Consumer acceptance of cheese,influence of different testing conditions

Consumer acceptance of cheese, influence of different testing conditions

Consumer acceptance of cheese, influence of different testing conditions Margrethe Hersleth a,b, * , Øydis Ueland a ,H  el  ene Allain a , Tormod Næs a,c a Matforsk, Norwegian Food Research Institute, Osloveien 1, N-1430  As, Norway b Agricultural University of Norway, Department of Chemistry, Biotechnology and Food Science, Box 5003, N-1432  As, Norway c University of Oslo, Box 1072, Blindern, No-0316 Oslo, Norway Received 29 September 2003; received in revised form 24 February 2004 Available online 8 April 2004 Abstract The main objective of this study was to evaluate the effect of various test situations on hedonic responses to food. The experimental sample set consisted of six different cheeses selected on the basis of a sensory descriptive test of a total of 15 cheeses performed by a trained panel. A panel of 87 consumers evaluated the cheese samples for overall liking in three different testing conditions, at a laboratory, at a central location and at home. The consumer panel was divided into three groups, each with different order of the testing conditions. The main results showed that changing the environments and the degree of social interaction in the consumer tests had no significant effect on hedonic ratings for the products. Lack of a natural meal context during testing, similar expectations in the three testing situations and high familiarity of the product category may explain the high degree of consistency in hedonic ratings. Ó 2004 Elsevier Ltd. All rights reserved. Keywords: Context; Liking; Consumer; Cheese; Familiarity; Expectation 1. Introduction There is an increasing demand for new products and line extensions in the food market. Hedonic scales and preference tests are common instruments used when decisions on market introductions are made. The prac- tical benefits of such tests are, however, strongly dependent on the validity of the test, as well as the sensitivity and the reliability (K € oster, Couronne, L  eon, L  evy, & Marcelino, 2002; Lawless & Heymann, 1999; Meiselman, 1993). According to Schutz (1999) internal validity of a test means that the data collected apply to the people, products and situations in that particular test. External validity, on the other hand, means that the data can be used to make more general conclusions about other people, products and situations. The validity of hedonic tests is therefore closely associated with the uniformity and the consistency of the consumer sample (K € oster, 2003; K € oster et al., 2002), but is also closely connected to the area of context research (Bell & Meiselman, 1995; Meiselman, 1996; Rozin & Tuorila, 1993). In the present paper we will focus on a particular aspect of validity of hedonic ratings; namely how con- sistent consumers are when rating food products on a hedonic scale in different testing conditions. Most hedonic food testing is conducted in laborato- ries, at central locations or at home. A laboratory rep- resents the most controlled environment for testing. In the testing area, one can usually control most environ- mental variables, stimulus variables and to a certain degree social interaction. The arguments in favour of using a laboratory test are to allow for focusing on the sensory characteristics of the products without being influenced by external variables related to the eating environment or social settings. However, a sensory testing booth is very different from a real eating envi- ronment and the realism of the test can be questioned. Simple laboratory food preference measures have in some projects shown to be poor predictors of con- sumption (Cardello, Schutz, Snow, & Lesher, 2000; Kozlowska et al., 2003). Meiselman (1992) has pointed out these aspects in his call for ‘‘real foods to be tested by real people in a real dining situation’’. * Corresponding author. Address: Matforsk, Norwegian Food Research Institute, Osloveien 1, N-1430  As, Norway. Tel.: +47-6497- 0100; fax: +47-6497-0333. E-mail address: margrethe.hersleth@matforsk.no (M. Hersleth). 