P. Rozin (1981) note that individuals have a deep attachment to their culture’s characteristic seasonings and find it difficult to imagine food prepared without them. Food simply doesn’t taste good without the appropriate flavor principle. “Traditional flavorings are high-priority culinary items and immigrant groups typically go to great lengths and expense to procure them in foreign settings” (p. 6). E. Rozin and P. Rozin (1981) hypothesize that one of the main functions of flavor principles is to add sufficient familiarity to novel foods to decrease neophobia, thereby facilitating their introduction into a culture. Anecdotal evidence illustrates this supposition. As a part of efforts in the early twentieth-century by home economists and other food specialists to increase variety into the Japanese diet by adopting
EFFECT OF FLAVOR PRINCIPLES
211
foods from abroad, cookery columns of women’s magazines encouraged middleclass housewives to try Western dishes. Among the strategies employed to facilitate this was the addition of Japanese seasonings to make the dishes more familiar (Cwiertka, 1998). More recently, E. Rozin and P. Rozin (1981) asked women in a Mexican village to imagine serving three novel foods (alfalfa, sardines and a common shrub). The women could do so only after picturing each food seasoned with their traditional flavor principle, consisting of chilli and lime. The study which follows was designed to examine experimentally the neophobiareducing function of flavor principles. Thus, an important task was to decide how exactly to operationalize the flavor principle construct. In order to do so, we first attempted to abstract from the notion of flavor principles, a cultural concept, their important elements so we could translate these into the concrete operations of the psychology laboratory. E. Rozin’s (1983) description suggests that a flavor principle is pervasive to the foods of a particular culture, is eaten by members of the culture frequently, typically many times a day, and is strongly liked by members of the culture. Thus, flavor principles are ubiquitous, familiar and palatable. In addition, because they are habitually eaten in association with particular foods, they must have some degree of “combinatorial” appropriateness with these foods. Of course, these characteristics of flavor principles are not independent. For example, familiarity and liking are known to be highly correlated (Birch & Marlin, 1982; Zajonc, 1968), and ubiquity and combinatorial appropriateness are presumably also related. That is, the more culinary situations in which a flavor principle appears, the more foods with which it is likely to be combined and, therefore, the greater its range of appropriateness. However, there must be limits on the extent of a flavor principle’s combinatorial appropriateness. A general limit can be seen in the fact that flavor principles are not often used in beverages or sweet dishes (E. Rozin, 1983). For the purposes of the present study, we decided that two of these characteristics of flavor principles, familiarity and palatability, were most important. Our goal was to test the hypothesis that the presence of a flavor principle could provide enough dietary familiarity to induce individuals to try an unfamiliary food; thus, familiarity appeared to be a characteristic crucial for inclusion. Because familiarity and palatability are highly correlated and because palatability is such an important aspect of food choice, we also chose to incorporate it into our operational definition of the flavor principle construct. We made the decision not to include ubiquity and appropriateness in our operational definition of flavor principles. Although ubiquity is a characteristic of flavor principles in general, this seemed less important to our specific goal of using them to decrease neophobia toward particular novel foods. Appropriateness also seemed less important, since with novel foods, individuals would presumably not have well-defined ideas of the appropriateness of the combination. With these considerations in mind, we decided on an approach in which participants, unselected for their membership in any particular cultural group, would chose for themselves “flavor principles” which were individually appropriate. We put the term in quotation marks at this point to indicate that we are now referring to our experimental analogues of flavor principles. In the experiment to be described, these consisted of common and readily available sauces and condiments for which participants indicated high degrees of familiarity and liking. Thus, the “flavor principles” used in our study are analogous to what E. Rozin (1983) describes in that they incorporate the characteristics of familiarity and palatability. However, even on these dimensions, they are obviously far from perfect replicas of flavor
212
C. STALLBERG-WHITE AND P. PLINER
principles; they must be much “weaker” than real ones since it is unlikely that our subjects eat them as frequently or feel as positively toward them as would be the case with traditional seasonings. The basic prediction of the study is that the addition of a familiar and liked “flavor principle” to an unfamiliar food will increase individuals’ willingness to taste it, in comparison to their willingness to taste the same novel food in the absence of the flavor principle. Since individuals have little reluctance to approach familiar foods, addition of a familiar flavor principle to a familiar food should have little effect; thus, we expect an interaction between the novelty of the food and the presence of a flavor principle. We expected that the addition of the flavor principles would increase individuals’ willingness to try novel foods. We were not sure, however, how the addition of the flavor principles would change the novel foods or individuals’ perceptions of them to produce increased willingness. In other words, we wondered how flavor principles might mediate changes in willingness. At least two different mediators seemed possible. We introduced the flavor principle notion originally in