JFS: Sensory and Nutritive Qualities of Food
Preference Mapping of Domestic/Imported Jasmine Rice for U.S.-Asian Consumers S. SUWANSRI, J.-F. MEULLENET, J. A. HANKINS, K. GRIFFIN
ABSTRACT: Three domestic and 12 imported commercial Jasmine rice varieties were evaluated by a trained sensory panel and by 105 Asian families who live in the State of Arkansas. Results showed that consumers preferred imported over domestic products. According to consumers, the sensory characteristics most important to the acceptance of cooked Jasmine rice were, in order of decreasing importance, color, favor, aroma, stickiness, and hardness. Using descriptive data, we evaluated predictive models of Jasmine rice’s overall acceptance. These models allowed us to identify important sensory characteristics that encouraged Asian consumers to accept Jasmine rice. Such characteristics included flavor (that is, aroma, aromatics, feeling factors, and aftertaste), texture, and visual attributes. Data collected here could be useful to the U.S. rice industry in developing an understanding of the drivers of Jasmine rice acceptance. Keywords: Jasmine rice, Asian consumers, preference mapping, descriptive sensory analysis
Introduction is, consumption of 150 to 200 lbs per capi- discern differences in the quality of the ICE (ORYZA SATIVA L.) IS A GOOD SOURCE ta) (Goodwin and others 1992). rice they consume, there is a need to as- Rof complex carbohydrates and protein In the past ten y, domestic aromatic sess the likes and dislikes of this popula- and is a staple food for a large portion of rice has accounted for less than 1% of the tion segment and to evaluate the key sen- the world’s population. Over 100 coun- aromatic rice sold in the United States due sory differences between domestic and tries, on every continent except Antarctica, to its dull off-white color and less pro- imported aromatic rice. The objectives of grow rice (IRRI 1993). In the United States, nounced aroma than imported aromatic this study were to assess the acceptability rice production began in South Carolina in rice (Rister and others 1992). The U.S.A. of both domestic and imported Jasmine 1690 (Adair 1972). Nowadays, the U.S. rice Rice Council’s preliminary test of the mar- rice by U.S.- Asian consumers and to corre- production represents around 1.5% of the ket potential for domestic rice varieties late the U.S.- Asian consumers’ accep- world’s rice production. The major rice pro- among U.S.-Asian consumers suggested tance of Jasmine rice to descriptive senso- ducing states are Arkansas, California, that visual and aromatic attributes of both ry data so that the acceptance drivers Louisiana, Mississippi and Texas. uncooked and cooked rice were important could be identified. Annual per capita rice consumption in to the U.S.- Asian consumers’ preferences the United States has risen from 10.25 lbs (Goodwin and others 1992). Imported Jas- Material and Methods in 1980 to 27.1 lbs in 2000 (U.S.A. Rice Fed- mine rice appearance was linked to milling IFTEEN DOMESTIC AND IMPORTED JAS eration 2000). The U.S.A. Rice Federation practices (that is, water mist milling), Fmine rices were purchased in February has set a consumption goal of 30 lbs per which are different from those of the stan- 2000 from either local specialty stores or capita by 2004 (U.S.A. Rice Federation dard U.S. milling process. Additionally, producers. All imported samples pur- 2000). The demand, in particular, has in- rice breeders from the U.S. Dept of Agri- chased were labeled as the “New Crop creased for whole grain foods and for pro- culture attempted to determine the num- 2000” and the “Premium Quality.” Accord- cessed foods using milled, brown, pre- ber of genes controlling the pronounced ing to Narainakorn (1998), these labels in- cooked, and specialty or aromatic rice, aroma of aromatic rice and developed new dicated that the imported Jasmine rices such as Jasmine rice imported mainly from varieties of fragrant rice by adding the aro- were freshly milled and packed from rice Thailand (U.S.A. Rice Federation 1999). ma gene to existing rice cultivars. The re- paddies harvested in 1999 and mixed with Jasmine rice is mostly consumed by sults indicated that a new aromatic rice va- other rice varieties to a maximum of 30 % Asians who emigrated from Southeast Asia riety, Jasmine 85, contained almost twice weight by weight. A list of Jasmine rice Sensory and Nutritive Qualities of Food (Pinson 1994). Asian Americans are the as much of the aromatic chemical 2- samples evaluated is given in Table 1. fourth largest ethnic group in the United Acetyl-1-pyrroline as the other U.S. aro- Three-digit random numbers were gener- States and the fastest growing population matic varieties (Pinson 1994). In a study ated according to Cochran and Cox (1957) in all regions (Campbell 1996). It is pro- reporting on Asian consumers’ preferenc- and assigned randomly for all of the 15 jected that this ethnic group will increase es for rice, Meullenet and others (2001) in- Jasmine rice samples. Three-lb samples by more than 11 million by 2015. Addition- dicated that imported Jasmine rice was were placed in plastic bags (Glad®, Zi- ally, rice consumption by Asian ethnic preferred more than any other rice, in- ploc® Bags) and used for both consumer groups is usually 10 times more than the cluding domestic Jasmine. testing and descriptive analysis. The cod- average for the total U.S. population (that Since the U.S.- Asian population can ed and sealed samples were placed in
2420 JOURNAL OF FOOD SCIENCE—Vol. 67, Nr. 6, 2002 © 2002 Institute of Food Technologists
jfsv67n6p2420-2431ms20010552-MP.p65 2420 8/22/2002, 2:33 PM Jasmine Rice Preference Mapping . . .