0950-3293/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.foodqual.2004.02.009 Food Quality and Preference 16 (2005) 103–110 www.elsevier.com/locate/foodqual An alternative to a laboratory test is a central loca- tion test (Lawless & Heymann, 1999) organised in a mall, at a school, in a canteen or as in this project at a club house belonging to a soccer team (see description of the facilities in Section 2.4). In a central location area one can partly control environmental variables, stimulus variables and the degree of social interaction. It is pos- sible to increase the realism of the test by making the test area more natural and home-like. The effect of social interaction, physical environment and freedom with re- spect to food choice was studied by King, Weber, Meiselman, and Lv (in press). They found that intro- duction of context effects in a central location test can improve the ability to predict actual liking scores in a real life environment. Food testing by consumers in their home is consid- ered to be more optimal compared to laboratory testing and central location testing with regard to realism dur- ing tasting and eating a product sample. In such tests the products can be tested under its normal conditions of use (Meilgaard, Civille, & Carr, 1991). However, a home test does introduce many practical challenges. The po- tential for product misuse is very high (Stone & Sidel, 1992). Other challenges are variability in preparation, variability from the time of use and variability from other foods or products used with the test product (Meilgaard et al., 1991). Home tests are time consuming to organise and it can be difficult to assess the testing procedure and the validity of the resulting data. The main objective of the present study was to compare and evaluate the effects of various commonly used test situations on hedonic responses to a particular food. The same set of products was presented to a group of consumers in three different locations; in a labora- tory, at a central location and at home. A selection of semi-hard and hard cheeses from the Norwegian market was chosen as products. The cheeses are well-known varieties in the market and typically used for breakfast, lunch and supper. 2. Materials and methods The study consisted of two parts, a conventional sensory profiling of fifteen cheese samples and a con- sumer test with six cheese samples served in three ses- sions. 2.1. Products This study was a part of a larger preference mapping project on hard and semi-hard Norwegian cheeses. For sensory profiling 15 different cheeses commercially available at the Norwegian market were selected. The consumer test was performed on six of the cheeses se- lected from the total number as described below. All the presented cheese samples in the descriptive test and in the consumer test came from the same batches. The samples were therefore treated as homogeneous in the statistical analysis. 2.2. Descriptive sensory analysis All cheeses were evaluated by a panel using descrip- tive sensory profiling as described in ISO 6564:1985. The sensory panel consisted of 11 panellists selected and trained according to guidelines in ISO 8586-1:1993 and the laboratory was designed according to guidelines in ISO 8589:1988. The surface area of the cheese samples was removed and 50 g was served to each panellist. The serving temperature of the cheese samples was 13 °C. The panellists developed a test vocabulary describing differences between samples and they agreed upon a consensus list of 22 attributes for profiling. A continu- ous, non-structured scale was used for evaluation. The left side of the scale corresponded to the lowest intensity of each attribute (value 1.0) and the right side corres- ponded to the highest intensity (value 9.0). In a pre-test session, the panellists were trained in the use of the scale by testing samples that were considered as extreme on selected attributes typical for the product. Each panellist did a monadic evaluation of the samples at individual speed on a computerised system for direct recording of data (CSA Compusense, version 5.24, Canada). Two replicates were performed for each cheese variety. All samples and replicates were served in a randomised order. The average response over replicates and asses- sors for each significant attribute were used in the multivariate analyses. Based on the results from Principal Component Analysis (PCA), six of the cheese samples (marked in Fig. 1) were selected and used for consumer testing. The selection of samples had to meet the following two cri- teria. (1) For the purpose of the preference mapping project some important market strategic elements had to be met. (2) For the purpose of the present paper the six cheese samples should be sufficiently different to create differences in hedonic liking. The marked samples in Fig. 1 represent a compromise between these two crite- ria. 2.3. Experimental design for consumer study The experimental design for the consumer study is shown in Table 1. A panel consisting of 87 consumers evaluated six cheeses in three different testing condi- tions: In individual sensory booths at a laboratory, at a club house belonging to a soccer team and at home. The quoted number of consumers participated in all three sessions. The consumer panel was divided into three groups (A–C), each with a different order of tasting conditions (shown in Table 1). 