Development of a "new" shelf-stable salsa with microwave vacuum (MIVAC) dehydrated cilantroкод для вставкиСкачать
INFORMATION TO USERS This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer. The quality of this reproduction is depeudent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely afreet reproduction. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book. Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6” x 9” black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order. UMI A Bell & Howell Information Company 300 North Zeeb Road, Ann Arbor MI 48106-1346 USA 313/761-4700 800/521-0600 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ABSTRACT DEVELOPMENT OF A “NEW” SHELF-STABLE SALSA WITH MICROWAVE VACUUM (MIVAC) DEHYDRATED CILANTRO A new shelf-stable salsa was developed using microwave vacuum (MTVAC) dehydrated cilantro. It was predicted that the MTVAC dehydration process would retain the fresh characteristics of the herb when added to a processed product and would be preferred by consumers. A standardized formulation without cilantro was developed as a control. An affective test was used for consumer acceptance. Analytical testing was used to reformulate the control and determine an “optimum” level of MTVAC dehydrated cilantro to add to the control. An affective test was conducted to determine if the reformulation with cilantro was preferred. The first affective test showed a 93% acceptance of the control recipe. The “optimum” amount of MTVAC dehydrated cilantro was determined to be at a 0.5% concentration level by a group recognition threshold. The second affective test showed a slight preference for the reformulation with cilantro, but no significant statistical preference was measured. Karla Marie Carlsen May 1997 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. DEVELOPMENT OF A “NEW” SHELF-STABLE SALSA WITH MICROWAVE VACUUM (MTVAC) DEHYDRATED CILANTRO by Karla Marie Carlsen A thesis submitted in partial fulfillment of the requirements for the degree of Master o f Science in Agriculture in the School of Agricultural Sciences and Technology California State University, Fresno May 1997 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UMI Number: 13 86269 UMI Microform 1386269 Copyright 1997, by UMI Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. UMI 300 North Zeeb Road Ann Arbor, MI 48103 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. APPROVED For the Department of Enology, Food Science, and Nutrition: ft & (-^ L;_____________________________ Dennis A. Ferris (Chair) Enology, Food Science, and Nutrition N. Joamfe Caid Enology, Food Science, and Nutrition J)n ,nAj-£j <A,Udby Sandra S. Witte Enology, Food Science, and Nutrition For the Graduate Committee: Dean, Division of Graduate Studies Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. AUTHORIZATION FOR REPRODUCTION OF MASTER'S THESIS K ^ I grant authorization for the reproduction of this thesis in part or in its entirety without further authorization from me, on the condition that the person or agency requesting reproduction absorbs the cost and provides proper acknowledgment of authorship. ________ Permission to reproduce this thesis in part or in its entirety must be obtained from me. Signature of thesis writer: ‘Ktf.iCi. ^ Cl. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ACKNOWLEDGMENTS I would like to thank my major professor, Dr. Dennis Ferris, for guidance, support, and patience throughout this project; and Dr. Sandra Witte and Mrs. Joanne Caid who served on my committee and encouraged me to further my education. I am grateful to the California Tomato Research Institute for funding this study as well as other opportunities for students at California State University, Fresno. And last, but not least, I would like to thank my husband, Carl, and my children, Peter, Kelly, and Bridget, for their steadfast love and support during my many hours of unavailability; and my parents for volunteering countless hours of child care and household maintenance. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. TABLE OF CONTENTS Page LIST OF T A B L E S ............................................................................................ vii LIST OF F I G U R E S ........................................................................................... viii IN TROD U CTION ........................................................................................... 1 REVIEW OF LITER A TU R E.......................................................................... 4 C ila n tr o ................................................................................................. 5 Microwave Vacuum ( M T V A C ) ......................................................... 6 Sensory E valuation................................................................................ 7 METHODOLOGY.................................................................................................. 10 Phase 1-Control Formulation and Consumer Acceptance Testing . 10 Phase 2-Reformulation of Control Salsa and Analytical Testing 14 . . Phase 3-Affective Consumer T estin g .......................................................... 22 RESULTS AND D IS C U S S IO N ........................................................................... 25 Results o f Phase 1 .......................................................................................25 Results o f Phase 2 .......................................................................................27 Results o f Phase 3 .......................................................................................34 C O N C L U S IO N S ..................................................................................................37 Phase 1-Control Formulation and Consumer Acceptance Testing . 37 Phase 2-Analytic-Discriminative T esting.................................................... 39 Phase 3-Affective Consumer Testing Consumer Preference T e s t i n g .................................................... 44 REFEREN CES........................................................................................................45 A PP E N D IC E S ........................................................................................................50 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. VI APPENDIX Page A. CONSENT WAIVER F O R M .................................................... 51 B. TEST SCORE SHEETS............................................................... 54 C. RAW DATA INCLUDING OPEN-ENDED COM M ENTS............................................... 58 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LIST OF TABLES Table Page 1. Triangle Test results for groups 1 and 2 .......................................... 28 2. Recognition threshold scores: Number correct out o f three sample s e t s ................................................................ 30 3. Converted data for individual recognition threshold: % correct above chance for each p a n e lis t ....................................32 4. Calculated individual and group t h r e s h o l d s .................................... 32 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LIST OF FIGURES Figure Page 1. Affective test for consumer a c c e p t a n c e .......................................... 26 2. Estimate of approximate threshold levels for individual judges . 3. Affective test for consumer preference................................................35 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 33 INTRODUCTION Of all new food products introduced on supermarket shelves in the last year, 3,698 were condiments (Domblaser, 1996). Salsas have shown the most growth in this category, replacing catsup, as the number one selling condiment in the United States today (Cousminer, 1996). According to a Neilsen Consumer Survey, the average household purchases 5.3 containers of salsa per year, compared with only 4.2 containers of catsup. This translates to salsa sales of over $600 million dollars per year (Mogelonsky, 1995; Charlet, 1994). One reason that accounts for the rapid growth and acceptance o f salsa is the changing demographics as the Hispanic population of the United States expands. An increase in the number and popularity of Mexican restaurants has exposed consumers to hot, spicy and bold flavors (Toops and Kevin, 1996). Another reason is the fashionable trend called “fusion cuisine” gaining notoriety among chefs from coast to coast. This fusion, or mixture of ingredients, flavors or cooking techniques of two or more cultures is the result of increased interest in ethnic and novel foods among consumers who want "something different." Still another reason is salsa’s low fat composition, which provides additional benefits to the consumer perception of “freshness” associated with vegetable products and disease prevention (Hollingsworth, 1996). While tomato is the basis for traditional salsa, variations with fruits and vegetables are now being served as an accompaniment to main dishes on a variety of restaurant menus. One traditional ingredient, the herb cilantro, is borrowed from both Mexican and Oriental cuisine, and is readily adaptable, offering new and exciting flavor to many dishes. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Because salsa is often offered as an accompaniment to tortilla chips, there is a connection to the salty snack market, which is also rapidly growing in popularity (Mogelonsky, 1995). Due to the growing demand for salsa and other products containing fresh herbs, it is advantageous to the food processing industry to produce a shelf-stable, processed product that contains cilantro, or other herbs, without losing the volatile compounds that give salsa its fresh flavor appeal. The purpose of this study was to develop a "new" consumer-acceptable, processed, shelf-stable salsa that contains microwave vacuum (MTVAC) dehydrated cilantro. Microwave vacuum dehydration allows for food products to achieve a very low moisture content, while retaining attributes, such as flavor, color and texture, which are compromised during conventional dehydration processes. Using MTVAC technology preserves cilantro and imparts the characteristic flavor and odor of cilantro to the processed salsa. The study was completed in three phases. Phase 1 was the product development stage. The objectives of this phase were: (1) to develop a standardized formulation for salsa without MTVACed cilantro to use as a control and (2) to use affective testing to verify consumer acceptance of the control formulation. Phase 2 was the determination of an “optimum” amount of cilantro to add to the control using analytical testing. The objectives in Phase 2 were: (1) to use a Triangle Test to screen subjects to serve on a trained panel and (2) to use this panel to verify a recognition threshold o f MTVACed cilantro added to the control formulation. Phase 3 was the determination of a preference between the control formulation without cilantro and the formulation with MTVACed cilantro. The objective of Phase 3 was to determine consumer preference through affective testing. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 The null hypothesis is that the target market, identified as customers at California State University, Fresno, Farm Market, will prefer a standardized salsa formulation made with MTVACed cilantro to the same formulation without MTVACed cilantro. The alternate hypothesis is that the same target market will not prefer a standardized salsa formulation made with MTVACed cilantro to a standardized salsa formulation without cilantro. The resulting standardized salsa formulation produced will be considered for eventual scale-up at the Center for Food Science and Nutrition Research Food Processing Laboratory. Surveys conducted at the California State University, Fresno, Farm Market over the past 3 years have indicated a 17% increase in interest in salsa, with a shelf-stable product preferred over a refrigerated product (Carlsen, 1995; Ferris, 1997). After scale-up, the finished product will be sold at the California State University, Fresno, Farm Market. University product development research often generates interest among food industry representatives. The use of MTVAC dehydrated cilantro in salsa is novel and has application in many other condiment or sauce formulations (Ferris, 1996). It is also hoped that this interest can lead to a collaborative working relationship between the food industry and the university as they realize the position that university facilities can serve in benchtop and scale-up operations. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. REVIEW OF LITERATURE Salsas are a combination of chopped vegetables and/or fruits to which any o f a variety of different herbs, such as cilantro, are added. The high degree of versatility of salsa includes use as a cooking sauce, a condiment or garnish for meats, poultry, eggs or fish, a dressing for salads and grains, and as an appetizer with chips or vegetables. In most formulations, tomatoes, cilantro, hot peppers and onions are key flavor components. The combined ingredients are not a puree, but are distinct pieces and are often uncooked (Cousminer, 1996). Tomato salsa can be made as a fresh product that needs to be refrigerated and used within a day or two. In order to prolong the shelf life beyond this time, the salsa must be processed to destroy any microorganisms responsible for spoilage. Because it is considered a high-acid food (Cousminer, 1996), tomato salsa can be produced using a high-pressure retort system at low (195-212°F) temperature or the "Hot-Fill-Hold" method (Ronsivalli and Vieira, 1992). The "Hot-Fill-Hold" method is preferred because it does not "over-process" the product, retaining firm texture and shape of the individual ingredients. Processing specifications using the Hot-Fill-Hold Method require time and temperature parameters to insure safety, while preserving quality of the finished product. According to current standards, minimum heat treatment required for high acid foods is 190° to 200°F in the coldest point (Rodrigo et al., 1990). Thermo-resistant bacteria, such as Bacillus coagulans, are found to be controlled in a tomato product at pH 4.5 and below when processed at 185° - 225°F for 2.56 minutes (Rodrigo et al., 1990). While tomatoes can withstand the temperatures reached under this type o f minimal processing without a loss o f flavor or texture, cilantro can undergo an undesirable flavor change (Giese, 1994). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Cilantro Traditionally, the herb cilantro (Coriandrum sativum L.) is used to impart a flavor which is characteristic in Mexican cuisine. Cilantro is the green, leafy portion of the coriander plant also known as Chinese parsley (Labensky and Hause, 1995) and is similar to the parsley family (Carum petroselinum) (Ockerman, 1978; Rosengarten, 1969). It is normally used in its fresh state as a seasoning or garnish. A study by Potter (1996) showed a significant difference in leaf oil during five distinct stages of growth, from the vegetative period to fruit set. An increase in (E)-2-Decenal was found to be present at the bud formation and blooming stages of the plant. (E)-2-Decenal is an irritant which may add an objectionable flavor. These oil compositional differences underscore the need to obtain cilantro which has been harvested at an early stage of growth, prior to fruit set, when levels of (E)-2-Decenal are low (Potter, 1996). The cilantro plant is delicate and the volatiles that produce the odor and flavor in this herb are easily vaporized (Giese, 1994). Under normal heat processing conditions o f canning, the color, flavor and aroma characteristics of fresh cilantro undergo changes or can be lost. Many processed salsas on the market today do not include cilantro in their ingredients because o f these undesirable qualities. Because the flavor and aroma o f cilantro is widely associated with fresh salsa, incorporation o f this herb was considered an important objective. Due to the problems associated with adding fresh cilantro during the processing stage, it was believed that adding the herb produced using MTVAC technology would achieve the desired flavor. In addition, the low moisture content o f the dehydrated product would retard growth of naturally occurring yeasts and molds (Blumenthal, 1990). Attributes desired in any dehydrated product include minimum change from fresh Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 product character, lower final moisture content, and reconstituted character similar to fresh in terms of flavor, color and texture (Clary, 1996). Common dehydration processes, such as air convection and freeze drying, result in a product that may retain some of these qualities, but is lacking in others. Air convection employs heated air which removes moisture from a product by evaporation or vaporization. However, the high heat may result in a loss o f color or browning effect. Flavor may be affected as well if temperatures reached cause volatile oils to evaporate or undergo heat-related reactions. Similarly, the freeze drying process, while utilizing low pressure and low temperature to remove moisture without volatile loss, may cause cell damage due to expansion and rupture of cells upon freezing. The result of this damage is loss of texture upon re-hydration (Clary, 1996). Cilantro can be dehydrated in the Dried Foods Technology Laboratory at California State University, Fresno, using MTVAC technology. MTVAC technology is described below. Microwave Vacuum (MTVAC) The patented MTVAC dehydration process allows fresh fruits, vegetables, and herbs to be dried to low moisture levels with minimal loss of their volatile components through the combined microwave vacuum dehydration process, (McKinney and Wear, 1987; McKinney et al., 1983). Within the MTVAC chamber, foods are exposed to radiant microwave energy at wavelengths approximately 300 MHz as generated by a magnetron tube (Clary, 1994). As the food absorbs the microwave energy, the polar water molecules attempt to rotate while sodium and chloride ions naturally present in the food product move in a linear fashion. This molecular movement within the food product produces heat, which drives the moisture from the inside of the food to the outside where it Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 vaporizes (Clary, 1994). This is opposite to other dehydration processes, which remove moisture from the product from the outside to the center through osmosis (Clary, 1996). Some of the factors that affect the rate of mass transfer of moisture include surface area, vacuum/pressure, time and temperature (Clary, 1996). Because the MTVAC chamber operates without pressure, the boiling point of water is lowered and moisture loss is achieved at a much lower temperature (Clary, 1996). The MTVAC process combines low atmospheric pressure and microwave energy resulting in relatively quick, low temperature dehydration of a product with little or no loss of volatile oils or heat damage to the plant cell walls (Clary, 1996). The MTVAC process has already been successfully used to produce dehydrated cilantro that retains its fresh appearance when reconstituted (Clary, 1996). Sensory Evaluation Sensory evaluation methods gather information through the use of reliable and valid testing of products in order to make decisions (Meilgaard et al., 1991). Four key factors are crucial: definition of the problem, test design, instrumentation, and interpretation of the test results (Meilgaard et al., 1991). A clear definition of the problem includes both project and test objectives, allowing the researcher to select the testing methods which will generate suggested solutions to the problem (Lawless and Claassen, 1993; Meilgaard et al. 1991). This is especially important in food product development, where multiple objectives, i.e., information regarding a food or ingredient characteristic and consumer perception, may be common goals (O’Mahony, 1986). Since the “instruments” many times are human subjects, the researcher is required to maximize efforts to control variability and bias (Meilgaard et al., 1991). Consumer testing has increased in use and, when used properly, has been found to Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 be very effective in product development, maintenance, and improvement (Meilgaard et al., 1991). Types o f Tests The diverse tests used in sensory evaluation can be divided into two major groups, affective and analytical-discriminative. Dependent on the primary goal, affective tests fall into two categories: acceptance or preference. Both tests seek to evaluate a general personal response (acceptance and/or preference) by a representative sample of potential consumers (Meilgaard et al., 1991). Acceptance testing determines the “affective status” of a product, i.e., how much a product is liked/accepted. Assessment by means o f a hedonic rating scale, preferably balanced with equal number of positive and negative choices in increments of equal size, results in more useful information (Meilgaard et al., 1991). Preference testing requires the subject to make a forced choice between two or more products, but may not indicate whether the product is liked. Training may be used for acceptance or preference tests; however, testing usually occurs on a large scale using untrained subjects (Sensory Evaluation Division o f the Institute of Food Technologists (SED, IFT, 1981). Analytical-discriminative tests are usually administered in a laboratory using trained, experienced testers. Similar to chemical or physical analysis, the goal is to obtain precise and reproducible results (Lawless and Claasen, 1993). Analytical-discriminative tests can further be broken down into difference and sensitivity. The objective is to identify and quantify food attributes based on difference or similarity (SED, IFT, 1981). Subjects must be experienced in methodology and possess the ability to identify, recognize, and recall sensory characteristics to assure repeatable results (Meilgaard et al., 1991). The validity of Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9 the results depends on individual subject sensitivity and reproducibility; the environment in which the test occurs; and control over physical and psychological influences (Amerine et al., 1965). Panel training is an important aspect in analytical testing. Orientation prior to the actual test should include familiarizing the subjects with the test procedure and scoring, and explaining product characteristics (Meilgaard et al., 1991). The researcher provides only enough information necessary to instruct the subjects without introducing bias (Meilgaard et al., 1991). The Triangle Test for difference is frequently used for panel selection based on the ability to make a distinction between three samples (Amerine et al., 1965). The subject must choose the odd sample from three in which two are identical and one is different. The Recognition Threshold Test for sensitivity is used to determine a concentration at which trained panelists identify and recognize a certain substance. This threshold is higher than the absolute threshold, which is the lowest possible concentration which is detectable (Meilgaard et al., 1991). Thresholds are among the most difficult data to obtain and define. Results are difficult to reproduce, even given the same panel on a given day (Meilgaard et al., 1991; Blumenthal, 1990). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. METHODOLOGY Phase 1-Control Formulation and Consumer Acceptance Testing The first objective of Phase 1 was to develop and standardize a formulation for a new salsa. The purpose of standardization was to establish a formula that would “yield a known quantity of a known quality at a known cost” (Knight and Kotschevar, 1989). Careful weighing or measurement of ingredients; attention to correct ingredient ratio; and documented handling, preparation and yield ensure reproducibility of a control formulation. The second objective in Phase 1 was to use affective consumer testing to determine acceptance of the control salsa formulation, using a single sample and a nine-point hedonic rating scale (Meilgaard et al., 1991). The purpose was to establish consumer acceptance of the control formulation for salsa prior to reformulation in Phase 2 to include the MIVACed cilantro. Formulation The basic formulation for the tomato-based salsa was as follows: Tomatoes, Toma-Tek, aseptic in foil bag, 3/4"cut 779.0 grams (77.5%) Onions, white, cut I/2"dice 155.8 grams (15.5%) 70.4 grams (7.0%) Jalapeno peppers, Cal-Compack, brined, l/4"dice Use of a preprocessed tomato assisted in the standardization of color, size of cut, and pH. Initially, tomatoes were provided by TomaTek, Inc. in 15-lb. aseptic lined, foil sample bags. The processed tomatoes were steam peeled and diced in a 3/4" x 3/4" cut. Calcium chloride (CaCl2) at 800 to 900 ppm was added for firmness. Citric acid was included to adjust the acidity level to a pH between 3.7 to 4.2. The tomatoes were processed aseptically and packaged in a lined foil Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 11 pouch. Tomatoes were removed from the foil pouch and weighed by difference on a top loading Ohaus electronic balance, model #GT4000. Use of a brined product assisted in the standardization of the capsaicin (heat) level in the jalapeno peppers. In this way, a “medium” salsa was produced with assurance of a standardized heat level that was consistent batch to batch. Jalapeno peppers were purchased through Whole Herb Company, Sonoma, California, and packed by Cal-Compack Foods, Inc., Santa Ana, California. The jalapeno peppers were cut into a 1/4" x 1/8" dice, standardized to a pungency level in the range of 100 - 200 ASTA Heat Units (ppm) and packed in brine containing 13-14% sodium chloride (NaCl) and 2.5-3.5% acetic acid, with a pH in the range of 2.8 - 3.7. The jalapeno peppers were drained in a wire strainer for 10 minutes, then weighed by difference as above. Since seasonal variation is not as great, it was felt that fresh onions would better match competing salsa brand ingredients. White onions were purchased from Save Mart supermarket in the Fresno, California area. They were peeled, diced into 1/2" pieces, then weighed by difference as above. Tomato solids and juice, onions, and jalapeno peppers were combined in a 3-quart saucepan and brought to "Hot-Fill-Hold" temperature (195-210°F), stirred occasionally, and held at this temperature for 10 minutes (Ronsivalli and Vieira, 1992). The pH of the mixture was taken and recorded prior to heating. The salsa was packed in 11-ounce clear, glass canning jars at the end of the holding time. Jars were washed and dried prior to filling. New fids were held in hot (180°F) water for a minimum of 10 minutes until jars were filled. The filled jars were inverted and held no longer than 5 minutes, then cooled in a refrigerator (35 40°F). After cooling, pH and salsa yield were measured and recorded. In order to limit microbiological activity and insure a safe product, a calibrated Oakton pH Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 12 meter, model #WD-35615-62, was used to test representative samples to assure the salsa was below pH 4.2. Consumer Acceptance Testing A Simple Acceptance Test of the control salsa without cilantro was conducted using a nine-point Hedonic Rating Scale. The objective was to determine overall acceptability of the control formulation. Scale ratings were converted to numerical scores: 9, like extremely; 8, like very much; 7, like moderately; 6, like slightly; 5, neither like nor dislike; 4, dislike slightly; 3, dislike moderately; 2, dislike very much; 1, dislike extremely (Amerine et al., 1965). The mean and median of the total responses were required to be equal to or greater than 6 in order to accept the control formulation (Amerine et al., 1965; O’Mahony, 1986). Participant Selection Thirty untrained volunteers were recruited among shoppers at the California State University, Fresno Farm Market, to best estimate the preference of the target market (Meilgaard et al., 1991). Participants were selected on the basis of an affirmative answer to the question, “Do you eat salsa?” (ASTM, 1995c; Meilgaard et al., 1991; SED, EFT, 1981). A Human Subject Review Committee determined the participants associated with this experiment to be “at minimal risk.” The consent waiver forms (see Appendix A),which include subjects signed consents to participate in the study, will be stored, on file, in the office of Dr. Dennis Ferris, Department of Enology, Food Science and Nutrition, California State University, Fresno, for 5 years. Testing occurred over a 1-week period. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 13 Sample Preparation Salsa was prepared according to the standardized control formulation on page 10 in the test kitchen, Sensory Laboratory, Family and Food Science Building at California State University, Fresno. One tablespoon of control salsa was placed in 2-ounce white paper portion control cups at 35-40° F. A single sample was presented to each panelist on a tray with a plastic spoon, score sheet and pencil. A cup of tepid water was offered to cleanse the palate before tasting. Testing Procedure The testing was held in the California State University, Fresno, Farm Market. A storage cabinet was used to hold samples, and provide a place for participants to write on score sheets. Opened jars o f salsa were kept chilled in a cooler during the testing process to maintain temperature control. A single sample appropriate for determining acceptability of a new product was presented (SED, IFT, 1981). Participants were asked to taste the salsa and mark their response on the score sheet using a nine-point Hedonic Rating Scale to describe their opinion (Appendix B). Comments regarding what in particular was liked or disliked about the product were solicited at the bottom of the score sheet. Statistical Analysis The Consumer Acceptance Test was analyzed using a one-tailed, one sample T-test. The results of the T-test were compared to Table T4, page 329 of Sensory Evaluation Techniques (Meilgaard et al., 1991). Results of the Consumer Acceptance Test were also presented graphically (SED,IFT, 1981; Meilgaard et al., 1991). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 14 Phase 2 - Reformulation of Control Salsa and Analytical Testing When the control formulation for salsa was found to be acceptable by affective testing of the consumer group, the control formulation was modified to include MTVAC dehydrated cilantro. One purpose of Phase 2 was to provide a basis for determining the “optimum” amount of MTVACed cilantro to be added to the control formulation. A Triangle Test was used as a screening to select a panel of judges to measure the Recognition Threshold of cilantro added to the control salsa. This panel of judges were trained to participate in an analytical discriminative Recognition Threshold test. Reformulation of Control Salsa The reformulated salsa was as follows: Tomatoes, San Benito brand, canned T'dice 779.0 grams (77.1%) Jalapeno peppers, Cal-Compack, brined, sliced 155.8 grams (15.4%) 70.4 grams (7.0%) 5.0 grams (0.5%) Onions, white, cut l/2"dice MTVACed cilantro, stem and leaves, chopped Toma-Tek, Inc. was no longer able to supply 3/4" size, aseptically processed tomatoes in the amounts necessary for the remainder of the study and for future scale-up. The new ingredient supplier, S.E. Rykoff, provided San Benito brand canned tomatoes; however, they were only available in a T' size cut. The canning process and increase in size of cut were two unavoidable changes. Other ingredient specifications, color and pH, were maintained as well as calcium chloride and citric acid levels. Tomatoes, white onions and jalapeno peppers were weighed using the same procedure as stated in Phase 1 on page 11; ingredients were combined in a 3-quart saucepan; and brought to “Hot-Fill- Hold” temperature for 7 minutes before the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 15 MTVAC cilantro was added. A pH reading was recorded prior to heating, using the same instrument and method as previously stated. The cilantro for use in this trial was preprocessed in the Viticulture and Enology Research Center MTVAC Food Processing Laboratory located on the California State University, Fresno, campus in March, 1996. Cilantro stems and leaves were washed, dried, soaked in citric acid at pH 3 for 1 hour to retard enzymatic activity, then chopped to process the stems and leaves separately. The cilantro was processed in the MTVAC at 130°F for 70 minutes and 3,000 watts. After dehydration, the cilantro was vacuum packed in lined foil pouches to limit exposure to light, air, and moisture. Desiccant pouches were added to stabilize moisture levels. The pouches were opened to remove the cilantro and then closed by folding over the top and taping shut using masking tape. The cilantro leaves were crumbled and the stems were chopped by hand to approximately 1/4" size. The MTVACed cilantro was weighed by difference using the same balance as before and added to the control salsa formulation in a 0.5% concentration by weight. This