the context of research which implies that increases in familiarity are likely to be the mediator; that is, the addition of an extremely familiar sauce could increase the familiarity of the novel food, causing its anticipated taste or appearance to be more familiar. This increased familiarity would lead to greater willingness. However, it also seemed possible that hedonic changes could be important. That is, the addition of a hedonically positive sauce could increase the hedonic aspects of the novel food, causing it to (be expected to) look or taste better. Some of our dependent measures attempted to clarify this issue. Finally, we assessed participants’ degree of adventurousness with respect to food in two different ways. Pliner and Hobden (1992) have demonstrated that there are stable individual differences in the extent to which individuals report themselves willing to approach novel foods and in the extent of experience with exotic foods. A previous study testing the effectiveness of a modeling manipulation to reduce food neophobia found that it was differentially effective, depending on subjects’ levels of food neophobia (Hobden & Pliner, 1995). In the present study, we examined the extent to which the flavor principle manipulation interacted with food adventurousness, as assessed by the Food Neophobia Scale (FNS) and by a measure of experience with various ethnic cuisines, shown in a previous study to be negatively related to the FNS (Pliner & Hobden, 1992).
M Pilot Study Before conducting the study proper, we conducted a questionnaire-based pilot study which had two main purposes. The first was to present participants with a fairly extensive list of sauces so we could determine which were rated as most familiar and highest in palatability; these would comprise the flavor principles to be made available for choice in the study proper. The second purpose of the pilot study was to obtain some preliminary data to test our flavor principle hypothesis. The study was conducted by students in an upper-level undergraduate Psychology course who administered questionnaires to 98 participants (primarily students) with a mean age
EFFECT OF FLAVOR PRINCIPLES
213
of 26·8. The first page of the questionnaire listed 19 sauces (e.g. Italian tomato sauce, teriyaki sauce, beef gravy) and requested ratings (on 9-point scales) of familiarity and liking. Once a participant had completed this page, the experimenter briefly examined it, and used the ratings to select as a flavor principle the sauce with the highest familiarity rating (and, if there were more than one, the one of these with the highest liking rating). After selecting this flavor principle, the experimenter wrote its name in the appropriate place on the second page of the questionnaire, as described below. This sauce is referred to subsequently as the participant’s “chosen” flavor principle. The second page of the questionnaire listed and described four foods, two novel and two familiar. Following the name and description of each of the first two foods (one familiar, one novel) was the question (answered on 9-point bipolar rating scales) “how willing would you be to taste a small amount of (food)?” For the third and fourth foods, the willingness question read “how willing would you be to taste a small amount of (food) with ——?” Written into the blank was the name of the particular participant’s chosen flavor principle. Subjects also completed two measures of their adventurousness with respect to food. One requested ratings of the frequency of eating foods from 10 relatively exotic (e.g. Korean, Vietnamese) and not-so-exotic (e.g. Chinese, Italian) cuisines, and the other was the Food Neophobia Scale (Pliner & Hobden, 1992). The results indicated that our list of sauces was sufficiently diverse to accommodate nearly all subjects. For 90%, the chosen flavor principle had been rated as “9” on the familiarity dimension and for 95%, it had been rated as “8” or “9” (M=8·84). Furthermore, subjects liked their chosen flavor principles (M=8·81). More important, the results of the pretest indicated that by having available a relatively small number of sauces in the study proper, it would be possible to provide a sauce high in familiarity and palatability for nearly all participants. The five sauces most frequently chosen as participants’ flavor principles according to the criteria above represented those chosen by more than 60% of participants. For the remaining participants, whose chosen flavor principle was not one of the top five, at least one of the top five received high liking and familiarity ratings. Thus, by having available only the top five sauces in terms of their rated familiarity, it would be possible to provide a familiar and liked flavor principle for participants in the study proper. To test the flavor principle hypothesis with our pilot data, we examined the participants’ reported willingness to taste the four foods (two novel, two familiar; one of each with, and one without, the chosen flavor principle) described in their questionnaires. We anticipated that subjects would be generally less willing to taste novel (vs. familiar) foods; the addition of a familiar flavor principle to the novel, but not the familiar, food should increase willingness to taste it. Thus, results in accordance with the prediction would appear in the form of an interaction between the two (within subjects) independent variables, novelty of food and presence of sauce. As expected, there was a highly significant main effect of novelty of food, F(1,99)=67·05, p<0·001; participants were more willing to try the familiar than the novel food (means: 7·8 and 5·9, respectively). However, the predicted interaction did not approach significance. In a second analysis, we characterized participants as relatively more or less experienced with respect to exotic foods, by dividing them at the median in terms of their frequency of eating the five least commonly eaten of the 10 cuisines desribed above. We then added status (above or below the median) on this new variable as a between subjects variable in a 2×2×2 analysis which also included novelty of food and presence of sauce as above. In addition to a novelty