104 M. Hersleth et al. / Food Quality and Preference 16 (2005) 103–110 2.4. Consumer testing The consumers were recruited from a local soccer club and were selected according to the following cri- teria: Likers of hard/semi-hard cheese, 30–50 years old, not employee at Matforsk (Norwegian Food Research Institute) or the nearby Food Science Department at the Agricultural University of Norway. The consumers were given the following information about the test during recruiting: ‘‘We seek participants for a consumer test of hard/semi-hard cheese. The test implies evaluation of 18 cheese samples and the test is therefore composed of three sessions. Due to practical circumstances and capacity we will arrange the test at three different loca- tions’’. The three sessions were carried out within eight days (Monday, Thursday and the following Monday). The consumers arrived at the laboratory and at the central location between 6–8 p.m. The test was a blind test i.e. the detailed identity (brand) of the products was un- known to the participants. However, the Norwegian market for cheeses is still dominated by one large cheese producer (Tine BA). Thus, when eating cheeses, most Norwegians associate this product category with the ‘‘Tine’’ brand. The consumers were presented with six coded samples of 200 g semi-hard or hard cheese in each session. The cheeses were evaluated for overall liking on a nine-point hedonic scale anchored with ‘‘Like Ex- tremely’’ and ‘‘Dislike Extremely’’ and with a neutral centre point of ‘‘Neither Like nor Dislike’’ (Peryam & Pilgrim, 1957). The consumers could choose between the use of a cheese slicer or a knife and they were requested to remove the surface area of the sample before tasting. No more information concerning the products or the experiment was given. Serving order in the laboratory session and in the central location session was varied according to a cyclic design balanced for order and carry-over effects (MacFie, Bratchell, Greenhoff, & Vallis, 1989). Water was served as the beverage in the laboratory and at the central location test. The similarities and differences between the experi- mental conditions are shown in Table 2. Laboratory evaluations took place in individual booths and social interaction was discouraged. The serving temperature of the cheese samples was 12–14 °C. The central loca- tion test took place at a club house belonging to the local soccer club from which the consumers were re- cruited. This environment may therefore have given the -1.0 -0.5 0 0.5 1.0 -0.5 0 0.5 1.0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Whiteness Colorhue Saturation Porosity Brightness Elasticity Int.odour Acid odour F.acid odour Int.flavour Salty flavour Bitter flavour F.acid flavour Acid flavour Nutty flavour Sharp flavour Hardness Solubility Fattyness Juiciness Stickiness Graininess PC1 PC2 Fig. 1. PCA bi-plot of significant sensory attributes (p < 0:05) and sensory data for the 15 cheese samples. PC1 account for 55% and PC2 for 26%. Cheese chosen for consumer study are marked. Table 1 Experimental design for consumer study Group Number of consumers Session (Monday week 1) Session (Thursday week 1) Session (Monday week 2) A 30 Laboratory Home Central location B 27 Home Central Location Laboratory C 30 Central location Laboratory Home M. Hersleth et al. / Food Quality and Preference 16 (2005) 103–110 105 consumers a more familiar sensation than the laboratory did. The club house is normally used for various social arrangements in the soccer club such as meetings, serv- ing food, celebrations etc. The testing took place in a meeting room furnished with dining tables with chairs and sofas with armchairs. A kitchen is connected to the meeting room with a service hatch. The consumers ar- rived between 6 and 8 p.m. and were seated around the same dining table, up to a maximum of eight persons at the same time. They were allowed to talk with each other, but not to discuss the samples. The home test was arranged as a one time trial, i.e. the consumers did not use the product over a period of time. At home the consumers were allowed to taste the cheeses together with family and/or friends, but they were instructed to report their personal liking only. The consumers were allowed to taste the cheese sample with bread, crisp- bread or biscuits and self-elected beverages. In addition to a standard score sheet used in all three sessions the consumers were asked to provide the fol- lowing information from the home test: (1) Were you alone or together with family/friends when you tasted the cheese? (2) Did you consume bread, crisp-bread or biscuits with the cheese? (3) Did you consume water or other beverages with the cheese? If the consumers had any cheese left after tasting they were requested not to consume the left over until the last tasting session was finished. 