concentration was selected based on former experiments (Duran-Guiterrez, 1995). The MTVACed cilantro was added to the control salsa formulation and the temperature was maintained for 3 more minutes. The researcher reasoned that the added cilantro would come up to “Hot-Fill-Hold” temperature in the salsa without prolonged exposure to heat. Condition 1 - Panel Selection for Discriminative Tests A Triangle Test was used to determine if a participant could detect a difference between control samples and those containing MTVACed cilantro. This was a screening process for serving on the threshold panel (Meilgaard et al., Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 16 1991). This screening of subjects is useful during recruitment because a subject is eliminated early on if they are unable to discriminate between a sample with the desired ingredient and a sample that does not contain this ingredient (Meilgaard et al., 1991). Fatigue during testing due to the chemical irritant properties of the capsaicin present in jalapeno peppers was an important consideration. In order to minimize the trigeminal effects of this stimulus, the smallest required number of repetitions for each testing situation was adopted (Amerine et al., 1965). This panel was assembled and trained to determine a recognition threshold for the “optimum” addition of cilantro to the control salsa formulation. Panel Selection Panel selection for the Triangle Test was based on the ability to recognize and identify the flavor of the herb cilantro when added to the control salsa. Participants were recruited from faculty and experienced students of the California State University, Fresno, Department of Enology, Food Science, and Nutrition; Food Services department; Viticulture and Enology Research Center Dried Foods Technology Laboratory; and other professionals from the food industry. Criteria included commitment to the testing schedule, general good health, and ability to discriminate flavor and aroma accurately. Fourteen individuals participated in the Triangle Testing, 6 men and 8 women ranging in age from 20 to 60. The participants were provided with the same consent waiver form described in Phase 1 (Appendix A). Panelist Training Panelists were trained in two separate groups. Both groups were familiarized with the test method and scoring procedure (Rutledge, 1992). Training for the first group of 9 panelists was held in Room 105 of the Family and Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 17 Food Science Building in a group setting. Seating for this training session was provided at tables and chairs arranged in a U-shape (Meilgaard et al., 1991; Rutledge, 1992). Training for the second group of 5 panelists was held in Room 118 of the Family and Food Science Building in a group setting. Seating for this session was round table style, commonly used in focus or smaller group settings (Meilgaard et al., 1991; Rutledge, 1990). Both groups of panelists viewed sample score sheets (see Appendix B) and were given scoring instructions. Panelists were instructed that samples would be presented in sets of three; to taste all samples from left to right; and indicate on the score sheet which sample was different. Panelists were allowed to return to the samples in a given set to re-taste if necessary. Instructions for the Forced-Choice Method were described during training: if a panelist could not tell the difference between samples, they must guess (Meilgaard et al., 1991; SED, EFT, 1981). Sample Preparation Samples of the reformulated control salsa were prepared in the same test kitchen using the “new” formulation on page 14. One tablespoon of control salsa without cilantro and 1 tablespoon reformulated salsa were placed in individual 2ounce clear plastic portion control cups. The cups were coded using random numbers from a random number chart (Meilgaard et al., 1991). The salsa was served at 35 - 40°F. Testing Procedure Sensory testing for the first group o f 9 panelists was performed in the individual testing booths in the Sensory Laboratory in the Food and Family Science Building, which accommodates up to 5 panelists at one time. Each booth is equipped with a water fountain and a pass-through door for presentation of Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 18 samples. The booths are arranged side by side and are divided by 4-foot panels to insure privacy (Meilgaard et al., 1991). The panelists entered and exited the tasting booth area o f the laboratory from the outside hallway. Red lighting was used to attempt to mask appearance differences between samples (Meilgaard et al., 1991). Testing for the second group of 5 subjects occurred in Room 118 of the Family and Food Science Building. Tepid water in cups was provided and samples were held covered until presentation. Available fluorescent lighting was used. Panelists were instructed to rinse their mouth with water prior to tasting the samples and rinse between samples. Unsalted crackers were provided for panelists to eat between sample sets (Gillette et al., 1984). Samples were arranged on trays in sets of three. In each set, 2 samples were identical and 1 was different. Six possible combinations were prepared. Samples were presented in balanced order so that the chance of panelists receiving an identical sample order was minimized (ABB, BAA, AAB, BBA, ABA, BAB) (Meilgaard et al., 1991). Each of the 14 participants received 3 sets o f samples for a total of 42 sample sets. Panelists were instructed to taste 3 samples in one set, then circle the number on the score sheet corresponding to the sample that was different. Testing occurred on 2 different days. Statistical Analysis The Triangle test was analyzed using Table T7 Triangle Test for Difference: Critical Number (Minimum) o f Correct Answers (Meilgaard et al., 1991). Each panelist group was analyzed separately. The results were presented separately in one table. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 19 Condition 2-Recognition Threshold Test A panelist’s ability to recognize the MTVACed cilantro sample correctly in three sets of three samples served in the Triangle Test was the criterion for serving on a Recognition Threshold panel. Panelists who met this criterion were asked to return to serve on the threshold panel. This panel was trained to determine a Recognition Threshold level of cilantro in the control salsa. This threshold was used to determine a guideline for adding an “optimum” level of cilantro to the standardized control formulation. The researcher reasoned that a concentration higher than that detected by a trained panel would be necessary for an untrained group to recognize. One concentration step above the established group threshold was selected to be appropriate and yet cost effective. Recognition Threshold Test A Recognition Threshold Test was used to determine the concentration at which the MTVACed cilantro could be recognized by a trained panel. The purpose of this test was to establish a guideline for adding an “optimum” level of cilantro to the control salsa formulation. According to ASTM, 1995, method E1432, the recognition threshold was defined as “the lowest concentration of a substance in a medium relating to the lowest physical intensity at which a stimulus is recognized by an individual panelist 50% of the time” (Meilgaard et al., 1991). The Recognition Threshold o f each individual judge was determined by the BestEstimate Threshold criterion (BET) using a trained panel (ASTM, 1995b). The group threshold concentration was determined based on the geometric mean of the individual threshold levels (ASTM, 1995b; Meilgaard et al., 1991). It was assumed that a one step higher concentration of cilantro will be recognized by 50% of the corresponding population (Meilgaard et al., 1991). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 20 Panelist Selection Five panelists were selected, based on their ability to perceive a difference in three out of three samples in the Triangle Test and availability. The panelists included 3 men and 2 women, approximately 20 to 45 years of age, experienced in sensory testing. The panel was retrained for familiarity with the new test procedures and with the cilantro stimulus using the re-hydrated cilantro to calibrate panelists. Panel Training Training was provided at the California State University, Fresno, Sensory Laboratory in the Family and Food Science Building to familiarize the panelists with the format o f the test, the task involved and the evaluation procedure (Meilgaard et al., 1991). The panelists were made familiar with the sample of control salsa; the sample of reformulated salsa; as well as a sample cup containing re-hydrated MTVACed cilantro. Panelists were instructed to rinse their mouth with water before tasting the first sample and between sets. Unsalted crackers were provided to eat between sample sets (Gillette et al., 1984). Samples were tasted from left to right. As in the Triangle Test, the Forced Choice Method was used: if a panelist could not tell which was different, they were instructed to guess. Panelists were allowed to return to each sample within a sample set in order to choose the sample in which the odor or flavor of cilantro was recognized (ASTM, 1995a). Sample Preparation Samples were prepared in the California State University, Fresno test kitchen described above by adding MTVACed cilantro to the control salsa formulation on page 10 in a series of concentration steps. The lowest concentration increments of Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 21 the substance began well below the level that the most sensitive panelist was expected to recognize (ASTM, 1995a). The ingredients, tomato, white onion, and jalapeno pepper, were weighed by difference using the same scale as before and placed into five individual 3-quart saucepans. Using the same instrument and method as before, a pH reading was taken for each of the five saucepans of salsa. Each of the saucepans were brought to “Hot-Fill-Hold” temperature and held for seven minutes (Ronsivalli and Vieira, 1992). MTVACed cilantro was weighed by difference using the same balance at 0.05%, 0.1%, 0.3%, 0.5%, and 0.7% (by weight) concentrations and was added to each of the saucepans. After the addition of the MTVACed cilantro, temperature was maintained for 3 more minutes. One tablespoon o f salsa at each concentration was placed in a 2-ounce clear plastic portion control cup and served at 35-40° F. All samples were coded using random numbers. Testing Procedures The test procedure used for the Recognition Threshold test was the 3Altemate Forced Choice (3-AFC) Method. If the panelist could not make a choice, they were instructed to guess (ASTM, 1995a,b). Each designated set of three samples contained one sample with cilantro and two samples without cilantro (blanks). Beginning with the lowest concentration of cilantro, the panelists were asked to identify the samples which contained the taste and odor of cilantro and record the results on the score sheet provided (Appendix B). Nine samples were presented on a tray at