214
C. STALLBERG-WHITE AND P. PLINER
main effect, there was a significant interaction involving all three variables, F(1,98)= 7·06, p<0·01. Examination of the means revealed the expected pattern for the participants below the median on our food experience measure: adding a sauce to a novel food increased willingness to taste it but had no effect on willingness to taste a familiar food (means: novel–no sauce=5·34, novel–sauce=6·11; familiar–no sauce=7·53; familiar–sauce=7·66). The pattern for the participants above the median on food experience was very different, showing no greater willingness to try a novel food with than without a sauce (means: novel–no sauce=6·28; novel–sauce=6·00; familiar–no sauce=8·02; familiar–sauce=8·25). When the data for the two groups of participants were analysed in separate analyses, there was a novelty by sauce interaction for those below the median in experience, F(1,44)=4·63, p<0·05, while for those above the median, the interaction did not approach significance (F<1). Contrasts on the means revealed that the relatively unexperienced individuals were significantly more willing to try a novel food with than without a sauce, t(98)=3·26, p<0·01. For the same participants, adding a sauce to a familiar food had no effect on willingness, and for relatively highly experienced participants, addition of a sauce had no effect on willingness to try familiar or novel foods. Thus, the flavor principle hypothesis received some empirical confirmation, albeit only for participants who were below the median on one of our food adventurousness measures. When we did a similar analysis, dividing participants at the median on the FNS, we did not obtain the same interaction. However, based on the results with the experience variable, we decided to proceed to the experiment proper. Participants Participants were 44 female and 20 male undergraduate students enrolled in an Introductory Psychology course, who participated for credit. They ranged in age from 18 to 33 with a mean of 20·3 years (SD=2·6). They were unselected for cultural group membership, and we did not collect information relevant to this characteristic. Overview Participants were led to believe at the time of recruitment that they would be taking part in an experiment designed to determine how various food combinations influence flavor-intensity perception. The experimental session consisted of three phases: (1) administration of questionnaires (questionnaire phase); (2) tasting and rating for flavor intensity either one or two novel sauces which were combined with a series of familiar foods (tasting phase); and (3) assessment of willingness to try novel and familiar foods alone or in combination with a flavor principle selected by the participant (rating phase). As in the pilot study, the main independent variables were novelty of food and presence of sauce and the dependent variable was willingness to taste. Procedure Questionnaire phase The purposes of the questionnaire phase were to select flavor principles for participants and to measure their degree of food adventurousness. They completed
EFFECT OF FLAVOR PRINCIPLES
215
a questionnaire on the pretext that “. . . eating habits, especially experience with certain cuisines, are known to influence flavor intensity perception . . .”. Participants rated the 19 sauces used in the pilot study for liking and familiarity, using 9-point bipolar scales (endpoints “not at all” and “extremely”) and rank ordered their top five sauces in terms of liking. The experimenter used this information to select a highly familiar and liked flavor principle for each participant for use during the willingness phase of the experiment. As noted earlier, the pilot data enabled us to determine the most familiar and preferred sauces for this population, and we kept supplies of the top-rated sauces on hand in the laboratory. Flavor principles were then selected according to the following criteria: (1) if the participant’s top-ranked sauce was among those on hand, choose that; (2) if not, choose the next ranked sauce. Using these criteria, we were able to accommodate all participants with the sauces on hand. On another section of the questionnaire, participants rated (9-point scales) their frequency of eating foods from 10 cuisines, as described in the pilot study; this provided the measure of experience with exotic cuisines. Finally, participants completed the Food Neophobia Scale (Pliner & Hobden, 1992). Tasting phase The purposes of the tasting phase were primarily to make it clear to participants that the study involved the actual tasting of foods and to provide a context in which we could subsequently ask them about their willingness to taste various foods both in combination with a sauce and alone. In the tasting phase, participants ingested small samples of familiar foods (such as rice, potato, celery and corn chip) both alone and accompanied by one or two commercially available cooking sauces (Creole Sauce or Ginger Tamari Sauce, both Master’s Choice brand), described as “flavor mediums”. In keeping with the cover story, participants made ratings of flavor intensity (9-point bipolar scales with endpoints labeled “extremely bland” and “extremely intense”) after they tasted the foods, both alone and in combination with the sauces. Willingness phase The main purpose of the willingness phase was to measure the dependent variable: willingness to taste novel and familiar foods both with and without flavor principles. To accomplish this, we led participants to believe that the experiment was over and that we were now pretesting foods for future use. They were told: “. . . that’s all the food I need you to taste for the experimental part of the session. What I’d like to do now is to get your opinion regarding a variety of foods and flavor mediums that we are thinking of using in another part of this study. This time, though, you won’t have to try any of the samples I’m about to show you, unless you are willing to do so.”