Table 1—Jasmine rice samples evaluated plastic airtight storage buckets and stored Product type Code Product Name Manufacturer/ Origin Brand/Importer at 4 °C in a commercial walk-in refrigerator until testing. Domestic dJ Jasmine 85 Lowell FarmTexas dS Jasmine Specialty Rice ArkansasMarketing dT Jasmati Rice Tec Texas Descriptive analysis Imported iA Jasmine Angel Rice Thailand A 9-member panel, employed by the iB Jasmine Bell & Flowers Thailand Univ of Arkansas, Dept of Food Science, iK Jasmine BKM Thailand and well trained in descriptive analysis of iC Jasmine C.T.F. Thailand ® iS Jasmine Dynasty Thailand rice (Spectrum Method , Meilgaard and iG Jasmine Golden Boy Thailand others 1999), evaluated 5 visual, 16 flavor, iM Jasmine Golden Camel Thailand and 11 texture attributes for all 15 Jasmine iD Jasmine Golden Cobra Thailand rice samples. Visual, flavor, and texture iI Jasmine I.T.C. Thailand lexicons were developed by the trained iR Jasmine Riviana Thailand iR1 Jasmine Rose Rice Thailand panelists during 3 orientation sessions iR2 Jasmine Rose Rice Thailand (Meullenet and others 1998). During these orientations, a frame of reference, which represented a variety of Jasmine rice products, was presented to the panel. These reference samples assisted the panel in developing descriptors and ap- propriate descriptive techniques. The vi- sual, flavor, and texture lexicons used to describe the attributes that were evaluat- ed, their definitions, and references used are listed in Tables 2, 3, and 4. Jasmine rice samples were retrieved from cold storage and allowed to temper for approximately 12 h before cooking. Electronic rice cookers (Rice-O-Mat, Na- tional Brand: model #SR-w10F-5 quart ca- pacity; Panason, Cypress, Calif., U.S.A.) were used to cook all the Jasmine rice sam- ples. All the rice samples were prepared using a 1.5:1 water:rice ratio or 300 g of rice in 450 ml of water. The rice sample and wa- ter were emptied into the cooker holding chamber, covered with the vented lid, and the rice cooker was switched on 21 min pri- or to the scheduled presentation time to the trained descriptive panel. The sam- ples remained covered throughout the cooking duration. When cooking was com- pleted, the removable holding chamber was immediately lifted out of the heating chamber to prevent overcooking or scorch- ing. Regarding the presentation to the pan- elists, the samples were dipped into pre- heated (165 °F) 6-oz glass bowls. The heat- ed glass bowls were then placed inside styrofoam cups and covered with coded watch glasses. Jasmine rice samples were presented at 160 °F, and panelists were in- structed to complete their evaluations be- fore the sample reached 140 °F; digital thermometers were provided to each pan- elist. The samples were unimodally (that is, one at a time) presented to the panel- Figure 1—Sample scores and variable loadings overlay for the principal com- ists in individual booths that featured ponents analysis of visual sensory attributes: (a) PC1 compared with PC2; (b) controlled lighting and positive air pres- PC1 compared with PC3. Score and loading plots obtained from sensory de- sure. Panelists scored all attributes on a 0 scriptive attributes using PCA model. Visual attributes data variation described was (a) 55% and 23% for PC1 and PC2, respectively; (b) 55% and 15% for PC1 to 15 numerical scale with one significant and PC3 respectively. Sample name abbreviations can be found in Table 1. digit. Panelists recorded the scores for
Vol. 67, Nr. 6, 2002—JOURNAL OF FOOD SCIENCE 2421 Sensory and Nutritive Qualities of Food
jfsv67n6p2420-2431ms20010552-MP.p65 2421 8/22/2002, 2:33 PM erence s parboiled(5.0) s parboiled (0.0). s parboiled (13.0) s parboiled (3.0) s parboiled s parboiled (1.5). ’ ’ ’ ’ ’ ThaiKitchenJasmin(11.0) Uncle Ben Home-use testing was conducted in the State of Arkansas in the towns of Fay- in the towns of Arkansas the State To Smith. and Fort Springdale, etteville, represent a true cross-section of the popu- 3 lation, Asian consumers living in these areas were recruited via a survey distribut- ed to local oriental supermarkets and res- taurants. The main criterion considered in selecting the subjects was whether or not the consumer and the family consumed Jasmine rice. This consumer group was represented by 15% Islanders (Filipinos and Indonesian), 13% Laotian, 12% Malay- sian, 23% Thai, 12% Continental (Sri Lan- ka, Bangladesh, and Burma), and 25% Taiwan). China and Chinese (Mainland Based on the survey results, the respon- dents were classified into 3 groups: light users (1 to 9 meals per wk, 27%), medium users (10 to 13 meals per wk, 30%), and heavy users (14 to 21 meals per wk, 43%). According to the census of population and housing survey by the Census State Data (Bell 1998), the approximate number of Asian families living in these 3 areas was 3,000 households. Therefore, the mini- mum number of respondents required to achieve the 95 percent level of confidence with an error that did not exceed ±10 per- cent was estimated to be 94 households (Rea and Parker 1992). Since, it was logisti- cally difficult to have all families evaluat- ing all 15 samples, the experiment was planned according to a balanced incom- plete block design arranged in a Latin each attribute on paper ballots. Crackers each attribute on paper (Nabisco Premium Unsalted) and water were provided for panelists to rinse their palates between samples. Sample serving order was randomized across treatment, and all panelists evaluated each of the 15 samples in duplicate on separate testing days. Home-use testing 8/22/2002, 2:33 PM Vol. 67, Nr. 6, 2002 67, Nr. Vol. Observe the sample and determine to whichObserve the sample kernels are split. the degree of roughness.Observe and determine Thai Kitchen Jasmine (2.5) Jasmine (4.0) Thai Kitchen The degree to which grain are cracked.The degree to which kernels are rough.The degree to which kernels shine.The degree to which Uncle Ben the degree of shine.Observe and determine Ben Uncle Ben Uncle Thai Kitchen Jasmine (10.0). The degree to which the sample is visually pure white. to which the sample The degree Uncle Ben together. to which grains stick The degree degree to which with fork to evaluate Fluff sample together. kernels stick Jasmine (1.5). Thai Kitchen — 2 2422 1 4 3 Sample scores and variable loadings overlay for the principal com- Descriptive visual lexicon for Jasmine rice for Jasmine lexicon visual Descriptive — — 5 JOURNAL OF FOOD SCIENCE 2422 Figure 2 (b) ponents analysis of flavor sensory attributes: (a) PC1 compared with PC2; de- PC1 compared with PC3. Score and loading plots obtained from sensory scriptive attributes using PCA model. Flavor attributes data variation described PC1 was (a) 25% and 15% for PC1 and PC2, respectively; (b) 25 and 14% for 1. Table can be found in Sample name abbreviations and PC3, respectively. Surface Roughness V Gloss V Stickiness between Grains V Stickiness Jasmine Rice Preference Mapping . . . . Mapping Preference Rice Jasmine 2 Table Term Definition Ref Degree of Whiteness V Degree of Grain Integrity V
Sensory and Nutritive Qualities of Food jfsv67n6p2420-2431ms20010552-MP.p65 Jasmine Rice Preference Mapping . . .
Table 3—Descriptive aromatics and basic tastes lexicon – Jasmine rice Term Definition Reference
Starchy A1, M1, S1 The aromatics associated with the starch Rice flour paste; of a particular grain source. Rice flour in water. Mixing ratio is 1:1.