each concentration step, arranged in three rows of three, with the sample containing cilantro in randomized leftcenter-right position to eliminate bias. Testing continued until all sets were Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 22 evaluated, or until a panelist correctly identified all three sets at two consecutive concentration steps (ASTM, 1995a). Sensory testing was performed in the individual testing booths in the California State University, Fresno Sensory Laboratory. Green lighting was used in the booths to mask visual differences between samples (Meilgaard et al., 1991). Statistical Analysis Results o f the Recognition Threshold test were analyzed using the Best Estimate Threshold calculation in Appendix X -l of the ASTM test for Determination of Odor and Taste Thresholds By a Forced-Choice Ascending Concentration Series Method of Limits (ASTM, 1995a). Individual thresholds were determined by converting data into percent correct above chance values and presented graphically. The group threshold was accepted to be the geometric mean of the individual thresholds (ASTM, 1995a,b). Phase 3-Affective Consumer Testing After the optimum amount of MTVACed cilantro was added to the control formulation, a second affective test was administered to determine if a consumer preference existed between the two salsa formulations. A Consumer Preference Test using Paired Preference was performed to determine if there was a preference for the reformulated salsa with cilantro over the control formulation. A One-Sided Comparison test was used to answer the original null hypothesis. To best approximate the preference of the target market, the test was conducted on 6 randomly selected days over a 2 week period. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 23 Participant Selection Participants were recruited from customers at the California State University, Fresno, Farm Market. One hundred and two customers were asked to taste the two salsa formulation samples. Panelists were asked to sign a consent waiver form (Appendix A) prior to tasting. Training Training was minimal. Participants were asked to taste both samples and record the sample number they preferred. Participants were allowed to choose which sample to try first, reducing bias. Although rinsing between samples was not practical in this setting, a cup of tepid water and unsalted crackers were offered to each participant with the samples and score sheet. Sample Preparation Samples were prepared using the control and reformulation recipes as described on pages 10 and 14 in Sensory Laboratory test kitchen at California State University, Fresno. One tablespoon of each salsa, control and reformulated, was placed in two clear, 2-ounce clear plastic portion control cups and served at 35 - 40° F. Samples were tasted with a plastic spoon. Jars of opened salsa used in the test were kept in a cooler throughout the testing period to control serving temperature. Random numbers were used to identify each sample. Testing Procedure The same storage cabinet in the Farm Market as described in the Consumer Acceptance Test was used to hold samples and provide a space for writing on the score sheets (see Appendix B). Panelists were presented with two coded samples at the same time, the control salsa and reformulated salsa with cilantro. The Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 24 samples were placed side by side, in random order. The chance a panelist would taste one formulation before the other was random. Statistical Analysis The Consumer Preference Test was analyzed using Table T8 for Comparison Test Difference: Critical Number (Minimum) of Correct Answers with the results presented graphically (SED, EFT, 1981; Meilgaard et al., 1991). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. RESULTS AND DISCUSSION Results of Phase 1 Formulation A standardized control salsa formulation was produced. This formulation was found to be acceptable by means of a Consumer Acceptance Test. No further modification to the control formulation was necessary. Consumer Acceptance Test Nominal scores for the Consumer Acceptance Test were converted to numeric values in order to determine a mean value of 7.1, a median value of 7 and a standard deviation of 1.16. The mean score o f 7.1 was compared to the criteria for acceptance score of 6 using a T-test resulting in a value of 5.19, which is greater than the tabular value of 1.697 required for significance at 0.05% confidence level. The median measure of central tendency is often used when the data are skewed by extreme values (O’Mahony, 1986). The results of the affective test for consumer acceptance are illustrated in Figure I. Although positively skewed, a normal distribution was suggested by the data. The largest percent of scores fall in the 7, “like moderately” (46.7%) and 8, “like very much” (40%) range. Criteria for acceptance of the control formulation was a mean score of 6, “like slightly”, or above. The mean value was 7.1 and median value was 7 , which are both above the score criteria for acceptance, 6 “like slightly”. The variance was 1.334. Based on the 7.1 mean value rating among the consumers tested, research proceeded without any changes to the formulation. It should be noted that this test was based Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 26 Consumer Acceptance Test Results Frequency 16 ■■■- dislike extremely dislike very m uch 1 2 dislike moderately 3 ■ dislike slightly neither like nor dislike 4 5 like slightly 6 . like m oderately 7 8 like very much like extrem ely 9 Numeric Scale of Scores Figure 1-Affective test for consumer acceptance Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 27 on a small sample o f customers (n = 30), which was referred to as a rough screening, common in benchtop studies (Fuller, 1994; SED,IFT, 1981). Raw scores including opened-ended comments are in Appendix C. Results of Phase 2 Reformulation o f Control Salsa The reformulation used the acceptable control formula producing a “new” formulation with a 0.5% concentration (by weight) o f MTVACed cilantro added. This amount was determined through discriminative recognition threshold testing of concentration levels above and below the 0.5% concentration. Triangle Test The scores for both groups of individual judges participating in the Triangle Test used for panel selection appear separately in Table 1. Scores for individual judges as well as overall scores were analyzed by Table T7 Triangle Test for Difference: Critical Number (Minimum) of Correct Answers (Meilgaard et al., 1991). Due to the difference in test conditions, the two panels must be considered independently, with raw scores reported in Appendix C (Meilgaard et al., 1991). O f the first group, 56%, or 5 out of 9 subjects, were able to correctly identify three out of three samples containing cilantro at the 5% significance level. In the second group, 40%, or 2 out of 5 subjects, were able to correctly identify three out of three samples containing cilantro at the 5% significance level. The second group, tested under fluorescent fighting, should have been able to visually detect the sample with cilantro, yet only 2 panelists out o f 5 scored correctly three out of three times. This suggests that perhaps there was an unaccounted factor influencing the panelists’ response; more training is needed; Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 28 Table 1 - Triangle Test results for groups 1 and 2 Triangle Test - Group 1 Panelist 1 2 3 4 5 6 7 8 9 # Correct/# Possible 3 out of 3 3 out of 3 3 out of 3 0 out of 3 2 out of 3 1 out of 3 1 out of 3 3 out of 3 3 out of 3 Will Serve on Threshold Panel Y Y Y 19 27 Total Correct Responses Total Possible Triangle Test - Group 2 Panelist 10 11 12 13 14 # Correct/# Possible 2 out of 3 3 out of 3 1 out of 3 2 out of 3 3 out of 3 Will Serve on Threshold Panel Y Y Total Correct Responses 11 Total Possible 15 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 29 or that cilantro at a 0.5% concentration in salsa is not perceptible to this group of subjects, even with visual clues. One possible reason that the first group scored higher than the second group involves the use o f colored lighting in the sensory booths. It was assumed that using red lighting would mask differences between the samples, but in actuality, the reverse was true. During testing, two judges reported that the red lights caused the samples containing cilantro to appear to have black particles. According to Meilgaard, theater gels or low-pressure sodium lamps which emit light at one wavelength are alternatives which remove colors, but do not eliminate differences in color intensity (Meilgaard et al., 1991). What is also interesting is that, while the red lighting effect provided a visual difference between samples with and without cilantro, there was still only 50% recognition by the judges. This could be explained by the psychological effects of stimulus error, in which irrelevant data causes error (Meilgaard et al., 1991). The panelists who reported the visual difference wondered if the test was purposely designed to “trick” them. To control this factor, blue and green lighting were tested (Meilgaard et al., 1991). Green lighting was found to be more effective in masking differences between the salsa samples with and without cilantro. From the two groups participating in the Triangle Test, seven panelists scores met the criteria for the threshold test, however only five were able to return to serve on the Recognition Threshold panel. Judges were selected based on their ability to correctly identify samples containing cilantro in the Triangle Test. Recognition Threshold Testing The “best-estimate” value of the individual panelists was calculated from each individual’s correct/incorrect responses; the group threshold is the geometric Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 30 mean of all the individual responses (ASTM, 1991a,b). Scores for the individual judges at each concentration step are shown in Table 2, representing correct choices in three sets. For example, if the judge scored correctly all three times, the overall score for that concentration step would be 3, with the probability of 1/3 that the correct choice was by chance alone. These data are then converted according to ASTM methods to represent the percent correct above chance, using the formula: % correct above chance = 100(3C-N)/2N, where N = the number of tests presented per panelist and concentration, and C = the number o f correct choices. Table 2-Recognition threshold scores: Number correct out of three sample sets Number Correct Panelist Number Concentration Steps 1 2 3 4 5 0.05% 0 1 2 1 0 0.10% 0 1 1 0 2 0.20% 0 1 1 3 2 0.30% 2 0 2 0 3 0.40% 2 1 1 3 3 0.50% 3 3 2 3 0.60% 3 3 2 0.70% 3 3 I Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 31 The converted scores are shown in Table 3, as indicated in ASTM Method E 1432-91, using the formula appearing below the table. After conversion, the scores are transformed so that 100% is plotted as 99.5%, 0% as 0.5%, and less than 0 as 0.1%. A best fit line is plotted through the points and the threshold is read as the concentration corresponding to 50% probability (ASTM, 1995b). Individual thresholds are the mean of the concentration step values at which the judge correctly identified the sample containing MTVACed cilantro at 50% above chance. Calculated threshold levels for each judge using the regression formula appearing on each graph is shown in Table 4. The estimate o f approximate threshold levels by individual judges is represented graphically in Figure 2. Using these calculated individual threshold values, the group mean is calculated to be the geometric mean o f the individual threshold means. Because this test required so many repetitions, fatigue and chemical irritation due to capsaicin levels in the samples were a concern. Capsaicin, a pungent amide o f vanillylamide present in jalapeno peppers, stimulates pain receptors in the mouth and throat, and causes an intense burning sensation (Amerine etal., 1965; Krajewska and Powers, 1988). Free nerve endings, usually mucosal membranes, are known as the trigeminal receptors (Amerine et al., 1965). Other physical responses that are common include tearing in the eyes and nasal secretion (Krajewska and Powers, 1988). Capsaicin does not impart a detectable flavor or odor to a food product; however, individuals have a wide range o f sensitivity to the substance (Krajewska and Powers, 1988). To control this problem, panelists were required to drink water and eat unsalted crackers between sample sets to cleanse the palate and prevent capsaicin build-up inside the mouth (Krajewska and Powers, 1988). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 32 Table 3-Converted data for individual recognition threshold: % correct above chance for each panelist at each concentration Panelist Number 4 5 Cone. Steps 50.00 0.50 0.10 0.05% 0.50 0.50 0.10 50.00 0.10% 0.10 0.50 0.50 99.50 50.00 0.20% 2.52 50.00 0.10 50.00 0.10 99.50 0.30% 2.40 50.00 0.50 0.50 99.50 99.50 0.40% 2.30 99.50 99.50 50.00 99.50 2.22 99.50 99.50 50.00 0.60% 2.15 99.50 99.50 0.50 0.70% 1 2 3 3.30 0.10 0.50 3.00 0.10 2.70 Log of Cone. Levels 0.50% Data converted per formula: % Correct above Chance = 100(3C-N)/2N N = number of tests presented per panelist and concentration (in this study, 3) C = number of correct choices (ASTM, 1995b) Table 4-Calculated individual and group thresholds Panelist Number 1 0.36 2 0.42 3 * 4 0.26 5 0.17 Group Threshold 0.29 * no threshold established for panelist #3 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 33 Individual Threshold Judge 1 Individual Threshold Judge 2 150.00 150.00 <D O c to g c .e 0 100.00 © 1< 8 oo * i 50.00 mm ■1 50.00 i 0.00 3.30 3.00 05% 100.00 > t 0.00 7 7 0 7 5 2 7 4 0 7 3 0 7 2 2 715 20% 30% 40% .50% .60% 70% 3.30 3.00 7 7 0 7 5 2 7 4 0 730 7 2 2 7 1 5 05% 10% .20% 30% 40% 50% 60% 70% Concentration Steps 10^fconcenlitration Steps y= 17 683X-42 y = 17 75x- 30.025 Individual Threshold Judge 3 0O) Individual Threshold Judge 4 150.00 150.00 100.00 O 100.00 c <0 5 aj 50.00 0.00 mm 50.00 3 30 3.00 7 7 0 7 5 2 7 4 0 7 3 0 7 2 2 7 1 5 05% 10% 20% 30% 40% 50% .60% 70% Concentration Steps 3.30 3.00 7 7 0 7 5 2 7 40 7 3 0 7 2 2 7 1 5 05% .10% .20% 30% .40% 50% .60% 70% Concentration Steps y = 19.823X -19.513 Individual Threshold Judge 5 150.00 0<D 100.00 1< lO 50.00 O 0.00 3.30 3.00 7 7 0 7 5 2 7 4 0 730 7 2 2 7 1 5 05% 10% 20% 30% 40% 50% 60% 70% y = 24.83x - 14.67 Concentration Steps Figure 2-Estimate of approximate threshold levels by individual judges Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 34 The error of expectation is a common problem in threshold testing. This error occurs when panelist anticipate the stimulus being tested and report it before it is actually perceived (Meilgaard et al., 1991). The error o f central tendency, where the odd sample is detected more often in the middle position of a set of three, and the pattern effect, where panelists try to detect a pattern in presentation, were also considered in the test design (Meilgaard et al., 1991). To control these factors, the method of presentation included random sample code numbers and random order of presentation and the Three-Alternate Forced Choice (3-AFC) method for scoring was adopted (Amerine et al., 1965, Meilgaard et al., 1991). While it would be expected that an increase in recognition as the concentration increased, judge 3 was inconsistent in recognition response to cilantro. Four judges scored were able to detect cilantro in the samples (50% correct recognition above chance at or below). Judge 3 was unable to score a threshold value. These data support a group recognition threshold of 0.29 (% concentration MTVACed cilantro), as shown in Table 4, at the 0.5% concentration level, based on the mean score for percent correct above chance (ASTM, 1995b). It can be inferred from these results that adding cilantro to the control formulation at 0.5% concentration will result in a product in which the flavor and odor of cilantro will be recognized by a large percentage of the population. Results of Phase 3 Consumer Preference Test The results of the affective test for consumer preference appear in a frequency chart in Figure 3. According to Table T8 Two-Sided Paired Comparison Test for Difference: Critical Number (Minimum) of Correct Answers, for a sample size n - 102, 61 participants would have had to have chosen the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 35 Consumer Preference Test Results 60 • Control Cilantro Control vs Cilantro Formulations Figure 3 - Affective test for consumer preference Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 36 reformulated salsa with cilantro in order to show a significant preference at the 5% confidence level. A Two-Sided Test is used to measure the responses that agree with the direction of interest, namely, a preference for the formulation containing cilantro (Meilgaard et al., 1991). This test, also conducted at the California State University, Fresno, Farm Market, compared the control salsa to the reformulated salsa with MTVACed cilantro. Of the 102 persons participating, 49% preferred the control, while 52% preferred the formulation with cilantro. To account for these results, at least two explanations can be offered. During the tasting session, there were occasions where more than one person was scoring the samples. This may introduce bias in the form of mutual suggestion, where one participant is influenced by the other participant, through words, facial expressions or gestures (Meilgaard et al., 1991). This can be controlled when using a panel format, but is more difficult to control in a consumer setting. Another factor that can influence response is the error of expectation, where the participant is given information about the sample which may affect a response based on preconception (Meilgaard et al., 1991). Based on some of the commentary responses, some participants may have anticipated that one sample was “hotter” than the other, when indeed the only difference between sample was the addition of cilantro (see Appendix C). Subjects were only given instructions to taste the two samples and indicate a preference. Because so many comments mentioned a heat (pungency) difference between the two samples, the subjects may have believed that this characteristic was the one under study. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CONCLUSIONS Phase 1-Control Formulation and Consumer Acceptance Testing Formulation The high ratings resulting from the Consumer Acceptance rough screening were well within limits of overall acceptance of the basic control formulation. Beside overall acceptance of the formulation, an assumption was made that the heat level was within an acceptable range. There are no recommended changes at this time. Recommendations for Further Study One factor that may be a problem in the scale-up of this formulation is the 1" size of the tomato cut. While this size yields the desirable chunky appearance and retains shape and texture during the salsa cooking time, it may be too large for the pump to accommodate at the California State University, Fresno Food Processing Laboratory. Finding an ingredient supplier that produces a product with the same specifications other than size will be important. The tomatoes used in the Consumer Acceptance Test were supplied by Toma-Tek, Inc., Firebaugh, California. Toma-Tek, Inc. was unable to supply the same aseptically processed tomatoes in the quantities needed for the other sensory tests required for this study, as well as the scale up process. To complete the sensory testing necessary, canned tomatoes packed by San Benito Foods and supplied by S.E. Rykoff were used. The decision to change tomato suppliers at this point was based on availability, cost, and quality. The variables introduced by this change included the use o f can processing instead of aseptic packaging and an Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 38 increase in cut from 3/4" to 1" in size. Visually, the tomatoes appeared chunkier; however, the color and flavor were comparable. All other specifications, including citric acid, sodium chloride, pH and yield, were maintained. Using a pre-processed tomato product enables a standardized color and Brix level of the tomato. If the formulation were changed to utilize fresh tomatoes, these parameters would have to be standardized and controlled. The white onions were purchased at a local supermarket and hand cut into 1/2" pieces. For scale-up processing, a standardized, mechanical process is necessary. A cost analysis study comparing the financial and time-saving benefits of a vertical or Buffalo chopper for manual processing, or a pre-cut onion product should be conducted for larger scale production. The jalapeno-pepper ingredient was selected to assure a standardized heat level. Because this product is packaged in a brine, there may be a salty or metallic aftertaste present. Future study should include investigation as to the objectionable qualities that this aftertaste may present over time. Consumer Acceptance Test Phase 1 cumulative scores o f the Consumer Acceptance Test show that 93% of the respondents rated the salsa with a score of 6, “like slightly,” or above, which satisfied the criteria for acceptance of the salsa formulation. Statistical analysis using a t-test showed that the sample mean value of 7.1 was significantly different than the value of 6 required for acceptance. From this information, the researcher inferred that the target market, defined as customers at the California State University, Fresno, Farm Market liked the control salsa. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 39 Recommendation for Future Studies The subjective nature of sensory testing makes reproducibility and quantification difficult (Amerine et al., 1965; O’Mahony, 1986; Meilgaard et al., 1991). Due to time or budget limitations, often too few subjects or trials are used (Roessler, 1984). For the Consumer Acceptance Test, 30 subjects represents the lower limits required for such a test. However, due to the favorable results of the subsequent Consumer Preference tests, the researcher feels that the target population has been adequately represented. A larger sample may be required in future studies, perhaps to break out segments of the market that prefer a hotter or milder formulation. For this study, the high level of acceptance justified continuation to the next phase. Phase 2-Analvtic-Discriminative Testing Re-formulation of Control Salsa The basic salsa formulation produced for the discriminative testing portion of the study included the addition of MTVAC dehydrated cilantro. The cilantro used was produced approximately 1 year prior and packaged in vacuum foil bags. Upon opening, the dry cilantro began equilibrating immediately, and upon rehydration in (room temperature) water had surprisingly bright green color, regained flavor and aroma. The dry leaf portion was crushed by hand and the stem portion cut. The handling of the dry cilantro stems was very awkward and recommendations for future treatment should include chopping prior to drying. Furthermore, to maintain the optimum dry state prior to addition to the salsa formulation, cilantro should be packaged in batch weight so that an entire packaged portion would be used at one time. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 40 Recommendation for Future Studies Further comparison tests should be conducted to determine if there is a detectable difference between the MTVACed cilantro, conventional air dried and freeze dried when used in the salsa formulation. Parameters should include flavor, aroma, color, texture, availability, cost and microbiological testing. Additional sensory testing is necessary after scale-up to insure that the standardized formulation maintains acceptable parameters. Shelf life studies are a prime consideration to maintain safety and acceptability of product characteristics. It is not known how the chemical composition of cilantro leaf oil, as well as the capsaicin present in the jalapeno peppers, will degrade over time. These components have the potential to develop off flavors or change into undesirable irritants. Possible interactions should be studied. While a clear glass jar allows the customer to view the product and lends marketability and safe packaging, it also allows light to interact with the product which will cause damage over time (Clary, 1996). A critical control point is the addition of the dehydrated cilantro to the tomato mixture. In the future, testing to determine yeast and mold counts which conform to standardized limits will be important because the cilantro is added at the end of the heating time. Other testing should include determining soluble solids, water activity, color measurement, and perhaps rheological properties. These tests could be used to establish processing parameters needed for full processing (Resurreccion and Shewfelt, 1985; Lopez et al., 1986; Rodrigo et al., 1990). Some of the consumer comments suggested that smaller tomato chunks or a thicker salsa is desirable. Smaller tomato dice or the addition of tomato paste may be alternatives to the original formulation. Other comments suggested that the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 41 control sample (without cilantro) seemed hotter and the cilantro sample seemed sweeter. Because both formulations were made on the same day, with the same ingredient specifications, one explanation is there may be some interaction between the cilantro and the capsaicin responsible for the heat sensation. The cilantro may have a potentiating effect on the sweetness of the tomato or onion or a masking effect on the jalapeno. These effects are certainly worth further investigation and if found to be true, adjustments should be made to the formulation. Triangle Test The Triangle Test was correctly used to screen potential subjects based on their ability to discriminate a given difference (Meilgaard et al., 1991). Out o f the 14 panelists participating in the Triangle Test in Phase 2, only 7 met the criteria of three out of three correct scores required to serve on the Recognition Threshold panel. Out of that 7, only 5 would commit to continue to serve on the Recognition Threshold panel, which allowed for the minimum amount of data needed to establish a group threshold. In the first group of 9 panelists, red lighting was selected to mask differences among samples. In this study, it had the opposite effect, as reported by 2 panelists. According to the 2 panelists, the samples containing cilantro were identifiable by dark specks, causing bias, yet only 5 of the 9 panelists scored three out of three correct. This suggests that not all of the 9 panelists noticed the difference caused by the lighting effects. It more importantly suggests that the panelists who observed the difference due to lighting may have scored according to visual clues rather than flavor or odor discrimination. This would allow for a panelist to meet the objective of the Triangle Test (ability to distinguish MTVACed Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 42 cilantro), but not possess the sensitivity to flavor and odor necessary to serve on a threshold panel. In this sense, the Triangle Test failed to correctly screen panelists for the sensory abilities needed for threshold testing. The panelists that were unable to distinguish a difference between samples were either unfamiliar with the characteristic flavor and odor of cilantro, in spite of training, or perhaps not sensitive to it at the 0.5% concentration level used for the Triangle Test. Consequently, more panelists were recruited to be screened so that a large enough pool would be available for the Recognition Threshold test. The sensory laboratory was not available to test the second group of 5 panelists for the Triangle Test. Testing occurred in a group rather than individual setting under different lighting conditions. The group setting may have allowed for bias as panelists could observe facial expressions, though no speaking was allowed. The fluorescent lighting allowed for visual clues between samples, yet only 2 of the 5 panelists detected a difference in three out of three sets. Again, this suggests that the visual difference was not noticed by some panelists. Recommendation for Future Studies For future studies, more training of the panelists to insure complete understanding of test objectives and scoring methods is recommended. Assessment of the reproducibility of results for an individual judge may require repeating the test over more than one day. This may be necessary to calibrate a panel within a narrower confidence range. To account for fatigue encountered due to the trigeminal response to the salsa, testing should be conducted over more than one day, under identical test conditions. If possible within budgetary and time constraints, a larger number of subjects should be tested so that a larger pool could be obtained. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 43 Because the green lighting was later found to be effective in masking the MIVACed cilantro added to the control formulation, it is recommended that this be used for any analytical-discriminative testing. Recognition Threshold Test Out of the 5 panelists who met the criteria for panel selection through screening, judge 3 scored inconsistently and was unable to detect the MIVACed cilantro in the salsa sample even at the highest concentration. No threshold was established for this judge. Higher concentrations should be introduced to this judge to establish an individual threshold; however, time constraints and availability of judge 3 prevented further testing during this trial. Further, this judge was in the first group of panelist screen using the red lighting in the sensory booth. It is the opinion of this researcher that this judge correctly identified MTVAC cilantro visually, rather than by flavor and odor, during the Triangle Test. If this is the case, this judge did not meet screening criteria as required by the Triangle Test objectives. This could be verified through repeating the Triangle test screening for this judge under green lighting conditions. Recommendation for Future Studies If this test were to be repeated, several factors should be investigated. One, determine if the screening procedure is adequate for determining that a subject can correctly identify a substance. Two, if time and budget permit, recruit a larger panel for threshold testing to better represent the population (Morrison, 1982). According to Brown, a panel of 25 or more is necessary to offset the variation between individual subjects (Brown et al., 1978). Three, because the effects o f fatigue and trigeminal response are factors due to the pungency o f this product, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 44 perhaps a sample without jalapeno peppers would be an alternative to assist the subjects in correctly identifying the stimulus of interest, cilantro. Phase 3-Affective Consumer Testing Consumer Preference Testing While it was concluded that the Alternative Hypothesis must be accepted for this study, Meilgaard suggests that a “failure to prove significant difference does not mean that proof o f similarity has been obtained”(Meilgaard et al., 1991). Results of the Consumer Preference Test showed only a slightly greater preference for the salsa with cilantro; however, there was no statistical difference in preference for that sample. These results may be explained by some of the psychological factors influencing sensory tests such as the error of expectation and mutual suggestion. The test location has several effects on the results obtained (Meilgaard et al., 1991). While less control can be obtained at a central location than a laboratory, the respondents are truly representative of the targeted population which lends validity to the results (Meilgaard et al., 1991). Finally, although no statistical difference was determined by the Consumer Preference testing, both salsa formulations were very well received. Based on these findings, it would be worth investigating the feasibility o f producing both formulations for sale at the California State University, Fresno Farm Market. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. REFERENCES Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. REFERENCES American Society o f Testing Methods, ASTM. 1995a. Standard Practice for Determination of Odor and Taste Thresholds by a Forced-Choice Ascending Concentration Series Method of Limits. Designation: E679-91. Vol. 15.1. American Society of Testing Methods, ASTM. 1995b. Standard Practice for Defining and Calculating Individual and Group Sensory Thresholds from Forced-Choice Data Sets of Intermediate Size. E 1432-91. Vol. 15.07. American Society of Testing Methods, ASTM. 1995c. Standard Test Method for Sensory Evaluation of Red Pepper Heat. E 1083-88. Vol. X. Amerine, M.A., Pangbom, R.M., Roessler, E.B. 1965. “Principles of Sensory Evaluation o f Food” . Academic Press, New York, NY. Blumenthal, D. 1990. The canning process; old preservation technique goes modem. FDA Consumer 24(7): 14-18. Brown, D.G., Clapperton, J.F., Meilgaard, M.C., and Moll, M. 1978. Flavor thresholds of added substances. Joum. of Amer. Soc. of Brew. Chem. 36: 7380. Carlsen, K. 1995. New Product Development Survey for California State University, Fresno Farm Market Item. Student survey and report, Department of Enology, Food Science and Nutrition, California State University, Fresno, Fresno, CA. Charlet, K. 1994. The heat is on! (new condiments). Prepared Foods 163(5): 75. Clary, C. 1994. Application of microwave vacuum and liquid media dehydration for the production of dried grapes. Ph.D Dissertation, Michigan State University Clary, C. 1996. Personal communication. California State University, Fresno. Cousminer, J.J. 1996. Sales of salsa. Food Technol. 