The set of foods used in the willingness phase consisted of four novel and four familiar foods. The novel foods were parval (an Indian vegetable), gathiya (an Indian snack chip), tapioca stick (an Asian noodle made from tapioca) and hominy (hulled corn kernels with the germ removed); the familiar foods were raw carrot, taco chips, cooked spaghetti and corn niblets. Participants were presented with one of the novel and one of the familiar foods with their chosen sauce and different novel and familiar foods with no sauce. Different participants received different foods, and the same food was presented to equal numbers of participants with and without a sauce. They
216
C. STALLBERG-WHITE AND P. PLINER
T 1 Willingness to try foods as a function of novelty of food and sauce condition in the study proper Food condition Sauce condition Plain Sauced
Familiar
Novel
8·19 (1·38) 64 7·06 (2·27) 64
6·03 (2·48) 64 6·48 (2·17) 64
rated their willingness to taste each of the four foods presented, with the understanding that their ratings would determine what they actually did taste. They also rated anticipated liking and anticipated familiarity for each. The food samples were displayed in small, clear plastic containers and were covered with clear, plastic lids, so that they could be seen but not smelled. The sauces were displayed in small plastic cups, also with lids. The samples remained behind a partition until they were shown. The names of the foods and sauces were announced as they were presented, and their names were also written on the lids of their containers. So, for example, when foods were presented in the “flavor principle” (or with sauce) condition, the sauce container was placed on top of the food container, and the two were shown together and described as “—— (food) with —— (sauce)”. At the beginning of the willingness phase, participants were told that they would be given fresh samples of the foods they were willing to taste, “. . . not these old pretest samples”. Just before food presentation began, participants were reminded that, when they completed the ratings, their responses would be analysed to see which of the foods and/or food and food–sauce combinations the experimenter would prepare for them to try.
R Willingness Measure Preliminary analyses revealed no effects of sex of participant even approaching significance; accordingly, data for males and females were combined for analysis. The main analysis was a 2×2 ANOVA on participants’ willingness ratings, with novelty of food and presence of sauce as within subjects independent variables. The results of the analysis revealed a significant effect of food novelty, F(1,60)=35·67, p<0·001; participants were more willing to try the familiar foods than the novel ones (means: 7·6 and 6·3, respectively). The only other significant effect was a significant novelty of food by presence of sauce interaction, F(1,60)=14·61, p<0·001. The means are presented in Table 1, where it can be seen that the addition of participants’ flavor principles made them more likely to taste a novel food, t(60)=1·60, p=0·06,
217
EFFECT OF FLAVOR PRINCIPLES
T 2 Expected liking and expected familiarity as a function of novelty of food and sauce condition in the study proper Food condition Measure Liking Familiarity
Sauce condition
Familiar
Novel
Plain Sauced Plain Sauced
7·7 6·3 8·8 8·6
5·0 5·7 1·8 2·8
and less likely to taste a familiar food, t(60)=3·87, p<0·01, in comparison to the unsauced versions. Our second analysis was a 2×2×2 ANOVA in which participants’ status on the food experience variable (above or below the median) was added to the familiarity and sauce variables. In this analysis, the effects of novelty and novelty by sauce obtained in the original analysis remained significant; however, the food experience variable did not interact with the familiarity and sauce variables. A similar analysis, with FNS scores added as a between subjects variable, also yielded significant effects of novelty and a significant novelty by sauce interaction but did not yield a significant second order interaction. Thus, in this study, unlike the pilot study, a flavor-principle effect was obtained for all subjects, not just those low in experience with exotic cuisines. Anticipated Liking and Familiarity We performed 2 (novelty of food) ×2 (presence of sauce) within subjects ANOVAs on participants’ reports of anticipated liking for and anticipated familiarity with the taste of the foods/food–sauce combinations. The analysis on liking yielded a main effect for novelty of food, F(1,59)=65·74, p<0·001. Participants had higher anticipated liking for familiar foods than for novel foods (means: 7·0 and 5·3, respectively). There was also