Cooked Grain A2, M2, S2 A general term used to describe the aromatics of raw Cereal grain: Bran Buds, or cooked grains, which cannot be tied to a special grain. Air popped popcorn, Cream of Wheat
Feedy A3, M3, S3 The aromatics associated with a mixture of grains Chicken feed; Bran buds reminiscent of animal feed (dusty, musty, sharp). Scorched A4, M4, S4 The aromatic associated with scorching. Scorched popcorn Wet Cardboard/Papery A5, M5, S5 The aromatics associated with early stages of oxidation. Place wet cardboard in a reference jar and sniff. Nutty A6, M6, S6 The aromatics associated with nuts or nutmeats, which cannot be tied to a specific origin. Toasted Wheat Germ Sweet Aromatic A7, M7, S7 The aromatic associated with materials that also have sweet tastes, such as molasses, caramelized sugar, 1-Vanilla; 1-Maltol cotton candy, maple syrup, maltol. Sulfury A8, ,, M8, S8 The aromatics associated with hydrogen sulfide, Boiled eggs,Struckmatch, boiled or rotten eggs. Sewer gas, Cooked cabbage Heated Oil A9, M9, S9 The aromatics associated with fresh oil that has been Heated vegetable oil heated: not indicative of any oxidized or “off” notes. Heated cottonseed oil
Metallic A10, M10, F1, S10 The aromatics associated with metals, tinny, or irony. Pineapple can (for A flat chemical feeling factor stimulated on the sniffing) tongue and teeth by metal (coin, tin foil). Tin foil (bite down on it several times) Burlap A11, M11, S11 The aromatics associated with burlap. Burlap rice bag Floral/Mint A12, M12, S12 The aromatics associated with a non-specific floral note and sometimes described as mint. Jasmine scent Woody A13, M13, S13 The aromatics associated with dry fresh cut wood; Wood chips, Beta ionone, balsamic or bark-like Toothpicks Dairy Note A14, M14, S14 The aromatics associated with an off or negative note reminiscent of soured or old dairy products. Texmati rice Hot Plastic A15, M15, S15 The aromatics reminiscent of hot rubbers, vinyl or plastic. –- Astringency F2 The chemical feeling factor associated with the shrinking Grape juice (Welch’s); and puckering of the tongue caused by substances such Tea bag / 1- hourh soak as tannins or alum.
Sweet B1, S16 The basic taste on the tongue stimulated by sugar and high Solutions of sucrose in potency of sweeteners. spring water Salt B2, S17 The basic taste on the tongue stimulated by sodium chloride. Solutions of sodium chloride in spring water Sour B3, S18 The basic taste on the tongue stimulated by acids. Solutions of citric acid in spring water Bitter B4, S19 The basic taste on the tongue stimulated by solutions or Solutions of caffeine in substances such as quinine and certain other alkaloids springwate
Abbreviations: A= Aroma B = Basic tastes, M =Aromatics F= Feeling Factors, S= Aftertaste
Square (Cochran and Cox 1957; Kuehl block 3 times. Therefore, homogeneous tremely” to “like extremely” was used to 1994). Thus, to obtain a data set that set of the samples and precise sample assess the acceptance of the various prod- would be planned according to the chosen comparisons would be obtained (Kuehl ucts (that is, overall acceptance, appear- design, 105 households (that is, 15 blocks ´ 1994). ance, aroma, flavor, and texture). A 5-point 7 replicates) were recruited (Rea and Park- A Jasmine rice sample was delivered “just right” scale was used to assess the er 1992; Cochran and Cox 1957). In this weekly to the respondents’ homes from 1 appropriateness of the intensity of more design, the number of Jasmine rice sam- March 2000 to 17 April 2000. The respon- specific attributes (that is, color, aroma, ples per block was less than the total num- dents received 800 g of the various Jas- flavor, stickiness, and hardness) from ber of samples (15 samples). In this case, a mine rice samples during 7 consecutive “much too little” to “much too much” (Meil- subset of 7 from 15 samples was arranged wk. Each Jasmine rice sample was accom- gaard and others 1999). and the order in which each household panied by a cooking instruction as well as would receive the rice sample was com- an evaluation form. The questionnaire Data analysis pletely randomized among the house- was designed to assess the acceptance of Descriptive analysis data were aver- holds. All 15 Jasmine rice samples were each sample by the respondents. A 9- aged by using the PROC MEANS proce- equally replicated (n = 49), and each sam- point verbal hedonic scale (Meilgaard and dure of SASâ (1999) across panelists and ple pair occurred altogether in the same others 1999) anchored from “dislike ex- replications and Fisher’s Least Square Dif-
Vol. 67, Nr. 6, 2002—JOURNAL OF FOOD SCIENCE 2423 Sensory and Nutritive Qualities of Food
jfsv67n6p2420-2431ms20010552-MP.p65 2423 8/22/2002, 2:33 PM rough) ..tight) .sticks) ..much) ...much) .difficult) .tight mass) ..difficult) …… . ………… mes and evaluate. ………………………… ……………………… ……………………… ……………………… ……………………… …………………… trained panel). Data was centered prior to PLS regression so that all results were in- terpreted in terms of variation around the mean. Predictive variables were stan- dardized by weighting with the standard deviation, so that all variables were given the same chance to influence the predic- tion of the consumer data, regardless of differences in variable magnitude. The full cross-validation method was used so that the same samples were used for both calibration and validation of the models. With full cross-validation, each sample is removed one at a time from the sample set, a new calibration is performed, and a predicted score is calculated for the sam- ple that has been removed. The Root ………………………… ace 1/2 teaspoon of sample in mouth. ace 1/2 teaspoon of sample â echnique teeth to evaluate. (none evaluate. (none to prepare sample for swallowing. Feel mass with tongue and quickly Feel mass with tongue mass is sticking evaluate how tightly the and holding together. (loose (easy lips. product remains on the (fall (loosemass sample mass. (easy 8/22/2002, 2:33 PM scales) using Partial ” just right “ ). In a similar manner, overall accep- overall a similar manner, ). In â Consumer scores were averaged using tance was predicted from sensory descrip- tive attributes, and acceptance of the appearance, flavor and texture predicted from the corresponding descriptive panel data (for example, appearance acceptance from visual descriptors evaluated by the Least Squares (PLS) regression (Unscram- bler data was determined (that is, model with a total residual variance close to 0 or a large total explained variance). the PROC MEANS procedure of SAS (1999) for each of the 15 samples (n = 49). Consumer overall acceptance was predict- ed from consumer attribute data (that is, 5-point Vol. 67, Nr. 6, 2002 67, Nr. Vol. The amount of product packed into the crowns sample 10-15 times, expectorate Chew of the teeth after masticationThe amount and degree of residue felt by the tongue when moves over the surface Swallow the sample and feel the surface to of the mouth with the tongue and feel the surface of the crowns of the The number of chews required to prepare theThe number of chews required to prepare sample for swallowing.mouth the in sample of teaspoon 1 Place and count the number ofchews required The amount of roughness perceived on thesurface of the chewed sample.the large lumps, are looking for You Hint: Chew sample with molars and evaluate The force required separating the jaws duringmastication. Chew sample 2-3 times and evaluate the the irregularities on the surface of the to pull your jaw apart. force required The degree to which the unchewed sample The degree to which holds or sticks together.the sample adheresThe degree to which lips.the to lips, release Compress sample between Pl chewed sampleThe amount that the holds together. Chew sample with molar teeth bumps, hills and valleys, etc. and evaluate the degree to which the ti 15 to up (smooth The force required separating the fingers separating the The force required the thumb sample between after compressing forefinger.and evaluate the force and forefinger and Compress thumb 5 kernels between required to separate your fingers. — of the mouth. 7 , version 7.5, â 2424 9 3 1 2 11 T 10 8
4 Descriptive oral texture lexicon for Jasmine rice for Jasmine lexicon texture oral Descriptive — JOURNAL OF FOOD SCIENCE Macro Roughness of Mass T Residual Film Toothpack T Toothpack 2424 ference (LSD) tests (a = 0.05) performed to determine significant differences be- tween samples. Principal component anal- ysis (PCA) (Unscrambler Camo, Norway) was used so that all com- plex multidimensional data sets (that is, descriptive variables) were visualized. The PCA can be translated as finding the lin- ear combination of the initial variables that contribute most to making the sam- ples different from each other (Camo 1999). The sensory attributes were stan- dardized prior to the analysis by weighing variables by their standard deviations. For interpretation of the results, the optimal number of the principal components that explained most of the information in the Toothpull T Toothpull Residul Number of Chews T Hardness T Initial Cohesion T Partial CompressionAdhesion to Lips T First Bitr Chew Initial T Manual Stickiness Jasmine Rice Preference Mapping . . . . Mapping Preference Rice Jasmine 4 Table Term Definition T
Sensory and Nutritive Qualities of Food jfsv67n6p2420-2431ms20010552-MP.p65 Jasmine Rice Preference Mapping . . .
