- Email: [email protected]
Development of a nutritious acceptable snack bar using micronized flaked lentils
Food Research International 43 (2010) 642–649
Contents lists available at ScienceDirect
Food Research International journal homepage: www.elsevier.com/locate/foodres
Development of a nutritious acceptable snack bar using micronized flaked lentils Donna Ryland *, Marion Vaisey-Genser, Susan D. Arntfield, Linda J. Malcolmson Faculty of Human Ecology, Department of Human Nutritional Sciences, University of Manitoba, Winnipeg, Manitoba, Canada R3T 2N2
a r t i c l e
i n f o
Article history: Received 26 February 2009 Accepted 23 July 2009
Keywords: Product development Micronized flaked lentils Snack bars Preference mapping Consumer acceptability Descriptive sensory analysis Mixture design experiment
a b s t r a c t Study objectives were to formulate a nutritious acceptable snack bar partially replacing oats with micronized flaked lentils (MFL), and to identify the sensory attributes that contribute to consumer acceptability. Six MFL snack bar formulations exhibiting a wide range of flavor and textural characteristics were developed from a mixture designed experiment. These bars and two commercial bars were evaluated by a consumer panel (n = 62) and a descriptive panel (n = 11). The highest mean acceptability values for one commercial sample (6.5) and three MFL bars (6.0) were not significantly different and corresponded to ‘like slightly’ on the 9-point hedonic scale. External preference mapping determined that sweetness, grainy and lentil flavors, hardness, cohesiveness, cohesiveness of mass and moistness had the greatest influence on consumer acceptability. MFL bars contained more dietary fibre, protein and iron in addition to an approximate sevenfold increase in folate over the all oat counterpart. Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction Consumers demand nutritious, convenient, tasty snacks that satisfy their hunger momentarily until the next meal. Snack bars, a food product that fits these criteria, continue to increase in sales according to the ACNielsen MarketTrack (Burn, 2007). Bars with a balanced nutritional profile of calories, fat, carbohydrate and protein as well as vitamins and minerals are being sought that include fibre and whole grains (King, 2006). Pulse products such as lentils are a source of complex carbohydrate, dietary fibre and essential amino acids, and low in fat. Women in particular would benefit from bars containing lentils as they are a good source of folic acid, a nutrient that prevents neural tube defects in newborns, as well as iron, a mineral essential for reducing anemia. Micronization is a process using electromagnetic energy transmitted as waves to penetrate food product. Optimum tempering levels and time, crucial for final product quality, impact the physical and functional properties. Sufficient moisture is required to gelatinize starch. High starch gelatinization decreases cooking time for lentils (Arntfield et al., 1997) ultimately allowing for their inclusion into food products as a food ingredient without a precook or grinding/milling step. The positive effects of micronization on nutritional properties include increased protein digestibility due to decrease in protein solubility (Arntfield et al., 2001; Zheng, Fasina, Sosulski, & Tyler, 1998), increased soluble dietary fibre * Corresponding author. Tel.: +1 204 474 0871; fax: +1 204 474 7592. E-mail address: [email protected] (D. Ryland). 0963-9969/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.foodres.2009.07.032
(Zhao, 2000) and decreased phytic and phenolic acids (Arntfield et al., 2001). Flaking of lentils coming directly from a micronizer increases their ability to take in moisture. Maurer, Fukuda, and Nielsen (2005) developed an acceptable snack bar from bean puree in proportions allowing for nutritional claims to be made for protein, fibre, folic acid and low fat, but no research has been published regarding whole lentils as a snack bar ingredient. The nutritional benefits of lentils cannot be realized unless the developed product is liked by consumers. Measuring the acceptability of the snack bar in terms of attributes is critical for optimizing the sensory characteristics, particularly taste. Using consumers for this task is difficult due to their misunderstanding of the meaning of the attributes that contribute to their acceptability (Greenhoff & MacFie, 1994). External preference mapping relates the product acceptability to attribute intensity measurements obtained from a descriptive panel. It separates products on the basis of sensory characteristics and consumer acceptance simultaneously. High weighted regression coefficients obtained through the partial least squares regression procedure pinpoint the attributes that contribute positively or negatively to consumer acceptance (Tang, Heymann, & Hsieh, 2000). This mapping technique can be successfully used in selecting product formulations that should be moved forward in the product development process (Jaeger, Wakeling, & MacFie, 2000). The objectives of this study were to develop a nutritious acceptable snack bar using micronized flaked lentils (MFL) in the formulation and to identify the sensory attributes that contribute to consumer acceptability for use in future optimization of bar formulations.
643
D. Ryland et al. / Food Research International 43 (2010) 642–649
Table 2 Nutritional composition of commercial snack bars used for focus groups (FG) and preference mapping (PM) studies.
2. Materials and methods 2.1. Materials Eston lentils, Canada No. 1 Grade, grown in 2000 were tempered (16.5% moisture for 16 h) and held to equilibrate at room temperature in a 2 metric tonne tempering bin. Tempered lentils were processed by InfraReady Products (1998) Ltd., Saskatoon, SK using a gas fired 2 ton capacity micronizer (Model MR20). Immediately after micronization, lentils were fed through a Turner Ipswich Flaker (Model 460) with roller mill set to produce a thickness of 1.35 ± 0.5 mm. Final moisture and temperature were 8% and 126 °C, respectively. All ingredients as well as the commercial snack bars used in the study were purchased from a retail supermarket except for liquid honey which was obtained from a local honey cooperative. 2.2. Preliminary recipe development Guidelines for the ingredient proportions for the initial product formulation as well as the method of preparation were obtained from a typical snack bar formulation (Murphy, 1990). The recipe development objectives for the health bar were to minimize the number of ingredients, keeping the fat and simple carbohydrate content as low as possible, incorporating approximately equal amounts of lentils and oats and using a straightforward method of preparation (Table 1). 2.3. Identification of attributes important for snack bar acceptability Four focus groups (n = 14) were conducted to identify the sensory characteristics of snack bars important to women (the demographic group benefiting the most from the folic acid advantage of lentils as an ingredient). Women were between 18 and 55 years of age and had eaten snack bars at least occasionally. Focus groups were led by an experienced facilitator and lasted 60–90 min. Discussion included descriptions of the attributes and their implied acceptability in four snack bars, two of which were prepared in the laboratory (Table 1) and two commercial samples (Table 2). These snack bars were purposely selected to encompass a range of flavors and textures to encourage everyone in the group to voice their honest opinion. Each sample was wrapped in a resealable plastic snack bag (16.5 8.25 cm) and labeled with a 3-digit random number. Samples were presented one at a time. Purified water was available for cleansing the palate as required. Flavor attributes generated from the focus groups included sweet, nutty, (black) pepper, uncooked and fruity. Texture attri-
Table 1 Formulations and preparation method of lentil bars presented at focus group sessions. Ingredient
Plain lentil bar
Cranberry lentil bar
SCMa (g) Honey (g) MFL (g) Granolab (g) Cut dried cranberries (g)
220 30 180 220 0
220 30 140 180 80
Preparation method: Blend together sweetened condensed milk and honey. Add MFL, granola and cranberries alternately in four batches mixing well after each addition. Spread evenly in a 21 cm square aluminum baking pan lined with aluminum foil sprayed with non-stick aerosol vegetable oil. Bake at 150 °C for 30 min. Cool on rack for 15 min, remove foil and cool another 15 min. Wrap in foil and place in resealable freezer bag for storage at 18 °C until required for testing. a Sweetened condensed milk (low fat). b Granola was prepared by combining 280 g quick cooking oatmeal, 80 g canola oil and 50 g liquid honey and baking at 120 °C for 90 min, stirring every 15 min.