a significant novelty by sauce interaction F(1,59)=26·54, p<0·001. Means for this interaction reveal that the effect of the addition of the sauce on anticipated liking depended on whether it was offered with a novel or a familiar food. Participants anticipated greater liking for a novel food when it was served with their own sauce than when it was served plain (means: 5·7 and 5·0, respectively), t(59)=2·59, p<0·05; the reverse occurred for a familiar food (means: 6·3 vs. 7·7), t(59)=4·67, p<0·01. With anticipated familiarity as the dependent variable, the same two effects emerged; participants rated familiar foods as significantly more familiar than novel foods, F(1,60)=790·78, p<0·001 (means: 8·7 and 2·3, respectively). This finding provides a check on our novelty manipulation; the foods we have been referring to as novel or familiar were perceived accordingly by our participants. There was also an interaction between novelty and presence of sauce, F(1,60)= 10·64, p<0·01; novel foods presented in combination with a familiar sauce seemed more familiar than when they were offered plain (means: 2·8 and 1·8, respectively), while for familiar foods there was very little difference (means: 8·6 vs. 8·8, respectively).
218
C. STALLBERG-WHITE AND P. PLINER
In order to examine the extent to which the flavor principle effect is mediated by changes in expected familiarity and/or palatability produced by the addition of the flavor principle, we did the following analysis. We computed scores for each participant consisting of the differences (for both the hedonic and familiarity ratings) between the sauced and non-sauced versions of the novel foods. These scores presumably represent how much more (or less) familiar/palatable the sauced version is than the unsauced version. We then regressed these scores on participants’ willingness to taste the sauced version of the novel food, using a stepwise procedure. Thus, in a sense, we are examining the rated “consequences” of adding flavor principles to novel foods as predictors of willingness to taste them. The analysis revealed that the hedonic difference score was the only significant predictor of participants’ willingness to taste the novel food, t=4·69, p<0·001, accounting for 28% of the variance (multiple R=0·52). Thus, the more that a participant expected the sauced version of the novel food to be more palatable than the unsauced version, the more willing he/she was to try the novel sauced food.
D We were able to provide evidence in favor of the flavor principle hypothesis; that is, the addition of a familiar and liked sauce to a novel food increased willingness to try it. This effect occurred in the pilot study, in which participants answered questions pertaining to their hypothetical willingness, only for the participants with less experience with the exotic ethnic cuisines, and in the study proper, in which participants were under the impression that their willingness ratings would determine what they would actually taste, for all participants. It is possible that when food choice is only hypothetical, individuals who are relatively adventurous/experienced with novel foods make neophilic choices, but when the choices become real, even these individuals need the added familiarity of a flavor principle to induce neophilic behavior. There is one respect in which the data from the study differed from expectation. We had anticipated that the addition of flavor principles to familiar foods would have no effect on willingness to taste them because individuals are not reluctant to approach familiar foods in the first place. If any effect could be expected, it would be that flavor principles might produce a slight increase willingness to try these foods to the extent that the flavor principles were perceived as more familiar and/or palatable than the foods themselves. What we found, however, was the reverse. The addition of flavor principles to familiar foods actually decreased participants’ willingness to try them. It is at this point that the issue of “combinatorial appropriateness”, mentioned in the Introduction as a characteristic of flavor principles, must be revisited. As mentioned earlier, we quite consciously decided not to include this appropriateness factor in our operational definition of flavor principles on the grounds that, for novel foods, individuals would probably not have well-defined ideas of appropriateness. In addition, the data from the pilot study showed no difference in willingness to try familiar foods with and without a sauce. Furthermore, in that study, as a rough measure of appropriateness, we asked participants to rate (on 9-point scales) the extent to which the addition of the sauce “. . . would clash with [the food’s] flavor”. For both the familiar and the novel foods the mean ratings were relatively low (3·9 and 4·5, respectively) and did not differ significantly. Thus,