Mean Square Error of Prediction (RMSEP) greatly from the common model (that is, Table 5—Means of descriptive is subsequently calculated. The predictive with all samples), it indicates that the sam- visual attributes models were optimized using the Jack-knif- ple or samples removed have seriously af- Visual Attributes ing method available as an option of the fected the common model. The approxi- Sample Unscramblerâ. Jack-knifing is a procedure mate uncertainty variance of the Code V1 V2 V3 V4 V5 that is designed to test the significance of regression coefficients can then be esti- dJ 7.64 6.34 2.11 3.12 8.67 the model parameters and is performed mated, and a t-test can then be performed dS 6.32 2.54 2.00 3.24 2.90 during cross-validation. During cross-vali- for each element relative to its estimated dT 8.51 3.90 1.91 3.26 7.42 dation, if a perturbed segment differs uncertainty variance, giving a significant iA 9.83 5.94 2.14 3.24 10.33 iC 9.76 5.75 2.28 4.02 10.66 iD 9.61 4.71 2.12 3.06 10.17 iG 10.68 6.47 2.18 3.56 11.51 iI 10.01 5.66 1.91 3.30 10.01 iK 10.07 4.82 1.98 3.28 9.20 iM 10.12 6.02 1.86 3.07 11.27 iR 9.91 5.07 2.19 3.64 9.10 iR1 10.01 5.81 2.08 3.49 11.11 iR2 9.99 5.25 1.79 3.53 9.50 iS 9.74 4.61 1.64 3.33 5.99 LSD3 0.47 0.99 0.44 0.61 1.09 1Definitions are presented in Table 2. 2Jasmine descriptions are found in Table 1. 3Least Significant Difference (a = 0.05)
level for each parameter. All parameters for which p < 0.05 were kept in the mod- el. This allowed for removal of predictive variables either not influencing the pre- diction or creating noise in the model. This procedure reduced “the uncertain- ty in the prediction models” (Camo 1999) and, in most cases, improved the Figure 3—Sample scores and variable loadings overlay for the principal compo- validation statistics. nents analysis of texture sensory attributes: PC1 compared with PC2. Score Overall consumer acceptance was and loading plots obtained from sensory descriptive attributes using PCA model. also predicted with similar modeling Texture attributes data variation described was 53% and 31% for PC1 and PC2, respectively. Sample name abbreviations can be found in Table 1. methods using consumer acceptance of color, aroma, flavor, and texture as pre- dictors. Finally, consumer acceptance of aroma, flavor, and texture was predicted using the corresponding descriptive analysis attributes. Similarly to work published by Meullenet and others (2001), the weighted regression coeffi- cients expressed numerically the rela- tionship between the predictors (that is, descriptive attributes) and the consum- er responses. Weighted regression coeffi- cients (that is, coefficient weighted by the variable mean value) were used toe- express their relative influence indepen- dently of differences in intensities of de- scriptive attributes.
Results and Discussion
Descriptive analytical results Descriptive attribute means are pre- sented in Tables 5, 6a to 6c, and 7 for vi- sual, flavor, and texture attributes, re- spectively. Fisher’s LSD tests were Figure 4—Predicted compared to observed overall acceptance by Asian con- sumers. Predicted scored obtained from sensory descriptive attributes using performed for each attribute to evaluate partial least squares regression. Sample name abbreviations can be found in significant differences (a = 0.05) between Table 1. samples. Mean scores of the sensory at-
Vol. 67, Nr. 6, 2002—JOURNAL OF FOOD SCIENCE 2425 Sensory and Nutritive Qualities of Food