Nutrient
Quaker chewy honey raisin (26 g serving) FG, PM (CB1)
Nature valley oats ‘n’ honey (23 g serving) FG
Quaker oatmeal to go oats and honey (47 g serving) PM (CB2)
Energy (cal) Protein (g) Fat (g) Carbohydrate (g) Sugars (g) Starch (g) Dietary fibre (g)
106 1.4 2.6 19 7.3 11 1.0
106 1.9 4.4 16 6 8.6 1.4
194 3.2 5.5 33 13 18 2.5
butes on holding included greasy and sticky. Chewy, hard, soft, crunchy, flexible, falls apart, sticks to teeth, dry, and a bit of moisture were comments noted for textural attributes in the mouth. Preliminary discussion yielded information regarding consumers’ preference for snack bars confirming that generally good flavor/ taste, chewier, harder, not crumbly texture, low fat, low calorie and healthy snack bars were preferred. Some but not all members expressed a desire for dried fruit.
2.4. Selection of bars for preference mapping analysis Bars for preference mapping analysis were selected based on the results from the focus groups. At least eight products with a wide range of attribute intensities are recommended to validate the external preference mapping procedure (Bower & Whitten, 2000; Greenhoff & MacFie, 1994). Two commercial bars (CB1 and CB2, Table 2) with the above criteria noted by the focus group members were selected to compare lentil bars with similar snack bars in the marketplace. A mixture experiment described by Dutcosky, Grossmann, Silva, and Welsch (2006) was designed to achieve the range of attribute intensities required for the six lentil bars employed for developing possible formulations. The concept bars presented to the focus group participants contained five major ingredients – MFL, granola (quick cooking oats, liquid honey, canola oil), low fat sweetened condensed milk, liquid honey and dried cranberries. Minimum and maximum levels for each of the five ingredients were set based on the formulation for the concept bars incorporating as many MFL as possible (to increase their nutritive advantage), having samples with and without dried cranberries, and increasing liquid honey overall to compensate for the larger amount of dry ingredients in the formulation (Table 3). Constraints were applied to some ingredients. For example, a minimum amount of liquid (low fat sweetened condensed milk and honey) was required to combine the ingredients. A black pepper attribute that was noted during focus group discussion was not desirable and since it was thought to be coming from the lentils, an upper limit was set for the lentils.
Table 3 Minimum and maximum ingredient proportions recommended based on proportions used in the formulations presented to the focus groups. Ingredient
Preliminary Proportion in formulation preliminary (g) formulation
Minimum Maximum recommended recommended proportion proportion
SCM Honey MFL Granola Cut dried cranberries
220 30 140 180 80
0.28 0.06 0.19 0.20 0
0.37 0.05 0.22 0.28 0.12
0.40 0.12 0.37 0.40 0.12
644
D. Ryland et al. / Food Research International 43 (2010) 642–649
Table 4 Proportions of ingredients for 25 lentil bar formulations. SCM a
0.34 0.28 0.40 0.37 0.35 0.28 0.28 0.28 0.28 0.28 0.28 0.28 0.40 0.34 0.28 0.40 0.28 0.40 0.37 0.28 0.33 0.37 0.28 0.40 0.28
Honey
MFL
a
Granola a
0.12 0.06 0.12 0.12 0.09 0.09 0.06 0.09 0.12 0.06 0.12 0.06 0.06 0.06 0.06 0.06 0.12 0.12 0.06 0.12 0.09 0.06 0.06 0.06 0.12
0.19 0.27 0.19 0.19 0.29 0.19 0.19 0.19 0.37 0.32 0.37 0.19 0.27 0.19 0.32 0.22 0.28 0.28 0.37 0.19 0.27 0.37 0.37 0.19 0.20
a
0.34 0.28 0.25 0.20 0.20 0.32 0.40 0.32 0.22 0.35 0.22 0.40 0.28 0.29 0.35 0.20 0.20 0.20 0.20 0.35 0.28 0.20 0.20 0.35 0.40
Dried cranberries 0.00a 0.12 0.05 0.12 0.07 0.12 0.07 0.12 0.02 0.00 0.02 0.07 0.00 0.12 0.00 0.12 0.12 0.00 0.00 0.07 0.03 0.00 0.09 0.00 0.00
a Proportion of ingredient in total formulation where the sum of all ingredients is approximately equal to 1.
The d-optimal mixture design incorporated all of the five ingredients in various proportions using Design–Expert software Version 6.0.1 (Stat-Ease, Inc., Minneapolis, MN). The 25-run design was not blocked and contained 15 runs to form the model, five runs to test for lack of fit and five replicates runs (Table 4). The formulation for each of the 25 runs was determined using the proportion for each ingredient multiplied by the batch size of 650 g. Granola was made for all of the 25 batches on the same day, divided into five lots and stored at 20 °C for 1 week. Baking was completed during 1 week with the five replicate runs baked on different days.
An expert panel consisting of seven members with extensive experience in the measurement of flavor intensity and texture attributes of various food products evaluated the 25 lentil bar formulations. A lentil bar sample containing 221 g low fat sweetened condensed milk, 58.5 g liquid honey, 182 g MFL, 195 g granola and 39 g cut dried cranberries was presented as a reference lentil bar sample. The flavor and texture attributes of the lentil bar was evaluated on a15-cm line scale (low = 0; high = 15). The intensity for sweetness, oat flavor, toasted, stickiness to touch, initial bite (soft = 0; hard = 15), cohesiveness, moistness (dry = 0; wet = 15) chewiness and tooth pack of the reference sample was marked at 7.5 on the line scale. Fruity was marked at 2.0. The 25 formulations were rated for intensity of all attributes in relation to the marked reference. The average value from the seven panelists for each attributes was entered into the Design–Expert program for all formulations. Numerical optimization was performed on this data set based on desired outcomes for the six lentil bars. Constraints were set within the mixture design to address the following criteria: lentil flavor was less acceptable than oat/grain flavor; sweetness decreased with the addition of dried cranberries; and liquid to dry ingredient ratios affected the hardness and chewiness properties of the bars. The criteria for each of the six bars were as follows: Bar 1 – low sweet, high fruit flavor; Bar 2 – high sweet, no fruit flavor; Bar 3 – high oat, low toasted, no fruit flavor; Bar 4 – low oat, high toasted, high fruit flavor; Bar 5 – soft, chewy, moist, moderate fruit flavor; Bar 6 – low softness, low chewiness, low moisture, no fruit flavor. Criteria entered into the Design–Expert numerical optimization procedure are shown in Table 5. In most cases the formulation with the highest desirability was chosen unless otherwise noted as the rationale for selection (Table 6).