2.5. Statistical analysis 2.5.1. Analysis of sensory data Analyses of Variance (ANOVA, two-way model with interactions and with assessor effects and interactions considered random) was performed on the descriptive sensory analysis data in order to identify the sensory attributes that differentiated between samples. To study the main sources of variation between the average sensory descriptive data, a PCA of the panel averages (significant attributes, p < 0:05) was performed (mean centred data, no standardisation) (Mardia, Kent, & Bibby, 1980). A PCA of the consumer scores of the cheese samples in each contextual situation was also performed in order to study the main tendencies in variation between products and situations. Full cross- validation (Martens & Næs, 1989) was used for valida- tion of the components. 2.5.2. Analysis of consumer data The statistical analysis of the consumer data was performed using Minitab version 13.3. We used the ANOVA model with main effects for consumer, product and context plus all the two-factor interactions. The main effects of the consumers and their interactions with other effects are considered random, the remaining ef- fects are fixed. A separate ANOVA was conducted for the first ses- sion only for each of the three groups (without any possible bias from previous sessions), group A in labo- ratory, group B in home and group C in central loca- tion. The model used was one with main effects for consumer and product (interaction was confounded with error). This model was also used to determine significantly different samples with respect to hedonic liking in each of the three testing situations. Three sep- arate ANOVA analyses were performed, one data set (87 consumers) for each of the testing locations. Tukey’s method for pairwise comparisons was used. 3. Results A PCA-plot (bi-plot) obtained from the significant sensory attributes for the 15 cheese samples is presented in Fig. 1. Principal component 1 (PC1) and principal component 2 (PC2) respectively accounted for 55% and 26% of the total variance. The cross-validation (results not shown) confirmed that interpretation of the first two components is valid. On the basis of the results from the descriptive test and the two mentioned criteria (Section 2.2) we selected the following cheeses for the consumer tests: 1, 4, 8, 9, 11, 14. We considered the chosen cheese samples to be sufficiently different in consistency and flavour attributes to be perceived as different by the consumers. The results from the ANOVA with all the three testing situations and all consumers included are shown in Table 3. As can be seen, there is a significant main effect for consumers (p ¼ 0:01) and a significant inter- action effect between consumer and product (p ¼ 0:01). Table 2 Similarities and differences among the three experimental conditions in the consumer study Laboratory Central location Home Controlled serving order a Yes Yes No Instructions about cutting Yes Yes Yes Temperature of cheese samples 12–14 °C? ? Social interaction No Yes (limited) Yes Bread, crisp bread, biscuits Nothing Nothing Freedom of choice Beverages Water Water Freedom of choice a Balanced for order and carry-over effects. 106 M. Hersleth et al. / Food Quality and Preference 16 (2005) 103–110 The effect of product is also significant (p ¼ 0:01), which means that type of cheese significantly influenced the rating of liking. Neither the effect of context nor its interaction with consumer are significant at the 5% level. The interaction between product and context is signifi- cant (p ¼ 0:01). The interaction plot is given in Fig. 2. As can be seen, the interaction effect is mainly due to cheese number 4 and cheese number 9. Both samples show a larger variation in average rating between test situations than the other samples. For cheese number 4, the laboratory environment gave the highest rating and for cheese number 9, the in-home situation gave the highest rating. The same type of ANOVA was also run for each of the three consumer groups (group A, B and C, results not shown) separately. Likewise, three separate ANO- VAs were run for each of the first testing situations. The results gave the same overall conclusion as for the full analysis, indicating no effect of the order of the testing situation. Table 4 shows data from the in-home testing. The majority of the consumers (80%) tasted the cheese samples without