jfsv67n6p2420-2431ms20010552-MP.p65 2425 8/22/2002, 2:33 PM 1 8/22/2002, 2:33 PM 1 1 Aroma Attributes Aromatic Attributes Vol. 67, Nr. 6, 2002 67, Nr. Vol. Basic Tastes, Feeling Factors, and Aftertaste Attributes Aftertaste Factors, and Feeling Basic Tastes, — 2426 M1 M2 M3 M5 M6 M7 M10 M11 M12 M13 M14 B1 B3 F1 F2 S1 S2 S5 S10 S11 S16 S18 0.64 1.00 0.90 0.76 0.23 0.41 0.62 1.11 0.62 0.13 0.83 0.52 0.41 0.62 0.22 0.44 0.63 0.32 0.68 0.43 0.40 0.36 Means of descriptive flavor attributes (Aromatics) Means of descriptive flavor attributes Means of descriptive flavor attributes (Aroma) attributes flavor of descriptive Means Means of descriptive flavor attributes (basic tastes, feeling factors, and aftertaste) Means of descriptive flavor attributes (basic tastes, feeling factors, and 0.81 1.09 1.02 0.54 0.54 0.62 0.50 0.13 1.25 0.74 0.25 0.74 A1 A2 A3 A5 A6 A7 A8 A9 A11 A12 A13 A14 — — — JOURNAL OF FOOD SCIENCE 2 2 1 3 3 3 Jasmine descriptions are found in Table 1. Table Jasmine descriptions are found in Least Significant Difference (a = 0.05) Definitions are presented in Table 3. Table in Definitions are presented Least Significant Difference (a = 0.05) Least Significant Difference (a = 0.05) Definitions are presented in Table 3. Table Definitions are presented in 1. Table Jasmine descriptions are found in Definitions are presented in Table 3. Table Definitions are presented in 1. Table Jasmine descriptions are found in DJDSDTIAIBIC 0.86ID 1.04IG 1.12II 0.38 0.73 0.92IK 1.40 0.83IM 0.95IR 1.44 0.79 1.02iR1 1.92 0.86 0.56iR2 1.40 0.52 1.13IS 3.01 0.59 0.99 1.12LSD 3.12 0.55 1.06 1.57 3.07 1.15 1.21 0.74 0.91 2.70 0.56 3.00 1.37 1.27 1.71 0.59 1.57 2.96 1.94 0.94 1.14 2.97 0.78 1.47 1.36 2.63 2.92 1.68 0.56 1.01 3.01 2.84 1.72 0.81 0.37 2.81 1.33 3.11 2.02 0.88 2.73 2.99 1.22 1.52 1.08 2.82 3.03 1.01 0.58 3.13 1.17 0.91 2.79 3.05 2.39 2.48 0.53 0.71 3.06 2.46 2.91 0.39 0.58 2.85 2.76 3.03 0.64 2.39 1.22 0.13 2.38 0.62 0.48 2.66 0.21 0.62 2.42 1.12 0.24 2.31 2.70 1.03 0.58 0.62 0.81 0.50 3.03 0.79 0.88 1.06 2.89 0.29 0.73 0.89 0.31 0.79 0.84 2.46 0.60 0.14 3.02 0.24 0.85 0.57 0.24 2.12 0.83 1.15 2.77 0.75 0.80 1.06 0.48 3.24 0.32 1.02 0.47 2.73 0.31 0.98 0.64 0.32 2.71 0.43 0.52 0.81 0.96 0.54 0.55 1.02 0.36 0.92 1.07 1.03 0.71 0.59 0.87 0.30 1.01 0.59 0.75 0.42 0.21 dJdSdTiAiBiC 3.73iD 1.58iG 2.88iI 2.35 3.27iK 1.98 3.61iM 2.63 3.46iR 0.77 3.59 2.75iR1 1.95 4.07 2.32iR2 1.53 2.54 3.41iS 1.40 0.19 3.60 3.10LSD 1.80 3.48 1.51 2.58 0.92 0.67 3.28 3.29 2.92 0.19 0.14 1.40 2.05 3.57 0.88 0.49 1.46 2.35 0.00 1.53 3.46 2.42 0.58 1.98 0.93 0.94 0.19 0.00 2.47 0.91 0.00 1.87 0.39 0.00 0.23 0.00 2.01 1.01 0.38 5.09 1.46 0.00 0.00 0.22 3.81 0.96 0.00 0.22 4.73 1.36 0.67 0.38 1.38 0.00 4.42 0.00 0.00 2.08 1.61 0.00 2.10 0.00 5.02 1.16 4.51 2.88 0.00 0.00 0.00 5.12 0.56 0.19 2.38 0.00 0.72 5.07 0.00 3.02 0.00 0.00 0.00 1.93 4.42 0.18 0.43 5.29 0.28 2.64 0.19 0.54 4.80 1.24 0.00 0.38 4.79 0.98 0.50 5.56 3.94 0.00 0.00 0.25 3.69 5.01 0.83 0.00 0.00 2.35 4.27 0.00 1.04 3.70 0.22 3.57 0.00 0.00 0.51 4.16 0.00 0.28 0.61 0.00 3.15 0.50 0.00 0.41 0.00 0.17 0.00 0.00 0.28 0.00 0.00 0.00 0.48 0.00 0.00 1.61 0.71 0.83 0.77 0.50 0.00 dJdSdTiA 2.47iB 0.79iC 1.57 1.83iD 1.98 2.52iG 2.18 2.27 0.87iI 2.08 1.80 2.44iK 3.12 1.14 1.54 0.33iM 2.87 0.63 0.67 0.36iR 2.32 0.31 0.84 2.37 2.30iR1 2.74 3.11 0.50 2.54 0.28iR2 1.26 1.79 0.28 0.39 2.06 0.23iS 0.96 2.46 2.05 0.93 0.00 0.32 2.77LSD 0.47 1.23 0.00 0.61 0.53 2.71 1.78 2.34 0.56 0.00 0.21 1.09 0.57 2.33 0.19 2.14 0.18 1.41 0.00 0.00 3.04 0.26 1.61 0.00 0.17 0.37 0.87 0.56 0.87 0.68 1.26 0.00 1.81 0.11 0.16 0.00 0.38 0.28 1.66 0.27 0.00 0.79 0.18 0.25 2.51 0.00 0.22 0.28 0.42 0.00 1.24 1.19 1.12 1.36 0.22 0.00 0.00 2.33 0.00 0.61 0.24 0.00 2.87 1.27 0.51 0.52 0.42 0.61 0.62 0.00 2.13 0.56 0.53 0.20 0.00 0.38 0.39 2.42 0.00 0.33 0.00 0.00 0.38 1.60 0.48 0.00 1.25 0.00 0.00 0.00 2.56 0.46 0.00 2.55 0.78 0.00 2.68 0.00 0.00 0.77 1.19 3.87 0.00 0.38 0.00 0.38 3.72 0.00 0.21 0.41 0.95 0.51 0.17 3.37 0.00 0.57 0.18 0.00 0.39 0.00 1.33 0.41 0.00 0.00 4.18 0.18 0.00 0.41 0.47 0.80 0.50 1 2 3 2426 Code 1 2 3 6c Table Sample Code 1 2 3 6b Table Sample Code Jasmine Rice Preference Mapping . . . . Mapping Preference Rice Jasmine 6a Table Sample Only detectable attributes are listed
Sensory and Nutritive Qualities of Food jfsv67n6p2420-2431ms20010552-MP.p65 Jasmine Rice Preference Mapping . . .