2.5. Presentation of bars for preference mapping analysis The six lentil bar formulations and the two commercial bars were rated for acceptability on a 9-point hedonic scale and for purchase intent on a 7-point scale. Seven important attributes determined by the focus groups including stickiness to touch with
Table 5 Criteria entered into the Design–Expert numerical optimization procedure for selection of six lentil bar samples. Sensory attribute
Bar 1 – low sweet high fruit flavor
Bar 2 – high sweet; no fruit flavor
Bar 3 – high oat low toasted no fruit flavor
Bar 4 – low oat high toasted high fruit flavor
Bar 5 – soft chewy moist moderate fruit flavor
Bar 6 – low softness low chewiness moist low moisture no fruit flavor
Sweetness Oat flavor Fruity Toasted Sticky to touch Initial bite Cohesiveness Moistness Chewiness Toothpack
Minimum Target = 7.2 Maximum Target = 7.8 Minimum Target = 7.4 Target = 7.6 Target = 7.4 Target = 8.5 Minimum
Maximum Target = 7.2 Minimum Target = 7.8 Minimum Target = 7.4 Target = 7.6 Target = 7.4 Target = 8.5 Minimum
Target = 7.6 Maximum Minimum Minimum Minimum Target = 7.4 Target = 7.6 Target = 7.4 Target = 8.5 Minimum
Target = 7.6 Minimum Maximum Maximum Minimum Target = 7.4 Target = 7.6 Target = 7.4 Target = 8.5 Minimum
Target = 7.6 Target = 7.2 Target = 1.4 Target = 7.8 Minimum Minimum Minimum Maximum Maximum Minimum
Target = 7.6 Target = 7.2 Minimum Target = 7.8 Minimum Maximum Maximum Minimum Minimum Minimum
Table 6 Bar formulations selected from the Design–Expert numerical optimization procedure for presenting to consumer and descriptive panels.
a
Sample
Description
Rationale
SCM
Honey
MFL
Granola
Dried cranberries
LB1 LB2 LB3 LB4 LB5 LB6
Low sweet high fruit flavor High sweet no fruit flavor Low toasted no fruit flavor High toasted high fruit flavor Soft chewy moist moderate fruit flavor Low softness low chewiness low moisture no fruit flavor
Only solution provided Maximize lentils Highest desirability Highest desirability Highest desirability Lowest SCM, lowest sticky
0.29 0.40 0.28 0.32 0.34 0.33
0.09 0.06 0.11 0.07 0.09 0.06
0.26 0.27 0.19 0.29 0.32 0.19
0.30 0.27 0.40 0.20 0.20 0.39
0.07 0.01a 0.02a 0.12 0.04 0.03a
Add to lentil proportion.
645
D. Ryland et al. / Food Research International 43 (2010) 642–649
fingers, sweetness, cranberry flavor (for those bars containing cranberries), cereal/grain flavor, firmness, chewiness, and moisture were measured for appropriateness on a 7-point ‘just right’ scale. Consumers consisted of 62 women from 18 to 50 years old who had eaten snack bars at least once. Samples were coded with 3-digit random numbers and presented to minimize carry over effects (MacFie, Bratchell, Greenhoff, & Vallis, 1989). Consumers evaluated the samples in individual panel booths. The same eight samples were also evaluated by a panel of eleven members trained according to the method of Stone and Sidel (1993) for quantitative descriptive analysis. Definitions, techniques for measuring attributes, line scale endpoint standards and reference sample attribute intensities were determined during seven 45-min sessions. Flavor attributes and their corresponding definitions included: ‘sweet’ taste associated with sucrose; ‘sour/fruity’ flavor associated with dried cranberry; ‘grainy’ flavor associated with an oat based product; ‘lentil’ flavor associated with MFL’s. Texture attributes and their corresponding definitions included: ‘stickiness to touch’ the degree to which the surface of the sample adheres to the finger after pressing gently on the sample for 2 s and releasing; ‘hardness’ the force required to bite completely through the sample placed between the incisors at a rate of one bite per second; ‘cohesiveness’ deformation undergone by the material before rupture when biting completely through the sample using the incisors; ‘cohesiveness of mass’ the degree to which the chewed sample holds together in a mass after 10 chews with the molars; ‘moisture absorption’ the amount of saliva absorbed by the sample after 10 chews; ‘chewiness’ time required to masticate a sample at a constant rate of force application (force required to penetrate a gum drop in one-half a second) to reduce it to a consistency suitable for swallowing; ‘adhesiveness to teeth’ the amount of sample adhering on/in teeth after the sample is swallowed. All attribute intensities were measured on a 15-cm line scale from 0 = low to 15 = high. The eight samples were given 3-digit random codes and presented in a Latin square design (Bower & Whitten, 2000). The eight sample evaluations were replicated (evaluated 1 day later) to determine panelist repeatability. Panelists sat in individual booths and evaluated the samples under red light. Purified water and saltines with unsalted tops were available for cleansing the palate before each sample for both the consumer and trained panels. Consumer and trained panelists were volunteers recruited according to the ethical protocols approved by the University of Manitoba Research Ethics Board. 2.6. Statistical analysis PROC GLM statistical analysis was performed to determine differences in acceptability, purchase intent ratings, and attribute intensities of the eight bars. Main fixed effects were panelist and sample for the consumer panel. For the trained panel, main fixed effects were panelist, replicate and sample with the interactions of panelist and sample and panelist and replicate performed in
order to check panelist consistency. As well the sample by replication interaction was used to determine consistency of samples from time to time. A level of probability less than 0.05 was declared significant. Tukey’s test was used for multiple comparisons of mean values when significant differences between samples were found. Multivariate analysis (PROC FACTOR) produced loadings which were plotted to determine the relationship between the samples and the attribute intensities. Internal preference mapping (relationship among consumers based on sample liking) (PROC PRINQUAL) and external preference mapping (the relationship between the consumer acceptability and attribute intensities) was determined using partial least squares regression (PROC PLS). All analyses were performed using the SAS statistical software (SAS Institute Inc., 1999).