any kind of bread, while 20% had bread, crisp bread or biscuits together with the cheese. More than half of the consumers were drinking water while testing the cheese samples and 16% used other beverages. About 50% of the consumers did the testing alone. Fig. 3 shows the average scores and the ranking of the cheese samples. Samples which are not significantly different at 5% level (Tukey’s method) are marked with lines. As can be seen, the ranking of the cheese samples was similar in the three testing situations. Samples 11 and 8 were the most preferred samples in all three sit- uations and sample 9 the least preferred. However, the clustering of samples in the three situations is different. For the central location test the overlap is the strongest, while for the in-home test the overlap is the weakest. For the laboratory testing, the mid group overlaps with both the high and low groups of samples. Detailed information about average ratings of cheese samples and the standard deviations together with the total averages of the standard deviation per cheese sample and per testing situation are shown in Table 5. The table shows fairly equal average standard deviations in liking for the three testing situations and more vari- ation in average standard deviations for cheese samples. Samples 1 and 11 have the smallest standard deviations while sample 9 has the largest. Table 3 Results from ANOVA of the experimental design shown in Table 1 Variable Degrees of freedom F -value P-value a Consumer 86 1.88 0.01 Product 5 8.93 0.01 Context 2 0.40 0.67 Consumer · Product 430 3.98 0.01 Consumer · Context 10 0.92 0.52 Product · Context 172 1.47 0.01 a P-value of 0.01 means equal to or less than 0.01. Fig. 2. Illustration of interaction effects between product and testing location. Average rating for each product in each testing location. H ¼ in home testing, CL ¼ central location, L ¼ laboratory. Table 4 Reported testing conditions from in-home testing Bread, crisp-bread or biscuits N ¼ 80 Without 80% With 20% Beverages N ¼ 79 Without 28% Water 56% Other 16% Social interaction N ¼ 74 Alone 49% Social group 51% Reported percent is calculated as a percent of N ¼ consumers who answered this particular question. Fig. 3. Average hedonic ratings and ranking of the cheese samples in the three testing situations. Different letters means different ratings at the 5% level of significance, calculated on the three different data sets separately. M. Hersleth et al. / Food Quality and Preference 16 (2005) 103–110 107 To further study the effect of testing condition, a PCA on the consumers scores of the six cheese samples (1, 4, 8, 9, 11, 14) in each location was performed (87 con- sumers as rows and 3 testing locations · 6 cheeses ¼ 18 samples as columns in the data matrix). The PCA score plot of the analysis is presented in Fig. 4. The two first principal components described 45% of the variation in the data. Fig. 4 shows that the principal components are more related to products than to testing conditions. Hedonic ratings for the same product in the three dif- ferent locations are relatively similar. This supports the findings above; the preference pattern among the con- sumers was independent of testing location. 4. Discussion The main finding of this study was that testing loca- tion had no significant effect on hedonic liking for a selection of semi-hard and hard cheeses (shown in Ta- bles 3, 5 and Figs. 2–4). In the following we will discuss similarities and differences between our findings and previous published results from related studies. Meiselman (1996) has pointed out the situation in which food is consumed as an important context factor. An example of a study where location was important for the hedonic ratings is the one conducted by Meiselman, Johnson, Reeve, and Crouch (2000). The main finding in this paper was the consistently higher ratings of res- taurant food over cafeteria food, when the food was identical. A study by Edwards, Meiselman, Edwards, and Lesher (2003) gave similar results. However, in these studies effects of contextual variables were studied in natural eating environments. Our study has another approach as we have investigated the effect of environ- ments and social interaction in consumer testing. The relationship between food or beverage intake on one side and social influence or facilitation on the other, has previously been established by for instance de Cas- tro, Brewer, Elmore, and Orozco (1990) and de Castro and de Castro (1989). Consistent with those findings Hersleth, Mevik, Næs, and Guinard (2003) found that hedonic ratings for Chardonney wines were higher in a reception room where some degree of socialising oc- curred, than