Table 7—Means of descriptive texture attributes Texture Attributes1 Sample 2 Code T1 T2 T3 T4 T5 T6 T7 T8 T9 T10 T11 dJ 6.89 6.46 9.71 4.86 6.39 5.39 6.22 2.86 17.00 2.28 1.98 dS 5.76 4.22 7.05 5.46 5.58 4.39 6.52 3.35 18.00 3.10 3.30 iA 7.17 6.94 10.74 4.99 6.68 5.76 6.31 3.03 17.39 2.36 2.12 iB 7.63 7.18 11.34 4.54 6.33 5.76 6.23 3.08 16.89 2.58 2.16 iC 7.12 6.74 11.21 4.51 6.60 5.72 6.02 2.87 17.22 2.09 1.91 iD 7.16 6.96 10.91 5.23 6.63 6.17 6.32 3.81 17.77 4.10 3.64 iG 7.17 6.41 11.23 4.88 5.76 5.39 6.12 3.01 17.78 2.21 2.03 iI 6.72 6.59 10.41 4.82 6.30 5.62 6.31 3.04 16.94 2.38 2.12 iK 7.17 6.72 11.28 4.93 6.13 5.68 6.08 3.09 17.61 2.08 1.82 iM 7.56 7.19 11.75 4.04 6.42 6.33 6.17 3.11 16.44 2.14 2.01 iR 7.24 7.25 9.77 5.65 6.49 5.91 6.37 3.64 17.53 4.95 3.86 iR1 7.52 7.01 11.77 4.43 6.37 6.05 5.98 3.02 17.94 2.03 1.99 iR2 6.85 6.82 10.81 4.61 6.64 5.31 6.32 3.26 16.78 2.43 2.03 iS 6.33 4.99 7.78 5.79 5.41 4.48 6.54 2.36 18.50 1.99 1.72 LSD 3 0.41 0.45 0.91 0.48 0.55 0.59 0.32 0.28 1.10 0.49 0.25 1Definitions are presented in Table 4. 2Jasmine descriptions are found in Table 1. 3Least Significant Difference (a = 0.05)
tributes that were found to be likely unde- tectable (that is, mean < 0.7 or detected
only by 1 panelist), heated oil (A9), hot plastic (A15, M15), sulfur note (M8), and flo- ral note (M12)) were excluded from the data set prior to PCA. The overlaid score and loading plots are presented in Figure 1a and 1b, Figure 2a and 2b, and Figure 3, for appearance, flavor, and texture at- tributes, respectively. PCA applied to visual descriptive at- tributes of 15 Jasmine rice samples indi- cated that 3 components explained a total of 93% of the variation. The map of Jas- mine rice samples (that is, score plot) andthe first 2 components is shown in Fig- ure 1a. Principal component 1 (PC1) was found to account for 55% of the variation in the data, while PC2 accounted for 23% of the variation. Most of imported sam- ples, except Dynasty (iS), Golden Cobra (iD), and BMK (iK), laid close to each other along the X-axis and were similar for all vi- sual attributes (Table 5). The highest de- gree of whiteness, stickiness between Figure 5—Weighted regression coefficients for the partial least squares re- grains, and gloss were found in Golden gression model predicting overall acceptance from visual, flavor, and texture descriptors. Descriptor name abbreviations are defined in Tables 2, 3, and 4 Boy (iG), and the highest degree of grain respectively. integrity and surface roughness was found for C.T.F. (iC). Dynasty (iS) and 2 domestic samples (that is, Specialty rice, dS, and Rice Tec, dT) laid farthest away from the other samples. It was also found ples, except Dynasty (iS), Golden Cobra dS, and Rice Tec, dT) laid farthest away that replicated samples (Rose Rice: iR1 along the X-axis and were similar for all vi- from the other samples. It was also found and iR2) were identical in their visual atvi- sual attributes (Table 5). The highest de- that replicated samples (Rose Rice: iR1 sual attributes (that is, loading plot) for gree of whiteness, stickiness between and iR2) were identical in their visual at the first 2 components is shown in Figure grains, and gloss were found in Golden tributes. However, Fisher’s LSD test at a = 1a. Principal component 1 (PC1) was Boy (iG), and the highest degree of grain 0.05 (Table 5) indicated that panelists per- found to account for 55% of the variation integrity and surface roughness was ceived these 2 samples (iR1 and iR2) dif- in the data, while PC2 accounted for 23% found for C.T.F. (iC). Dynasty (iS) and 2 ferently for glossiness. Along the second of the variation. Most of imported sam- domestic samples (that is, Specialty rice, component, degree of whiteness (V1),
Vol. 67, Nr. 6, 2002—JOURNAL OF FOOD SCIENCE 2427 Sensory and Nutritive Qualities of Food
jfsv67n6p2420-2431ms20010552-MP.p65 2427 8/22/2002, 2:33 PM , ) ) ) ), ). 7 5 7 6 14 11 , M , M 7 ), and 7 14 ), cohe- 3 ). Another ) was very 1 ), initial co- 10 1 ) were most ). 3 10 ). ), and the samples 9 18 ), cooked grain aroma ). In contrast, dT load- 5 12 ). For PC2, the attributes ). For ) was most highly nega- 8 ) and residual film (T) and residual 14 10 ), cohesiveness of mass (T 5 ), adhesion to lips (T 2 ) aromatics (that is, negatively) (that ) aromatics 6 ), macro roughness of mass (T The texture descriptive analysis re- Figure 4 is a graphical representation ) and adhesion to lips (T ), and metallic aftertaste (S ), and metallic aftertaste )(Figure 2b) loaded most highly on PC3. 4 1 2 7 imported sample (iM) exhibited high imported sample (iM) soured dairy aroma. The other 2 domestic types (dJ and dT) did not load highly on this PC. Thus, the third component gave more details about the similarities and dif- ferences of these domestic samples with imported Jasmine rice samples. The at- tributes sweet aroma/aromatic (A sults indicated that 2 components ex- plained 84% of the variation. The sample map (that is, score plot) and texture at- tributes (that is, loading plot) for PC1 and PC2 are shown in Figure 3. These compo- nents described, 53% and 31% of the vari- The attributes posi- ation, respectively. tively correlated with PC1 were hardness (T (A (Figure 2a). The sample score of 1 import- The sample score of (Figure 2a). and (iD) exhibited high starchy ed sample after- aromatics and metallic cooked grain was sig- domestic sample (dS) taste. One It ex- different from all others. nificantly nutty and woody aroma and hibited high aroma (A cardboard Wet feedy aromatics. tively correlated to PC2. Even though the tively correlated to PC2. aroma (A intensity of starchy of the observed consumer overall accep- and metallic aftertaste (S and metallic aftertaste and floral aroma (A and sour aftertaste (S Prediction of overall acceptance from descriptive profiles using partial least squares regression nutty (M nutty siveness (T number of chews (T M One domestic (dJ) and 2 imported prod- ucts (iC and iS) loaded similarly and in the vicinity of sweet aroma/aromatic (A ed close to dairy aroma/aromatic (A whose scores were the highest were dT, iS, dT, the highest were were whose scores and dS. Negatively correlated with this manual stickiness (TPC were and toothpull (T The samples whose scores were the high- Other imported sam- iR and iD. est were ples scored negatively on PC2; these were samples with higher manual stickiness (T with the highest positive loads were toothpack (T hesion (T highly positively correlated to PC2, while highly positively correlated dairy aroma (A high in iK (Tables 6a to 6c), the sample high in iK (Tables that this im- scores along PC2 suggested exhibited high wet ported sample also cardboard aroma (A 8/22/2002, 2:33 PM ) and 3 intensity ” ), feedy (M 13 ) (that is, positively) 8 just right “ For the flavor descriptive attributes, 3 components accounted for only 54% of the variation in the data. The score and loading plots are shown in Fig. 2a and 2b. PC1 was found to account for 25% of vari- ation in the data, while PC2 and PC3 were found to account for 15% and 14% of the The flavor descrip- variation, respectively. were starchy aroma/aromatic/aftertaste and sulfur aroma (A grain integrity than the other 2 domestic 5). products (Table and woody aroma (A ) 4 Vol. 67, Nr. 6, 2002 67, Nr. Vol. — ), and gloss 2 2428 Weighted regression coefficients for the partial least squares re- Weighted regression coefficients for Weighted regression coefficients for the partial least squares re- ) and surface roughness (V 3 — — JOURNAL OF FOOD SCIENCE ) loaded negatively, whereas grain in- grain whereas ) loaded negatively, 5 stickiness between grains (Vstickiness between grains (V loaded positively. Two imported samples Two loaded positively. iC, and Riviana, iR) exhibited more (C.T.F., grain integrity and surface roughness than the other samples. The plots of PC1 compared with PC3 (Figure 1b) showed that Jasmine 85 (dJ) scored similarly to some imported samples (Golden Cobra, in- iA) in terms of grain and Angel Rice, iD, that also showed The mean score tegrity. this domestic sample (dJ) exhibited higher 2428 Figure 7 form gression model predicting the acceptance of Jasmine rice appearance ab- the visual attributes evaluated by a descriptive panel. Descriptors name 2. defined in Table are breviations Figure 6 assessment scores for color, aroma, flavor, stickiness, and hardness. flavor, aroma, for color, assessment scores Jasmine Rice Preference Mapping . . . . Mapping Preference Rice Jasmine tegrity (V gression model predicting overall acceptance from the gression model predicting overall acceptance
Sensory and Nutritive Qualities of Food jfsv67n6p2420-2431ms20010552-MP.p65 Jasmine Rice Preference Mapping . . .