3. Results and discussion 3.1. Selection and quality of six lentil bar samples A large range of sensory characteristics for the snack bar sample set is desirable to predict which ones have the greatest impact on consumer acceptability. Six lentil bar formulations (Table 6) were selected to cover this range using the flavor and texture sensory responses from the 25 samples presented to the expert panel. Water activity, an indicator of storage stability, was measured on lentil bars after approximately 3 weeks storage at room temperature using an AW Spring–Novasina Swiss-made TH-500 instrument. Water activity for all six bars was 60.62 the upper limit set by Troller and Christian (1978) for yeast and mold tolerance. Nutrient profiles were generated with Owl Software (Tech WizardTM, 2003) for the six lentil bar samples and compared to bars containing only oats (Table 7). Energy and total carbohydrate levels were similar between the lentil/oat bars and the all oat bars. Lentil bars had less total fat and more dietary fibre, protein and iron compared to the corresponding all oat bars. The most striking difference was the folate content where the lentil bars contained about 6–9 times more of the nutrient. The amount of folate in one 30 g snack bar would provide 8–10% of the recommended daily amount of 400 lg for adults. 3.2. Sensory attribute intensities by descriptive panel Investigation of significant interactions determined that no particular panelist was consistently scoring the samples or replicates differently than the rest of the group. A significant sample by replication interaction for stickiness to touch could be due to ambient humidity variation between days of sensory testing. All attributes showed a significant sample effect (Table 8). Sour/ fruity flavor, lentil flavor, grainy flavor, stickiness to touch, sweetness and hardness attributes showed the greatest range in mean intensity values among all bars.
Table 7 Major nutrients in snack bars containing lentils and oats (L/O) compared to bars with only oats (O) 30 g serving size). Sample
LB1 LB2 LB3 LB4 LB5 LB6 a b
Folate (lg)
Energy (kcal)
Total fat (g)
Total carbohydrate (g)
Dietary fibre (g)
Protein (g)
Iron (mg)
L/Oa
Ob
L/O
O
L/O
O
L/O
O
L/O
O
L/O
O
L/O
O
87 84 98 77 82 94
91 88 101 80 86 97
2.3 2.1 3.1 1.6 1.6 3.0
2.7 2.5 3.4 2.1 2.1 3.3
13.2 12.4 14.3 12.0 12.9 13.3
13.9 13.2 14.9 12.8 13.8 13.9
3.1 3.0 2.8 3.2 3.4 2.9
1.6 1.5 1.6 1.5 1.5 1.6
4.1 4.4 4.0 4.1 4.5 4.2
3.2 3.5 3.2 3.1 3.3 3.4
1.0 1.0 1.0 1.0 1.0 1.0
0.6 0.6 0.7 0.6 0.6 0.7
34.8 36.2 29.3 38.6 42.8 30.4
4.6 4.5 4.8 4.3 4.6 4.8
Nutrient content of bars containing lentils and oats. Nutrient content of bars replacing lentils with oats.
646
D. Ryland et al. / Food Research International 43 (2010) 642–649
Table 8 Sensory attribute intensities (0 = low; 15 = high) for eight snack bars measured by descriptive panelA. Sensory attribute
LB1
Stickiness to touch Sweetness Sour/fruity flavor Grainy flavor Lentil flavor Hardness Cohesiveness Cohesiveness of mass Moistness Chewiness Adhesiveness to teeth
5.6 7.9 5.9 9.5 7.1 6.3 6.5 8.2 7.7 7.2 4.1
LB2 (2.5) (1.7) (2.2) (1.3) (2.4) (1.2) (1.8) (1.8) (1.6) (1.8) (1.7)
b bc b ab a b a a a a abc
5.0 6.6 2.7 8.5 5.8 4.2 6.0 7.9 7.4 5.8 3.9
LB3 (1.8) (2.0) (2.0) (2.1) (2.4) (1.5) (2.4) (2.3) (1.6) (1.7) (2.1)
bc d d c bc e a a a c abc
8.4 7.1 3.4 8.7 5.4 5.5 6.4 7.1 7.5 6.8 4.4
LB4 (3.0) (2.1) (3.0) (1.8) (2.5) (1.3) (2.2) (2.0) (1.6) (1.4) (2.3)
a cd cd abc c bc a a a abc ab
8.4 8.6 8.6 8.4 6.5 6.3 6.3 8.4 6.7 7.3 4.9
LB5 (3.0) (1.7) (3.5) (2.2) (2.8) (1.9) (2.2) (2.1) (2.0) (1.9) (2.4)
a b a c ab b a a ab a a
3.8 6.9 4.3 8.5 6.5 4.5 6.1 7.3 7.4 6.7 4.8
LB6 (1.5) (2.4) (2.4) (1.5) (2.8) (1.3) (2.1) (1.5) (1.4) (2.2) (2.6)
de d c bc ab de a a a abc a
7.8 6.7 2.6 9.5 5.3 6.1 6.2 7.4 7.3 6.9 4.5
(3.0) (1.9) (2.4) (1.5) (2.5) (1.8) (1.5) (2.2) (2.1) (1.9) (2.5)
a d d a c bc a a a ab ab
CB1
CB2
4.5 (2.2) cd 10.5 (2.2) a 6.7 (4.3) b 4.6 (3.0) e 1.7 (1.9) d 7.7 (1.9) a 6.4 (2.6) a 7.3 (1.9) a 6.4 (2.0) b 6.0 (2.1) bc 3.1 (1.4) c
7.7 (2.8) a 10.0 (2.4) a 2.4 (3.3) d 5.7 (3.6) d 2.4 (3.6) d 5.3 (3.5) cd 3.9 (2.8) b 6.5 (3.4) a 4.9 (2.1) c 7.2 (2.7) a 3.6 (2.1) bc
A Results are mean values of 11 panelists 1 replicate (n = 22) ± (standard deviation). Values with the same letter within the same row are not significantly different (p < 0.05).