in individual booths in a laboratory. The present study also focuses on differences in social interaction. In the laboratory no social interaction was possible, but the two other settings allowed for social influence (Table 2). It was observed that most consum- ers had standard small talk about everyday subjects during the central location test and, as shown in Table 4, more than 50 percent of the consumers performed the cheese testing at home together with family and/or friends. However, the increased element of social atmosphere did not contribute to change the hedonic ratings of the cheese samples. One reason could be a lack of natural meal context. The consumers knew, from the information given during recruiting, that they were participating in consumer testing. A consequence could be that they felt the cheese tasting as an arranged event in all three situations. Moreover, in the central location test the consumers were served plain cheese samples with water as beverages and in the home test only 20% stated that they had bread, crisp bread or biscuits together with the cheese and 56% chose to drink water. One possible explanation for the contrary effect of social interaction in the wine study (Hersleth et al., 2003) compared with the results from the present cheese study, could therefore Table 5 Average hedonic ratings with standard deviations from the three testing situations Product Average rating Standard deviation Average a Laboratory CL Home Lab. CL Home 1 6.82 6.89 6.91 1.62 1.51 1.76 1.63 4 6.79 6.52 6.40 1.78 1.90 1.99 1.89 11 6.99 7.05 7.16 1.79 1.55 1.60 1.65 14 6.30 6.24 6.17 1.83 1.95 1.79 1.86 8 7.05 6.95 7.14 1.91 1.97 1.64 1.84 9 5.91 5.82 6.14 2.23 2.32 2.17 2.23 Average 1.86 1.87 1.82 CL ¼ central location. a Total averages of the standard deviation per product and per testing situation. Fig. 4. PCA loading-plot on the consumers rating of six cheeses in the three testing. L ¼ laboratory, CL ¼ central location, H ¼ in home. 108 M. Hersleth et al. / Food Quality and Preference 16 (2005) 103–110 be that the arranged wine tasting sessions in the recep- tion room were experienced as more natural than the corresponding cheese testing sessions. Wine tasting for the consumers recruited in California was probably a more ‘‘natural activity’’ than cheese tasting for the consumers recruited in Norway. Another study of the effects of social interaction, physical environment and food choice freedom on consumption in a meal-testing environment was recently published by Weber, King, and Meiselman (2004). They found that social interaction alone had no detectable impact on food consumption. In this case, the reason possibly was that the meal duration was too short to increase consumption. Findings by Feunekes, de Graaf, and Staveren (1995) suggest that increased meal dura- tion is responsible for the increase in consumption ob- served with social interaction. Additionally, the subjects may not have felt comfortable enough with other par- ticipates to socialise freely. de Castro (1994) found in their study that family and friends had a larger impact on food intake than other companions. In the wine study (Hersleth et al., 2003) the wines were rated higher when served with food than without. However, since most consumers in the home situation chose to test the cheese as a single food item, we were not able to test the effect of eating other foods and beverages in this study. Expectations have been thoroughly investigated by Cardello (1992, 1995). When customers are served identical food in different locations, the customers’ existing expectations lead them to rate the food in a non- institutional setting higher than the food in an institu- tional setting (Cardello, 1995; Edwards et al., 2003). In these studies the expectations were both sensory-based and hedonic-based (Cardello, 1992) and the hedonic- based expectations were probably different in each location. It is likely that the expectations in our study primarily were sensory-based as we did blind testing of a single food item in a testing situation as opposed to branded testing, meal testing and/or testing under more natural eating conditions. These sensory-based expec- tations were probably quite equal in all the three testing situations and accordingly resulted in stable hedonic ratings of the cheeses. Another possible reason for the stability of the he- donic ratings in our study could be the consumers’ de- gree of familiarity with the presented products. A study by Pound and Duizer (2000) supports this assumption. Pound and Duizer (2000) measured consumer responses for overall liking of commercial chocolate in four types of testing situations (central location, in-home, teaching