tance of Jasmine rice samples in contrast to described only 3% and 15% of the vari- ed Jasmine rice. the predicted scores calculated using descrip- ation in consumer overall acceptance By evaluating the weighted regression tive attributes as predictors in a partial least and descriptive attributes, respective- coefficients relating the response variable squares regression model. Visual, flavor (that ly. Results showed that this consumer (that is, the overall acceptance by Asian is, aroma, aromatics, feeling factors, and af- group preferred most of the imported consumers) to each predictive variable tertaste) and texture attributes were used in Jasmine rices. One imported sample (that is, descriptive attributes), influential this analysis. Two principal components that (iS) and all domestically grown Jas- sensory attributes can be defined. Figure 5 explained most of variation in the sensory mine rice were found to be less accept- is a graphical representation of the weight- data were retained in the regression model. able (Figure 4). Meullenet and others ed regression coefficient for each of the de- PC1 described 95% of the variation in con- (2001) hypothesized that Asian con- scriptive sensory attributes retained in the sumer overall acceptance and 50% of the vari- sumers expected some specific senso- model predicting overall coked rice accep- ation in descriptive attributes, whereas PC2 ry characteristics pertaining to import- tance after stepwise selection using Jack- knifing. In this study, the most important senso- ry attributes of Jasmine rice were defined as those with weighted regression coeffi- cients greater than 0.1 (that is, absolute
value from Figure 5) such as sulfur (A8), woody (A13), wet cardboard (S5), and de- gree of whiteness (V1). Other important at- tributes were defined as those with regres- sion parameters greater than 0.03. Greater
scores for adhesion to lips (T3), starchy (A1), sulfur aroma (A8), burlap note (M11), degree of whiteness (V1), and glossiness (V5) were associated with greater overall acceptance while greater macro roughness of mass
(T7), floral note (A12), woody (A13) and feedy aroma (M3), and wet cardboard after- taste (S5) contributed to lowering accep- tance scores. Contrary to previous studies (Meullenet and others 2001), cooked grain intensity, bitter, and nutty notes were not significant in determining the overall ac- ceptance of Jasmine rice by Asian consum- Figure 8—Weighted regression coefficients for the partial least squares re- ers. It should not be surprising since all the gression model predicting the acceptance of Jasmine rice flavor from the fla- vor attributes evaluated by a descriptive panel. Descriptors name abbrevia- samples studied here were Jasmine aro- tions are defined in Table 3. matic rices, whereas non-aromatic and aro- matic samples were included in previous reports (Meullenet and others 2001). An in- crease in intensities for floral, woody, and wet cardboard resulted in a decrease in ac- ceptance. High intensities of floral, woody, and wet cardboard notes were exhibited in iS, dJ, dT, and dS. Consumers also com- mented on the foul smell when they tested any one of these samples. Although it would be hard to hypothesize the origin of the floral and woody notes, it seems that the rice samples exhibiting wet cardboard notes were slightly oxidized. This could be a result of storage, processing conditions, or inherent qualities of the cultivars. How- ever, since the samples selected were com- mercial samples, this question would re- main unanswered.
Prediction of overall acceptance from attribute acceptance Figure 6 is a graphical presentation of Figure 9—Weighted regression coefficients for the partial least squares re- weighted regression coefficients for a mod- gression model predicting the acceptance of Jasmine rice texture from the texture attributes evaluated by a descriptive panel. Descriptors name abbre- el predicting overall consumer acceptance viations are defined in Table 4. from color, aroma, flavor, stickiness, and
Vol. 67, Nr. 6, 2002—JOURNAL OF FOOD SCIENCE 2429 Sensory and Nutritive Qualities of Food
jfsv67n6p2420-2431ms20010552-MP.p65 2429 8/22/2002, 2:33 PM ) ) - 5 7 ) were 13 IMPORT ), cohe-
3 , M 13 ), initial cohe- 1 ) were found to ) were ) and wet card- 8 3 PREFERRED
), cohesiveness of mass 5 Conclusions ) detected was less than 1.3 ) detected ), and wet cardboard (S wet cardboard ), and ) notes were detectable in ) notes were detectable 7 3 5 CONSUMERS ), and roughness of mass (T ), adhesion to lips (T
4 2 ed Jasmine rice more than the do- SIAN ), and toothpull (T Manual stickiness (TManual 6 sion (T mestic products. The most important factors in determining the acceptance of Jasmine rice were, in decreasing or- aroma, flavor, color, der of importance, notes, resulted in a decrease in accep- in a decrease in notes, resulted aromat- though the sweet tance. Even ic note (A had negative impacts on the accep- tance of Jasmine rice texture by Asian consumers (Figure 9). These findings confirm PCA results (Figure 3) and confirmed results reported by Meul- lenet and others (2001), which indicat- ed that higher sample cohesiveness contributed to increased consumer ac- ceptance of texture, while higher sam- ple macro roughness of mass contrib- uted to decreased acceptance. This ex- model (RMSEP = 0.39), however, plained only 61% of the variation in consumer acceptability for Jasmine rice texture. Further research is neces- sary to confirm the factors that signifi- cantly affect the acceptance of rice texture. siveness (T A Prediction of Jasmine rice texture acceptance from texture descriptors undetectable in most imported Jas- undetectable in most in all domestic mine rice but present imported sample Jasmine rice and one (Dynasty). Feedy (M (T units (Table 6a), it had a significant 6a), units (Table of the consumer acceptance impact on (A notes Woody the aroma. both domestic and imported Jasmine both domestic and of these findings rice samples. Most the results ob- are in agreement with studies by Meul- tained from previous except for sul- lenet and others (2001), sulfur intensities fur notes. Similar rice using were detected in Jasmine the descriptive analysis from both studies, but the importance of this pa- rameter was found to be much more Re- study. in the present prominent sults from this study showed that sul- fur contributed as a major determi- nant of flavor acceptance, opposite to the previous reports. have positive impacts, whereas