Table 9 Variance, eigenvalues and loadings for sensory attributes measured by the descriptive panel determined by principal component analysis. Parameter
PCa1
PC2
PC3
Percentage variance Cumulative percentage variance Eigenvalue (k)
45.7 45.7 5.0
26.0 71.7 2.9
14.6 86.3 1.6
Sensory attribute Stickiness to touch Sweetness Sour/fruity flavor Grainy flavor Lentil flavor Hardness Cohesiveness Cohesiveness of mass Moistness Chewiness Adhesiveness to teeth
Loadings 0.25 0.83 0.15 0.92 0.96 0.33 0.68 0.70 0.88 0.14 0.79
0.71b 0.49 0.94 0.30 0.06 0.72 0.50 0.58 0.05 0.09 0.15
0.10 0.24 0.27 0.14 0.17 0.13 0.36 0.02 0.37 0.97 0.44
a
Principal component. Loadings over 0.40 make an important contribution to the principal component (Stevens, 1992). b
Multivariate analysis showed that three principal components explained 86% of the variability in the sensory attribute data (Table 9). The first principal component explained 46% of the variability and contained the majority of the flavor attributes – sweetness, grainy flavor, lentil flavor as well as textural components dealing with moistness, cohesiveness on initial bite and while chewing, and adhesiveness to teeth. Generally as sweetness increased grainy and lentil flavors decreased (Fig. 1). The second principal component explained 26% of the variability and contained sour/fruity flavor, stickiness to touch and hardness. The harder the snack bars the lower the stickiness to touch. Interestingly chewiness loaded exclusively on the third principal component (15% of the variability). Relating the samples to these sensory characteristics resulted in lentil bar (LB) samples 2, 3, 5 and 6 having the highest lentil and grainy flavors (in close proximity on the bi-plot – Fig. 1) and LB1 and LB4 having the highest sour/fruity flavor (Fig. 1). The commercial samples were sweeter overall (shown on the same side of the bi-plot) with CB1 exhibiting a high degree of hardness (sharing the same quadrant on the bi-plot) and CB2 having low cohesive properties (found in opposite quadrants).
3.3. Acceptability, purchase intent and ‘just right’ responses by consumer panel Consumer acceptability also showed a significant sample effect (Fig. 2). All samples had mean scores that corresponded to
descriptors from ‘neither like nor dislike’ to ‘like moderately’. No significant differences in acceptability were found between LB1, LB3, LB6 and CB2 with the three lentil bars receiving scores corresponding to ‘like slightly’. The average acceptability and purchase intent values for all lentil bars showed no significant differences although LB2, LB4 and LB5 had significantly lower values than both commercial bars. Among the lentil bar samples the percentage of responses in the ‘like’ categories (those responses P5) was highest for LB3 at 76% followed by LB1 at 66% compared to CB2 which received 79% of the responses in these same categories. According to data from the ‘just right’ scales LB6 received the same percentage of attributes with the majority of responses in the ‘just right’ category as CB1 the most liked of the eight samples (Table 10). In both samples sweetness was the attribute that needed improvement. For the commercial sample less sweetness was desired while for the lentil sample more sweetness was desired. LB1 and LB4 had the next highest percentage of attributes in the ‘just right’ category. Improvements for both these samples as with the remaining lentil bar samples included decreasing the cereal/grain flavor. Increase in cranberry flavor was noted as an improvement for LB1 with an increase in moistness noted for LB4. 3.4. Preference mapping Internal preference mapping showed that the 62 consumers divided into four groups (Fig. 3). The largest group, containing 29 members, found the commercial samples most acceptable. The next largest groups were equally divided between those that liked cranberry lentil bars (LB4, LB1 and LB5) and those that liked plain lentil bars (LB2, LB3 and LB6) with 15 and 14 members, respectively. Four consumers found none of the bars acceptable. Implications from these results from a product development perspective indicated that bars formulated as close as possible to the commercial samples in sensory quality would be most successful. The inclusion of fruit did not influence acceptability. External preference mapping using partial least squares regression determined that two factors represented 97% of the variability with 65% from the sensory attributes (x variables) and 32% from the consumer aspects (y variables) (Table 11). Of the 65% variability in the attribute data set, factor 1 represented 45% of the variability with the highest loadings for grainy and lentil flavors. Factor 2 received the highest loadings for stickiness to touch, sweetness, hardness, cohesiveness of mass, and chewiness. Within the consumer data, acceptability was more heavily loaded than purchase intent and contributed to factor 2. Regression coefficients, required for the prediction equations, are a combination of the correlations of all attributes (loadings) with the amount of liking associated with the attributes (weights) and were calculated as the product of the loadings and the weights
647
D. Ryland et al. / Food Research International 43 (2010) 642–649 2
PC2 = 26% 1.5
CB1 LB4
1
sour/fruity flavor hardness cohesiveness of mass sweetness
0.5
LB1
Loadings
cohesiveness
PC1 = 46%
0
moistness lentil flavor adhesiveness to teeth grainy flavor
-0.5 stickiness
LB3
to touch
-1
LB5 LB2 LB6
CB2 -1.5
-2 -2
-1.5
-1
-0.5
0
0.5
1
1.5
2
Loadings
Mean Acceptability/Purchase Intent Value
Fig. 1. Loadings from the first and second principal components for attributes and snack bars evaluated by the descriptive panela. aResults are mean values of 11 panelists with replicated measurements.
9 8 7 6
C 5.6
C 5.2
5 4
c
3
2.7
BC
BC 5.9
C 5.6 c
c
2.9
2.8
bc 3
AB
BC
6
A 7.1
6.5
6
bc
bc
3
3
ab
a 4.1
3.6
2 1
LB5
LB4
LB2
LB6
Acceptability
LB3
LB1
CB2
CB1
Purchase Intent
Fig. 2. Acceptability and purchase intent of eight snack bars evaluated by the consumer panela. aResults are mean values of 62 consumers. Mean values with the same uppercase letter within the attribute ‘‘acceptability” are not significantly different (p < 0.05). Mean values with the same lower case letter within the attribute ‘‘purchase intent” are not significantly different (p < 0.05). bAcceptability: 1 = dislike extremely; 5 = neither like nor dislike; 9 = like extremely. cPurchase Intent: 1 = never; 4 = sometimes; 7 = every opportunity.