50(1): 70-73. Domblaser, L. 1996. Lightweights drive condiment category. Prepared Foods 165(5): 59-62. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 47 Duran-Guiterrez. E. J. 1995. At the heat of salsa. Student report. Department of Enology, Food Science & Nutrition, California State University, Fresno. Ferris, D.A. 1996. California State University, Fresno revised tomato salsa project budget. August 8. Ferris, D A. 1995, 1997. California State University, Fresno farm market survey. Fuller, G.W. 1994. New Product Development: From Concept to the Marketplace. CRC Press, Boca Raton, FL. Giese, J. 1994. Modem alchemy: use o f flavor in foods. Food Technol. 48(2): 106116. Gillette, M.H., Appel, C.E., and Lego, M.C. 1984. A new method for sensory evaluation of red pepper heat. Joum. of F.S. 49(X): 1028-1033. Hollingsworth, P. 1996. Food shows spotlight subtle innovations. Food Technol. 50(7): 36. Knight, J.B. and Kotschevar, L.H. 1989. In “Quantity Food Production, Planning and Management” 2nd edition, p. 10. Van Nostrand Reinhold, New York. Krajewska, A.M.and Powers, J.J. 1988. Sensory properties of naturally occurring capsaicinoids. Joum. of F.S. 53(3): 902-905. Labensky, S.R. and Hause, A.M. 1995. In “On Cooking.” Prentice Hall Publishers, Englewood Cliffs, NJ. Lawless, H.T. and Claassen, M.R. 1993. Application o f the central dogma in sensory evaluation. Food Technol. 47(6): 139-146. Lopez, A, Williams, HL, and Cooler, FW. 1986. Essential elements in tomatoes and fresh and canned tomato juice. Joum. of F.S. 51(4): 1071-1072. McKinney, H.F. and Wear, F.C. 1987. Zoned microwave drying apparatus and process. U.S. Patent 4,640,020. February 4. McKinney, H.F., Wear, F.C., Sandy, H.L., Petrucci, V.E., and Clary, C.D. 1983. Process of making hollow dried grape. U.S. Patent 4,418,083. November 23. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 48 Meilgaard, M., Civille, G.V., and Carr, B. T. 1991. “Sensory Evaluations Techniques.” 2nd edition. CRC Press, Inc. Boca Raton, FL. Mogelonsky, M. 1995. Sauces Beyond Salsa. American Demographics. 17(5): 10. Morrison, G.R. 1982. Measurement of Flavor Thresholds. J. Inst. Brew. 88(3) 170-174. Ockerman, H.W. 1978. The Source Book for Food Scientists, p. 74. AVT Publishing Co. Westport, Conn. O’Mahony, M. 1986. “Sensory Evaluation of Food: Statistical methods and Procedures.” Marcel Dekker, Inc. New York and Basel. O’Mahony, M. 1995. Sensory measurement in food science: fitting methods to goals. Food Technol. 49(4): 72,74,76-78,80-82. Potter, T.L. 1996. Essential oil composition of cilantro. J. Agric. Food Chem. 44(7): 1824-1826. Resurrection, AVA, and Shewfelt, RL. 1985. Relationships between sensory attributes and objective measurements of postharvest quality of tomatoes. Joum. of F.S. 50(x): 1242-1245. Rodrigo, M., Martinez, A., Sanchis, J., Trama, J. And Giner, V. 1990. Determination o f hot-fill-hold-cool process specifications for cmshed tomatoes. Joum. of F.S. 55(4): 1029-1038. Roessler, E.B. 1984. Statistical Evaluation o f Experimental Data. In “Food Analysis: Principles and Techniques. Marcell Dekker, Inc. New York. Ronsivalli, L.J. and Vieira, E.R. 1992. “Elementary Food Science,” 3rd. ed. p. 311-313. Chapman and Hall, New York. Rosengarten, F. Jr. 1969. “The Book of Spices.” Livingston Publishing Co., Philadelphia, PA. Rutledge, K.P. 1990. Sensory evaluation: method for establishing and training a descriptive flavor panel. Food Technol. 44(12): 78-84. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 49 Rutledge, K.P. 1992. Accelerated training of sensory descriptive flavor analysis panelists. Food Processing 46(11): 114-118. Sensory Evaluation Division o f the Institute o f Food Technologists. 1981. Sensory evaluation guide for testing food and beverage products. Food Technol. 35(11): 50-59. Toops, D. and Kevin, K. 1996. News bytes. Food Processing. 57(8): 16. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. APPENDICES Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. APPENDIX A CONSENT WAIVER FORM Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. CONSENT FORM CALIFORNIA STATE UNIVERSITY, FRESNO DEPARTMENT OF ENOLOGY, FOOD SCIENCE AND NUTRITION CONSENT TO PARTICIPATE IN THE FOLLOWING STUDY I, (please print)___________________________________________ hereby willingly and voluntarily consent to participate as a subject in the research project entitled “Development of “Fresh Flavor” in a Mildly Processed Tomato-Based Salsa” conducted by Dennis Ferris, Ph.D. and Karla Carlsen at the California State University, Fresno. The California State University, Fresno Committee on the Protection of Human Subjects has reviewed and approved the procedures for this research, which is determined to be of minimal risk to participants. I understand the procedures for participating in the project are as follows: I will taste a sample of tomato-based salsa (recipe appears below) and record my results on an accompanying score sheet. To the best of my knowledge, I have no known allergies to any of the ingredients listed in the recipe. I understand that the results of the study will be kept on file at California State University, Fresno for five years in the office of Dr. Dennis Ferris and will be kept confidential. I understand that my participation in this study does not in any way affect my relationship with California State University, Fresno. I understand that I may quit or decline to participate in the study at any time. The salsa contains the following ingredients: Tomatoes jalapeno chili peppers White onion Dehydrated cilantro I fiilly understand the terms of my consent and agree to participate in the study. Signed Date Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. APPENDIX B TEST SCORE SHEETS Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Consumer Acceptance Test Name______________________ Date___ Type of Sample: Tomato Salsa • Please rinse your mouth before starting. Evaluate the product in front of you by looking at it and tasting it. Considering ALL characteristics (APPEARANCE, FLAVOR, and PUNGENCY) indicate your overall opinion by checking one box [ S ]. __ 1 Dislike Extremely LJ Dislike Very Much u □ □ a i __ i ! ! Dislike Dislike Neither Like Like Like Moderately Slightly like nor Slightly Moderately Very Much Like dislike (nl/nd) Comments: Please indicate WHAT in particular you liked or disliked about this product. (USE WORDS NOT SENTENCES). LIKED DISLIKED Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 55 Triangle Test Taster N o._______ Name_____________________________ Date Type o f Sample: Tomato Salsa Instructions: • Rinse with water before starting. • Taste the samples on the tray from left to right. Two samples are identical; one is different. Determine which is the odd/different sample. Circle the code of the odd sample. If no difference is apparent, you must guess. After completing one set, rinse palate with cracker and water before starting on the next set. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 56 Threshold Test Name___________________________ Date___ Type of Sample: Tomato Salsa Instructions: • Rinse with water before starting. • Please taste the samples on the tray from left to right. Two samples are identical; one is different. Determine which is the odd/different sample. Circle the code of the odd sample. If no difference is apparent, you must guess. After completing one set, rinse palate with cracker and water before starting on the next set. Circle the odd sample Set 1 ____ _____ _____ Set 2 ____ _____ _____ Set 3 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 57 Consumer Preference Test Name______________________ Date Type of Sample: Tomato Salsa Please taste both products. Which sample do you prefer? Please circle your choice. 451 862 Why do you prefer this sample? Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. APPENDIX C RAW DATA INCLUDING OPEN-ENDED COMMENT Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 59 Consumer Preference Test Comments: 451 Preferred 451 = Control Formulation without Cilantro right degree of heat seemed fresher, more interesting mix of flavors tastier, warm, not too hot 862 needs smaller (cut) veggies needs more spices ('Mexican') more chile not as hot as the other tastes better fresh tasting it tastes more fresh; 862 tasted roasted and a little salty tasted more spicy better flavor texture and taste, more tomato taste good bite, both are good, I'd buy both 862 has too much cilantro, 451 is hotter,more vinegar, tangy not as spicy, good overall flavor fresh tasting tastes zestier seems to be less spicy, less onion (in sample) less spicy, richer, more vegetable-like because it's hot I like the hot stuff! taste is lighter, more tangy, better flavor, but not as chunky 862 tasted sweeter, 451 is not too hot, nor was it bland, good blend spicy, appealing, liked both, but 451 a bit better sweeter, crunchier, fresher flavor is slightly different better taste because it's spicier not quite as hot as the second (862) one better smell and taste, looks more fresh (color) sharper taste, onions, crisper, heat level medium to hot, brings out flavor Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 60 Consumer Preference Test Comments: 832 Preferred 862 = Control Formulation without Cilantro right degree of heat seemed fresher, more interesting mix of flavors tastier, warm, not too hot 862 needs smaller (cut) veggies needs more spices ('Mexican') more chile not as hot as the other tastes better fresh tasting it tastes more fresh; 862 tasted roasted and a little salty tasted more spicy better flavor texture and taste, more tomato taste good bite, both are good, I'd buy both 862 has too much cilantro, 451 is hotter,more vinegar, tangy not as spicy, good overall flavor fresh tasting tastes zestier seems to be less spicy, less onion (in sample) less spicy, richer, more vegetable-like because it's hot I like the hot stuff! taste is lighter, more tangy, better flavor, but not as chunky 862 tasted sweeter, 451 is not too hot, nor was it bland, good blend spicy, appealing, liked both, but 451 a bit better sweeter, crunchier, fresher flavor is slightly different better taste because it's spicier not quite as hot as the second (862) one better smell and taste, looks more fresh (color) sharper taste, onions, crisper, heat level medium to hot, brings out flavor Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.