EFFECT OF FLAVOR PRINCIPLES
219
it appears that, on the average, participants found the sauces to be appropriately combined with the foods, both novel and familiar. As a result, when we designed the study proper we did not concern ourselves with appropriateness. Given the unexpected results of the the study proper, we decided to use the “clash” ratings from the pilot study to explore further the notion of appropriateness. Although the overall means were low, there were sizeable individual differences in responses to these questions, produced at least in part by the fact that different subjects received different food–sauce combinations. There were two reasons for this: first, different subjects had chosen different flavor principles, and second, participants received, at random, one of eight different novel foods. If participants are divided at the median in terms of their responses to this question about “clashing”, those above the median (i.e. those for whom the pairing of food and sauce was perceived as relatively inappropriate) were less willing to try the familar food with the sauce than without it (means: 7·11 and 7·72, respectively). In contrast, those below the median (i.e. those for whom the pairing was relatively appropriate) were slightly more willing to try the familiar food with the sauce than without it (means: 8·22 and 7·76, respectively). In other words, for familiar foods, appropriate sauces increased willingness slightly while inappropriate ones decreased it. In the case of novel foods, however, the relatively less food adventurous participants (for whom the flavor principle effect occurred) were more willing to taste the novel food with sauce than without, regardless of their ratings of the sauce’s appropriateness (means for appropriate sauces=6·48 and 5·48, respectively; means for inappropriate sauces= 6·36 and 5·00, respectively). In other words, both relatively appropriate and inappropriate sauces increased willingness to try novel foods. We suggest that this occurred for one of two reasons. One possibility is that ideas about appropriateness are less firmly held when a food is novel—what is appropriate is less clear. A second, not incompatible, possibility is that despite inappropriateness, a familiar and liked flavor principle is able to overcome neophobia. That is, it may be precisely the aspects of flavor principles we included in our operational definition that are important for inducing individuals to try novel foods. Appropriateness, although an important characteristic of a flavor principle, may be less important than familiarity and palatability in the context of this particular function. However, there are clearly limits on the “unimportance” of appropriateness; Pliner and Stallberg-White (1998) found that an extremely palatable and familiar, yet highly inappropriate, flavor principle did not increase children’s willingness to taste an unfamiliar food. Clearly, the sauces employed in our studies only approximated flavor principles as they were conceptualized by E. Rozin (1983). And, certainly, flavor principles may have functions other than the neophobia-reducing one tested in our studies. Other possibilities suggested by E. Rozin and P. Rozin (1981) include adding gustatory and/or visual zest to otherwise bland-tasting and/or -looking diets and providing a cultural “marker” in the sense of identifying individuals as members of a cultural group and separating that group from others. And here, aspects of flavor principles not incorporated into our operational definition might be relevant. For example, a particular means of seasoning food could be a more potent cultural marker, the more ubiquitously it appeared in the culinary repertoire of a particular culture. Or if a particular set of seasonings serves to reduce the blandness of a group’s diet, then the seasonings might be most appropriately combined with the bland elements of the diet and less with those with more distinctive flavors of their own. Our point here is to acknowledge that we are aware of the fact that our study utilized what
220
C. STALLBERG-WHITE AND P. PLINER