(Table 11). The higher the regression coefficient either in a positive or negative direction the more it contributed to the acceptability/ purchase intent. Regression coefficients for factor 1 were highest for sweetness, grainy and lentil flavors. This finding agrees with the report of Bower and Whitten (2000) that aromas and flavors have the most influence on consumer acceptability. Regression coefficients for factor 2 were highest for hardness, cohesiveness, cohesiveness of mass and moistness. The following proposed prediction equation for the ‘flavor factor’ (factor 1) was Y1 Y2 (Consumer acceptability Purchase intent) = 0.1974 (sweetness) + 0.2352 (grainy flavor) + 0.2835 (lentil flavor). A proposed prediction equation for the ‘texture factor’
(factor 2) was Y1 Y2 (Consumer acceptability Purchase intent) = +0.3468 (hardness) + 0.2368 (cohesiveness) + 0.1128 (cohesiveness of mass) + 0.1592 (moistness). To apply this to further product development work with MFL bars the descriptive panel would evaluate sweetness, grainy and lentil flavors, hardness, cohesiveness, cohesiveness of mass and moistness in new formulations and the mean scores entered into the equation. The higher the result the more likely the product formulation would be acceptable to a consumer group comparable to the one in this study. These new formulations could be generated in the numerical option of Design–Expert using data from the original mixture experiment. For example, setting levels for ingredients and attributes at a high level for sweetness and hardness and a low level for attributes related to grainy and lentil flavors, cohesiveness and moistness resulted in a solution selected with desirability of 0.59 and attribute intensities provided (Table 12). 4. Conclusions The development of a nutritious acceptable snack bar made from MFL followed the process from concept to final formulation. A d-optimal mixture design produced a 25 run experiment from which six lentil bar samples were selected using numerical optimization procedures based on criteria determined from focus group results. Regression coefficients were calculated from partial least squares analysis using consumer and quantitative descriptive panel data to predict consumer acceptance/purchase intent. Sweet, grainy and lentil flavors had the highest regression coefficients in factor 1 indicating their relatively high degree of importance to consumers in this snack bar study. Texture attributes of importance included hardness, cohesiveness, cohesiveness of mass and moistness. Compared to snack bars made from oats alone those formulated with lentils and oats contained 8–10% more folate providing about 10% of the daily required amount in a 30 g serving.
648
D. Ryland et al. / Food Research International 43 (2010) 642–649
Table 10 Percentage of attributes with majority of responses in ‘just right’ (JR) category and suggested improvements for snack bars.
a b
Sample
% attributes with majority of responses in JR category (total # of attributes measured)
Suggested improvements
CB1 LB6 LB1 LB4 CB2 LB3 LB5 LB2
83 83 71 71 50 50 43 17
;aSweetness "bSweetness ;Cereal/grain; "cranberry flavor ;Cereal/grain; "moistness ;Stickiness to touch; "firmness; ;chewiness ;Cereal/grain; "sweetness; ;stickiness to touch ;Cereal/grain; "sweetness; "cranberry flavor; "moistness ;Cereal/grain; "sweetness2, ;stickiness to touch; "firmness; ;moistness2
(6) (6) (7) (7) (6) (6) (7) (6)
Decrease (related to high frequency in ‘too much’ category). Increase (related to high frequency in ‘not enough’ category).
3
PC2 = 20%
Group 2 n=15
2
LB4 Loadings
1
Group 1 n=29
LB1 LB5
PC1=30%
CB2
CB1
0
LB6 LB3
Group 4 n=4
-1
-2
LB2 Group 3 n=14
-3 -3
-2
-1
0
1
2
3
Loadings Fig. 3. Groupings of consumers based on acceptability of snack bars using loadings from the first two principal componentsa. aResults are mean values of 62 consumers.
Table 11 Variance, loadings, weights and regression coefficients for attributes (x variables) and consumer acceptability data (y variables) for the first two factors from the partial least squares analysis. Parameter
Factor 1
Factor 2
Percentage variance (x variables – attributes) Cumulative percentage variance (x variables) Percentage variance (y variables – acceptability and purchase intent) Cumulative percentage variance (y variables)
45.2 45.2 88.1 88.1
20.2 65.4 8.9 97.0
Attribute
Stickiness to touch Sweetness Sour/fruity flavor Grainy flavor Lentil flavor Hardness Cohesiveness Cohesiveness of mass Moistness Chewiness Adhesiveness to teeth Acceptability Purchase intent a b
Factor 1
Factor 2 a
Loadings
Weights
RC
0.04 0.47 0.17 0.56 0.63 0.26 0.10 0.10 0.23 0.03 0.17 0.19 0.16
0.11 0.42 0.05 0.42 0.45 0.35 0.14 0.24 0.31 0.09 0.41 0.69 0.73
0.0044 0.1974 0.0085 0.2352 0.2835 0.0910 0.0140 0.0240 0.0713 0.0027 0.0697 0.1311 0.1168
Regression coefficient = loadings * weights. Loadings over 0.40 make an important contribution to the principal component (Stevens, 1992).
Loadings
Weights
RC
0.41b 0.75 0.39 0.29 0.07 0.51 0.37 0.47 0.37 0.44 0.21 0.49 0.28
0.22 0.01 0.25 0.13 0.01 0.68 0.64 0.24 0.43 0.09 0.14 0.80 0.60
0.0902 0.0075 0.0975 0.0377 0.0007 0.3468 0.2368 0.1128 0.1591 0.0396 0.0294 0.3920 0.1680
649
D. Ryland et al. / Food Research International 43 (2010) 642–649 Table 12 Formulation development, new snack bar formulation and sensory characteristics. Ingredient (I) Attribute (A)
Goal
Lower limit
Upper limit
Lower weight
Upper weight
Importance
Ingredient and attribute criteria SCM (I) Honey (I) MFL (I) Granola (I) Sweetness (A) Oat flavor (A) Toasted (A) Initial bite (A) Cohesiveness (A) Moistness (A) New formulation
Maximize Maximize Minimize Maximize Maximize Minimize Minimize Maximize Minimize Minimize
0.28 0.06 0.19 0.20 5.40 5.20 5.80 5.10 5.50 4.90
0.40 0.12 0.37 0.40 9.80 9.10 9.80 9.70 9.70 9.90 Snack bar sensory characteristics
1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1
3 3 3 3 5 3 3 5 3 3
Ingredient
Proportion
Attribute
Mean value from 0 to 15
SCM Honey MFL Granola
0.40 0.09 0.22 0.29
Sweetness Oat flavor Toasted flavor Initial bite Cohesiveness Moistness
8.2 7.3 7.7 6.1 6.5 7.4
Acknowledgements We thank the Natural Sciences and Engineering Council of Canada, SaskEnergy and Manitoba Hydro for financial support and InfraReady Products (1998) Ltd. for providing the micronized flaked lentils. Assistance with data analysis from Dr. K. Mount in the Department of Statistics at the University of Manitoba is also greatly appreciated. References Arntfield, S. D., Scanlon, M. G., Malcolmson, L. J., Watts, B. M., Cenkowski, S., Ryland, D., et al. (2001). Reduction in lentil cooking time using micronization: Comparison of 2 micronization temperatures. Journal of Food Science, 66, 500–505. Arntfield, S. D., Scanlon, M. G., Malcolmson, L. J., Watts, B., Ryland, D., & Savoie, V. (1997). Effect of tempering and end moisture content on the quality of micronized lentils. Food Research International, 30(5), 371–380. Bower, J. A., & Whitten, R. (2000). Sensory characteristics and consumer liking for cereal bar snack foods. Journal of Sensory Studies, 15, 327–345. Burn, D. (2007). On the rise. Food in Canada, 67(1), 28–32. Dutcosky, S. D., Grossmann, M. V. E., Silva, R. S. S. F., & Welsch, A. K. (2006). Combined sensory optimization of a prebiotic cereal product using multicomponent mixture experiments. Food Chemistry, 98, 630–638. Greenhoff, K., & MacFie, H. J. H. (1994). Preference mapping in practice. In H. J. H. MacFie & D. M. H. Thomson (Eds.), Measurement of food preferences (pp. 137–166). London: Chapman & Hall.