hard- ness (T board (S feedy (M feedy ), 13 8/22/2002, 2:33 PM ) were ) posi- , M 4 8 ), feedy 13 13 ), grain in- ), grain 2 , M 13 ) notes contribut- 5 ), woody (A 7 ), woody (A ) and sulfur note (A 7 1 , S 1 , M 1 ), and surface (V roughness 3 s LSD test at a = 0.05 (Tables 6a to 6c). s LSD test at a = 0.05 (Tables ’ s Jasmine (dS). This sample was found to (dS). s Jasmine ’ ), and wet cardboard (S The intensity of the cooked grain note Increase in intensities of attributes, such The prediction of flavor acceptance from 3 (that is, equivalent to popcorn-like aroma in (that is, equivalent to popcorn-like aroma our methodology) was found to be different among the Jasmine rice samples according to Fisher Some samples exhibited significant differ- ences in terms of starchy and sulfur notes, but did not for cooked grain intensities. This result showed that cooked grain note or pop- corn-like is not the only significant flavor characteristic of Jasmine rice as previously re- ported (Buttery and others 1983; Pinson 1994). Furthermore, the intensity of this at- tribute was not significantly higher in Jas- mine 85 (dJ) than in other domestic samples (that is, dS and dJ). These results are incon- sistent with those of Pinson (1994), who re- ported that Jasmine 85 contained nearly twice as much of the aromatic chemical re- sponsible for the popcorn-like aroma than other U.S. aromatic varieties. as sweet aromatic (A exhibit very low intensities for starchy and exhibit very low intensities for starchy 6a to 6c), undetectable sulfur notes (Tables re- thus, the lower overall acceptance score ported by consumers (Figure 4). ed to lowering acceptance (Figure 8). The ed to lowering acceptance (Figure 8). starchy and sulfur notes were readily detect- ed in most of the samples, except in Specialty rice not important contributors (Figure 7). These contributors (Figure not important Meullenet results from findings confirmed Goodwin and others and others (2001) and others (1992) conducted (1992). Goodwin and market potential for do- a study assessing the among Asian Americans mestic rice varieties one of important at- and concluded that for Asian consumers tributes of cooked rice study the present from Results was the color. 63% of Asian consum- showed that 31% and for the appear- ers expressed a discontent the domestic rice sam- ance of dJ and dS; 2 of or brownish yellow color ples exhibited a dark 75.5% of prior cooking. However, and after consumers reported that one of the domestic rices (dT)Jasmine had a very ap- acceptable pearance after cooking. flavor attributes suggests that starchy at- flavor attributes suggests that starchy tributes (A were shown to be the most important param- important the most shown to be were of Jas- the acceptance eters in determining consumers, appearance by Asian mine rice (V between grain while stickiness Prediction of Jasmine rice flavor acceptance from flavor descriptors tively impact the aroma and flavor accep- tance among Asian consumers while higher sweet aromatic (A tegrity (V (M ) ” 5 to ” ) by ) for Vol. 67, Nr. 6, 2002 67, Nr. Vol. ” ” scale) — ” by 61%, ” (n = 49). A just about ” “ ) and gloss (V 1 just right 2430 “ much too little “ somewhat too light just about right “ “ to somewhat too little ” “ ). Stickiness and hard- ). Stickiness somewhat too brownish ” “ ) for Jasmine 85 (dJ), Spe- ) for Jasmine and 47%, 57%, and 40% as just about right ” ” just about right “ “ (that is, 30% to 63%). Three import- JOURNAL OF FOOD SCIENCE ” while 51%, 39% and 39% rated them while 51%, 39% and 39% rated ” much too light Consumers indicated that the most im- indicated that the most Consumers Jasmine rice sam- Most of the imported Degree of whiteness (VDegree “ somewhat too little just about right much too little “ of the Jasmine rice samples. of the Jasmine overall in determining their portant factor in order for Jasmine rice were, acceptance flavor, color, importance, of decreasing intensi- and hardness aroma, stickiness, and aroma, flavor, 6). Greater ties (Figure to a greater overall stickiness contributed color and hard- acceptance, while a yellow lower overall accep- ness contributed to tance scores. rice (dT) Jasmine Tec ples including Rice 85% of this consumer were rated by 67% to ( group as satisfactory color (n = 49). Only 22% to 44% of consumers color (n = 49). Only 22% the domestic samples rated the color of 2 of (dJ and dS) as 2430 Prediction of Jasmine rice appearance from visual descriptors Jasmine Rice Preference Mapping . . . . Mapping Preference Rice Jasmine scores hardness (5-point significant proportion of respondents rated these 2 samples as 61%, and 63% and 81%, 63%, and 75.5% in terms of their aroma and flavor intensities these three However, (n = 49), respectively. most acceptable Jasmine rices were rated for aroma and flavor as too low in intensity ( “ (dT), (dS), and Jasmati re- cialty Jasmine 39%, 43%, and 55% of Flavor, For spectively. consumers gave a rating of 30.5%, 26%, and 24% and 14.5%, 28.5%, and the 3 For respectively. 20% of consumers, (Jasmine rice samples, domestic Jasmine 85, dT) dS; and Jasmati, Jasmine, dJ; Specialty least preferred by consumers, their aroma and flavor intensities were not rated satis- Aroma, only 45%, 28.5%, and For factorily. 41% of consumers rated the intensity as “ in intensity ( too low ness seemed to play a small role in deter- mining overall acceptance of Jasmine rice by an increase However, this consumer group. in stickiness and a decrease in hardness contributed to a greater overall acceptance. More detailed information about the drivers of liking can be obtained by using descrip- tive attributes as predictors of the accep- and tex- aroma, flavor, tance of appearance, ture. yellow ed Jasmine rice samples, Golden Boy (iG) and replicates of Rose Rice (iR1and iR2), were rated as right, in intensity ( as too low to
Sensory and Nutritive Qualities of Food jfsv67n6p2420-2431ms20010552-MP.p65 Jasmine Rice Preference Mapping . . .
stickiness, and hardness, respectively. In- References preference mapping for Asian consumers living in the United States. J Sensory Stud 16:73-94. Adair CR, 1972. Production and utilization of rice in crease intensities of yellowness, floral, Narainakorn S. 1998. Jasmine rice overview at Jas- Rice Chemistry and Technology, AACC, St. Paul, mine Rice Net. Available from:
Vol. 67, Nr. 6, 2002—JOURNAL OF FOOD SCIENCE 2431 Sensory and Nutritive Qualities of Food
jfsv67n6p2420-2431ms20010552-MP.p65 2431 8/22/2002, 2:33 PM