might be considered to be an incomplete or limited instantiation of the flavor principle construct—what might be considered to be a far cry from the real thing. However, our intention was also limited—namely, to study in the laboratory the neophobia-reducing effects of flavor principles. And for that purpose we chose the aspects of flavor principles which appeared to be most applicable: familiarity and palatability. Thus, although our laboratory experiment did not capture the flavor principle construct in its entirety, it may have captured the elements of it which are most relevant to the neophobia-reducing function of flavor principles. Although humans’ omnivorous heritage inclines them to be neophobic, it is clearly in their interest to be flexible on this issue. Flavor principles afford a means of “certifying” a new food as familiar and safe to eat, thereby providing a cultural solution to the omnivore’s dilemma. This is yet another example of a cultural solution to nutritional problems produced by incompatibility between the available foods and the biological characteristics of humans. Other examples would include techniques for detoxification of toxic but otherwise nutritious foods such as cassava and soybeans (Katz, 1982) and the treatment of foods to free otherwise unusable nutrients (Katz et al., 1974). Thus, flavor principles could be thought of as one of many means by which humans negotiate between their biology and their environment. R Birch, L. L. & Marlin, D. W. (1982). I don’t like it; I never tried it: Effects of exposure on two-year old children’s food preferences. Appetite, 3, 353–360. Birch, L. L., McPhee, L., Shoba, B. C., Pirok, E. & Steinberg, L. (1987). What kind of exposure reduces children’s food neophobia? Looking vs. tasting. Appetite, 9, 171–178. Cwiertka, K. (1998). A note on the making of culinary tradition—an example of modern Japan. Appetite, 30, 117–128. Harper, L. V. & Sanders, K. M. (1975). The effect of adults’ eating on young children’s acceptance of unfamiliar foods. Journal of Experimental Child Psychology, 2, 206–214. Hobden, K. & Pliner, E. (1995). Effects of a model on food neophobia in humans. Appetite, 25, 101–114. Katz, S. H. (1982). Food, behavior and biocultural evolution. In L. M. Barker (Ed.), The psychobiology of human food selection. Westport, CT: Avi Publishing Company, Inc. Katz, S. H., Hediger, M. L. & Valleroy, L. A. (1974). Traditional maize processing techniques in the New World. Science, 184, 767–773. Martins, Y., Pelchat, M. L. & Pliner, P. (1997). “Try it; it’s good and it’s good for you”: Effects of taste and nutrition information on willingness to try novel foods. Appetite, 28, 89–102. MacFarlane, T. & Pliner, P. (1997). Increasing willingness to taste novel foods: Effects of nutrition and taste information. Appetite, 28, 227–238. Pelchat, M. L. (1985). The effect of context on neophobia in college students. Unpublished manuscript. Pelchat, M. L. & Pliner, P. (1995). “Try it. You’ll like it”: Effects of information on willingness to try novel foods. Appetite, 24, 153–165. Pliner, P., Eng, A. & Krishnan, K. (1995). The effects of fear and hunger on food neophobia in humans. Appetite, 19, 77–87. Pliner, P. & Hobden, K. (1992). Development of a scale to measure the trait of food neophobia in humans. Appetite, 19, 105–120. Pliner, P. & Melo, N. (1997). Food neophobia in humans: Effects of manipulated arousal and individual differences in sensation seeking. Physiology and Behavior, 61, 331–335. Pliner, P., Pelchat, M. & Grabski, M. (1993). Reduction of neophobia in humans by exposure to novel foods. Appetite, 20, 111–123. Pliner, P. & Stallberg-White, C. (1998). “Pass the ketchup, please”: Flavor principles increase children’s willingness to taste novel foods. Unpublished manuscript.
EFFECT OF FLAVOR PRINCIPLES
221
Rozin, E. (1983). Ethnic cuisine: The flavor principle cookbook. Lexington, MA: The Stephen Greene Press. Rozin, E. & Rozin, P. (1981). Culinary themes and variations. Natural History, 90, 6–14. Rozin, P. (1984). The acquisition of food habits and preferences. In J. D. Matarazzo, S. M. Weiss, J. A. Herd, N. E. Miller & S. M. Weiss (Eds), Behavioral health: A handbook of health enhancement and disease prevention, pp. 590–607. New York, NY: John Wiley. Shulze, G. & Watson, N. V. (1995). Comments on “Flavor neophobia in selected rodent species”. In R. Wong (Ed.), Biological perspectives on motivated activities, pp. 229–230. Norwood, NJ: Ablex Publishing Corp. Tuorila, H., Meiselman, H., Bell, R., Cardello, A. V. & Johnson, W. (1994). Role of sensory and cognitive information in the enhancement of certainty and liking for novel and familiar foods. Appetite, 2, 231–246. Wong, R. (1995). Flavor neophobia in selected rodent species. In R. Wong (Ed.), Biological perspectives on motivated activities, pp. 229–264. Norwood, NJ: Ablex Publishing Corp. Zajonc, R. (1968). Attitudinal effects of mere exposure. Journal of Personality and Social Psychology Monograph Supplement, 9 (2, Part 2), 1–27. Received 20 January 1998, revision 28 September 1998, accepted in revised form 3 June 1999