Jaeger, S. R., Wakeling, I. N., & MacFie, H. J. H. (2000). Behavioural extensions to preference mapping: The role of synthesis. Food Quality and. Preference, 11, 349–359. King, J. (2006). Nutrition bar update. Nutraceuticals World, 9(1), 32–36. MacFie, H. J., Bratchell, N., Greenhoff, K., & Vallis, L. V. (1989). Designs to balance the effect of order of presentation and first-order carry-over effects in hall tests. Journal of Sensory Studies, 4, 129–148. Maurer, G., Fukuda, G., & Nielsen, S. (2005). Development of bean-based granola bars and cereal. Cereal Foods World, 50(1), 27–32. Murphy, P. (1990). Countlines and cereal bars. In E. B. Jackson (Ed.), Sugar confectionery manufacture (pp. 300–310). Glasgow: Blackie and Son Ltd. SAS Institute Inc., (1999). SAS/STATÒ user’s guide. Cary, NC: SAS Institute. Stat-Ease, Inc., (2000). Design–Expert Software, Version 6.0.1, Minneapolis, MN: StatEase, Inc. Stevens, J. (1992). Applied multivariate statistics for the social sciences. Hillsdale, NJ: Lawrence Erlbaum Associates. Stone, H., & Sidel, J. L. (1993). Sensory evaluation practices. San Diego, CA: Academic Press. Tang, C., Heymann, H., & Hsieh, F. (2000). Alternatives to data averaging of consumer preference data. Food Quality and Preference, 11, 99–104. Tech WizardTM, (2003). Owl Software, Version 3. Columbia, MO: Owl Software. Troller, J. A., & Christian, J. H. B. (1978). Water activity and food. New York: Academic Press. Zhao, Y. (2000). The effects of pretreatment and micronization on the quality and cookability of lentils. Masters Thesis, University of Manitoba, Winnipeg, MB, Canada, pp. 78–82. Zheng, G. H., Fasina, O., Sosulski, F. W., & Tyler, R. T. (1998). Nitrogen solubility of cereals and legumes subjected to micronization. Journal of Agricultural and Food Chemistry, 46, 4150–4157.
Contents lists available at ScienceDirect
Food Research International journal homepage: www.elsevier.com/locate/foodres
Development of a nutritious acceptable snack bar using micronized flaked lentils Donna Ryland *, Marion Vaisey-Genser, Susan D. Arntfield, Linda J. Malcolmson Faculty of Human Ecology, Department of Human Nutritional Sciences, University of Manitoba, Winnipeg, Manitoba, Canada R3T 2N2
a r t i c l e
i n f o
Article history: Received 26 February 2009 Accepted 23 July 2009
Keywords: Product development Micronized flaked lentils Snack bars Preference mapping Consumer acceptability Descriptive sensory analysis Mixture design experiment
a b s t r a c t Study objectives were to formulate a nutritious acceptable snack bar partially replacing oats with micronized flaked lentils (MFL), and to identify the sensory attributes that contribute to consumer acceptability. Six MFL snack bar formulations exhibiting a wide range of flavor and textural characteristics were developed from a mixture designed experiment. These bars and two commercial bars were evaluated by a consumer panel (n = 62) and a descriptive panel (n = 11). The highest mean acceptability values for one commercial sample (6.5) and three MFL bars (6.0) were not significantly different and corresponded to ‘like slightly’ on the 9-point hedonic scale. External preference mapping determined that sweetness, grainy and lentil flavors, hardness, cohesiveness, cohesiveness of mass and moistness had the greatest influence on consumer acceptability. MFL bars contained more dietary fibre, protein and iron in addition to an approximate sevenfold increase in folate over the all oat counterpart. Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction Consumers demand nutritious, convenient, tasty snacks that satisfy their hunger momentarily until the next meal. Snack bars, a food product that fits these criteria, continue to increase in sales according to the ACNielsen MarketTrack (Burn, 2007). Bars with a balanced nutritional profile of calories, fat, carbohydrate and protein as well as vitamins and minerals are being sought that include fibre and whole grains (King, 2006). Pulse products such as lentils are a source of complex carbohydrate, dietary fibre and essential amino acids, and low in fat. Women in particular would benefit from bars containing lentils as they are a good source of folic acid, a nutrient that prevents neural tube defects in newborns, as well as iron, a mineral essential for reducing anemia. Micronization is a process using electromagnetic energy transmitted as waves to penetrate food product. Optimum tempering levels and time, crucial for final product quality, impact the physical and functional properties. Sufficient moisture is required to gelatinize starch. High starch gelatinization decreases cooking time for lentils (Arntfield et al., 1997) ultimately allowing for their inclusion into food products as a food ingredient without a precook or grinding/milling step. The positive effects of micronization on nutritional properties include increased protein digestibility due to decrease in protein solubility (Arntfield et al., 2001; Zheng, Fasina, Sosulski, & Tyler, 1998), increased soluble dietary fibre * Corresponding author. Tel.: +1 204 474 0871; fax: +1 204 474 7592. E-mail address: [email protected] (D. Ryland). 0963-9969/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.foodres.2009.07.032
(Zhao, 2000) and decreased phytic and phenolic acids (Arntfield et al., 2001). Flaking of lentils coming directly from a micronizer increases their ability to take in moisture. Maurer, Fukuda, and Nielsen (2005) developed an acceptable snack bar from bean puree in proportions allowing for nutritional claims to be made for protein, fibre, folic acid and low fat, but no research has been published regarding whole lentils as a snack bar ingredient. The nutritional benefits of lentils cannot be realized unless the developed product is liked by consumers. Measuring the acceptability of the snack bar in terms of attributes is critical for optimizing the sensory characteristics, particularly taste. Using consumers for this task is difficult due to their misunderstanding of the meaning of the attributes that contribute to their acceptability (Greenhoff & MacFie, 1994). External preference mapping relates the product acceptability to attribute intensity measurements obtained from a descriptive panel. It separates products on the basis of sensory characteristics and consumer acceptance simultaneously. High weighted regression coefficients obtained through the partial least squares regression procedure pinpoint the attributes that contribute positively or negatively to consumer acceptance (Tang, Heymann, & Hsieh, 2000). This mapping technique can be successfully used in selecting product formulations that should be moved forward in the product development process (Jaeger, Wakeling, & MacFie, 2000). The objectives of this study were to develop a nutritious acceptable snack bar using micronized flaked lentils (MFL) in the formulation and to identify the sensory attributes that contribute to consumer acceptability for use in future optimization of bar formulations.