laboratory and formal sensory laboratory) and found that testing in all four locations gave similar results. All identifying features were removed from the chocolate prior to testing, so the panellists were not influenced by the brand of the chocolate. However, products tested in the experiment were well-known milk chocolates (Nestl  e, Hershey and Cadbury). The cheeses in our project were also well-known products used in Norway in various meal contexts from early childhood. The consumers may consequently have developed a quite stable preference pattern for these cheeses. The consumers in our study rated the same set of cheeses three times. A possible reason for no effect of the testing situation could therefore be that the consumers suspected the purpose of the study and scored accord- ingly. This would, however, require that the consumers remembered the scores given in the previous sessions. Statements from the consumers during the testing and results from the ANOVA of the acceptance ratings of the cheese samples in the first testing session only, without any possible bias from previous sessions, indi- cate that this was probably not the case. The aspect of product familiarity may also be rele- vant for a discussion of consumer’s consistency. If we do not assume any effect of testing situation, we may con- sider our experiment as repeated hedonic measurements in three separate occasions. K € oster et al. (2002) claim that ‘‘one may not expect that the same population will say the same thing when offered the same set of samples on two separate occasions’’. In our study the consumers showed a clear consistency in ratings between sessions. This result may also be explained by the consumers familiarity with the presented product category, this in contrast to hedonic testing of products with a higher degree of novelty (K € oster, 2003; K € oster et al., 2002; L  evy & K € oster, 1999). Although we have demonstrated a relatively clear consistency in hedonic ratings between sessions, some variation between cheese samples was present as shown in Table 5. Samples 1 and 11 provided the most con- sistent judgements (least standard deviation) and were the most liked samples (in addition to sample 8) in all three testing situations. Sample 9 provided the least consistent judgements (largest standard deviation) and was the least liked sample in all three testing situations. McEwan (1997) published a study about consistency in results over test methods, considering also repeatability of individual consumer judgements. McEwan pointed out, in good correspondence with our results, that the sample that provided the most consistent judgements was one of the most liked samples. Moreover, in our experiment, the consumers were probably least familiar with cheese number 9 as this cheese has the lowest sales figures in the market. Consequently, the relatively large standard deviation across testing situation for sample number 9 and the observed interaction between product and context mainly due to the same sample (Fig. 2), indicate that effects of contextual factors in consumer testing both are dependent on the declared liking for the tested products and the degree of familiarity with the presented product category. M. Hersleth et al. / Food Quality and Preference 16 (2005) 103–110 109 5. Conclusion Changing the environments and the degree of social interaction during a consumer test had no significant effect on hedonic ratings for a selection of semi-hard and hard cheeses in this study. Lack of a natural meal con- text during testing, similar expectations in the three testing situations and high familiarity of the product category may explain the high degree of consistency in hedonic ratings between sessions. The differences be- tween our results and results from related studies should encourage more studies of contextual variables in ar- ranged testing environments and natural eating envi- ronments with product categories of different levels of familiarity. Acknowledgements The authors would like to acknowledge Tine BA, Norway, who contributed with products in this project. In particular we want to express our gratefulness to Mari Austvoll Ilseng for good advice and recommen- dations during the project. In addition, we want to thank the referees for valuable comments during the preparation of the paper. References Bell, R., & Meiselman, H. L. (1995). The role of eating environments in determining food choice. In D. Marshall (Ed.), Food choice and the consumer (pp. 292–310). Glasgow: Blackie Academic & Professional. Cardello, A. V. (1992). Consumer expectations and their role in food acceptance. In H. J. H. MacFie & D. M. H. Thomson (Eds.), Measurement of food preferences (pp. 253–297). Glasgow: Blackie Academic and Professional. Cardello, A. V. (1995). Food Quality: Relativity, context and consumer expectations. Food Quality and Preference, 6, 163–170. Cardello, A. V., Schutz, H. G., Snow, C., & Lesher, L. (2000). Predictors of food acceptance, consumption and satisfaction in specific eating situations. Food Quality and Preference, 11, 201–216. de Castro, J. M. (1994). Family and friends produce greater social facilitation of food intake than other companions. Physiology & Behavior, 56, 445–455. de Castro, J. M., Brewer, E. M., Elmore, D. K., & Orozco, S. (1990). Social facilitation of the spontaneous meal size of humans occurs regardless of time, place, alcohol or snacks. Appetite, 15, 89–101. de Castro, J. M., & de Castro, E. S. (1989). Spontaneous meal patterns of humans: influence of the presence of other people. American Journal of Clinical Nutrition, 50, 237–247. Edwards, J. S. A., Meiselman, H. L., Edwards, A., & Lesher, L. (2003). The influence of eating location on the acceptability of identically prepared foods. Food Quality and Preference, 14, 647–652. Feunekes, G. I. J., de Graaf, C., & Staveren, W. A. (1995). Social facilitation of food intake is mediated by meal duration. Physiology & Behavior, 58, 551–558. Hersleth, M., Mevik, B.-H., Næs, R., & Guinard, J.-X. (2003). Effect of contextual factors on liking for wine––use of robust design methodology. Food Quality and Preference, 14, 615–622. King, S., Weber, A., Meiselman, H.L., & Lv, N. (in press). The effect of meal situation, social interaction, physical environment and choice on food acceptability. Food Quality and Preference. Kozlowska, K., Jeruszka, M., Matuszewska, I., Roszkowski, W., Barylko-Pikielna, N., & Brzozowska, A. (2003). Hedonic tests in different locations as predictors of apple juice consumption at home in elderly and young subjects. Food Quality and Preference, 14, 653–661. K € oster, E. P. (2003). The psychology of food choice: some often encountered fallacies. Food Quality and Preference, 14, 359–373. K € oster, E. P., Couronne, T., L  eon, F., L  evy, C., & Marcelino, A. S. (2002). Repeatability in hedonic sensory measurement: a concep- tual exploration. Food Quality and Preference, 14, 165–176. Lawless, H. T., & Heymann, H. (1999). Sensory evaluation of food. Principles and practices. New York: Chapman & Hall. L  evy, C. M., & K € oster, E. P. (1999). The relevance of initial hedonic judgements in the prediction of subtle food choices. Food Quality and Preference, 10, 185–200. MacFie, H. J. H., Bratchell, N., Greenhoff, K., & Vallis, L. V. (1989). Designs to balance the effect of order of presentation and first-order carry-over effects in hall tests. Journal of Sensory Studies, 4, 129–148. Mardia, K. V., Kent, J. T., & Bibby, J. M. (1980). Multivariate analysis. London: Academic Press. Martens, H., & Næs, T. (1989). Multivariate calibration. Chichester, UK. McEwan, J. A. (1997). A comparative study of three product acceptability trials. Food Quality and Preference, 8(3), 183–190. Meilgaard, M. C., Civille, G. V., & Carr, B. T. (1991). Sensory evaluation techniques. Boca Raton: CRC Press, Inc. Meiselman, H. L. (1992). Methodology and theory in human eating research. Appetite, 19, 49–55. Meiselman, H. L. (1993). Critical evaluation of sensory techniques. Food Quality and Preference, 4, 33–40. Meiselman, H. L. (1996). The contextual basis for food acceptance, food choice and food intake: the food, the situation and the individual. In H. L. Meiselman & H. J. H. MacFie (Eds.), Food choice acceptance and consumption (pp. 139–263). Glasgow: Blackie Academic and Professional. Meiselman, H. L., Johnson, J. L., Reeve, W., & Crouch, J. E. (2000). Demonstrations of the influence of the eating environment on food acceptance. Appetite, 35, 231–237. Peryam, D. R., & Pilgrim, F. J. (1957). Hedonic scale method of measuring food preferences. Food Technology, 11, 9–14. Pound, C., & Duizer, L. (2000). Improved consumer product devel- opment. Part one. Is a laboratory necessary to assess consumer opinion? British Food Journal, 102(11), 810–820. Rozin, P., & Tuorila, H. (1993). Simultaneous and temporal contextual influences in food acceptance. Food Quality and Preference, 4, 11–20. Schutz, H. G. (1999). Consumer data––sense and nonsense. Food Quality and Preference, 10, 245–251. Stone, H., & Sidel, J. L. (1992). Sensory evaluation practices. San Diego, CA: Academic Press, Inc. Weber, A. J., King, S., & Meiselman, H. L. (2004). Effects of social interaction, physical environment and food choice freedom on consumption in a meal-testing environment. Appetite, 42, 115–118. 110 M. Hersleth et al. / Food Quality and Preference 16 (2005) 103–110