643
D. Ryland et al. / Food Research International 43 (2010) 642–649
Table 2 Nutritional composition of commercial snack bars used for focus groups (FG) and preference mapping (PM) studies.
2. Materials and methods 2.1. Materials Eston lentils, Canada No. 1 Grade, grown in 2000 were tempered (16.5% moisture for 16 h) and held to equilibrate at room temperature in a 2 metric tonne tempering bin. Tempered lentils were processed by InfraReady Products (1998) Ltd., Saskatoon, SK using a gas fired 2 ton capacity micronizer (Model MR20). Immediately after micronization, lentils were fed through a Turner Ipswich Flaker (Model 460) with roller mill set to produce a thickness of 1.35 ± 0.5 mm. Final moisture and temperature were 8% and 126 °C, respectively. All ingredients as well as the commercial snack bars used in the study were purchased from a retail supermarket except for liquid honey which was obtained from a local honey cooperative. 2.2. Preliminary recipe development Guidelines for the ingredient proportions for the initial product formulation as well as the method of preparation were obtained from a typical snack bar formulation (Murphy, 1990). The recipe development objectives for the health bar were to minimize the number of ingredients, keeping the fat and simple carbohydrate content as low as possible, incorporating approximately equal amounts of lentils and oats and using a straightforward method of preparation (Table 1). 2.3. Identification of attributes important for snack bar acceptability Four focus groups (n = 14) were conducted to identify the sensory characteristics of snack bars important to women (the demographic group benefiting the most from the folic acid advantage of lentils as an ingredient). Women were between 18 and 55 years of age and had eaten snack bars at least occasionally. Focus groups were led by an experienced facilitator and lasted 60–90 min. Discussion included descriptions of the attributes and their implied acceptability in four snack bars, two of which were prepared in the laboratory (Table 1) and two commercial samples (Table 2). These snack bars were purposely selected to encompass a range of flavors and textures to encourage everyone in the group to voice their honest opinion. Each sample was wrapped in a resealable plastic snack bag (16.5 8.25 cm) and labeled with a 3-digit random number. Samples were presented one at a time. Purified water was available for cleansing the palate as required. Flavor attributes generated from the focus groups included sweet, nutty, (black) pepper, uncooked and fruity. Texture attri-
Table 1 Formulations and preparation method of lentil bars presented at focus group sessions. Ingredient
Plain lentil bar
Cranberry lentil bar
SCMa (g) Honey (g) MFL (g) Granolab (g) Cut dried cranberries (g)
220 30 180 220 0
220 30 140 180 80
Preparation method: Blend together sweetened condensed milk and honey. Add MFL, granola and cranberries alternately in four batches mixing well after each addition. Spread evenly in a 21 cm square aluminum baking pan lined with aluminum foil sprayed with non-stick aerosol vegetable oil. Bake at 150 °C for 30 min. Cool on rack for 15 min, remove foil and cool another 15 min. Wrap in foil and place in resealable freezer bag for storage at 18 °C until required for testing. a Sweetened condensed milk (low fat). b Granola was prepared by combining 280 g quick cooking oatmeal, 80 g canola oil and 50 g liquid honey and baking at 120 °C for 90 min, stirring every 15 min.
Nutrient
Quaker chewy honey raisin (26 g serving) FG, PM (CB1)
Nature valley oats ‘n’ honey (23 g serving) FG
Quaker oatmeal to go oats and honey (47 g serving) PM (CB2)
Energy (cal) Protein (g) Fat (g) Carbohydrate (g) Sugars (g) Starch (g) Dietary fibre (g)
106 1.4 2.6 19 7.3 11 1.0
106 1.9 4.4 16 6 8.6 1.4
194 3.2 5.5 33 13 18 2.5
butes on holding included greasy and sticky. Chewy, hard, soft, crunchy, flexible, falls apart, sticks to teeth, dry, and a bit of moisture were comments noted for textural attributes in the mouth. Preliminary discussion yielded information regarding consumers’ preference for snack bars confirming that generally good flavor/ taste, chewier, harder, not crumbly texture, low fat, low calorie and healthy snack bars were preferred. Some but not all members expressed a desire for dried fruit.
2.4. Selection of bars for preference mapping analysis Bars for preference mapping analysis were selected based on the results from the focus groups. At least eight products with a wide range of attribute intensities are recommended to validate the external preference mapping procedure (Bower & Whitten, 2000; Greenhoff & MacFie, 1994). Two commercial bars (CB1 and CB2, Table 2) with the above criteria noted by the focus group members were selected to compare lentil bars with similar snack bars in the marketplace. A mixture experiment described by Dutcosky, Grossmann, Silva, and Welsch (2006) was designed to achieve the range of attribute intensities required for the six lentil bars employed for developing possible formulations. The concept bars presented to the focus group participants contained five major ingredients – MFL, granola (quick cooking oats, liquid honey, canola oil), low fat sweetened condensed milk, liquid honey and dried cranberries. Minimum and maximum levels for each of the five ingredients were set based on the formulation for the concept bars incorporating as many MFL as possible (to increase their nutritive advantage), having samples with and without dried cranberries, and increasing liquid honey overall to compensate for the larger amount of dry ingredients in the formulation (Table 3). Constraints were applied to some ingredients. For example, a minimum amount of liquid (low fat sweetened condensed milk and honey) was required to combine the ingredients. A black pepper attribute that was noted during focus group discussion was not desirable and since it was thought to be coming from the lentils, an upper limit was set for the lentils.
Table 3 Minimum and maximum ingredient proportions recommended based on proportions used in the formulations presented to the focus groups. Ingredient
Preliminary Proportion in formulation preliminary (g) formulation
Minimum Maximum recommended recommended proportion proportion
SCM Honey MFL Granola Cut dried cranberries
220 30 140 180 80
0.28 0.06 0.19 0.20 0
0.37 0.05 0.22 0.28 0.12
0.40 0.12 0.37 0.40 0.12
644
D. Ryland et al. / Food Research International 43 (2010) 642–649
Table 4 Proportions of ingredients for 25 lentil bar formulations. SCM a
0.34 0.28 0.40 0.37 0.35 0.28 0.28 0.28 0.28 0.28 0.28 0.28 0.40 0.34 0.28 0.40 0.28 0.40 0.37 0.28 0.33 0.37 0.28 0.40 0.28
Honey
MFL
a
Granola a
0.12 0.06 0.12 0.12 0.09 0.09 0.06 0.09 0.12 0.06 0.12 0.06 0.06 0.06 0.06 0.06 0.12 0.12 0.06 0.12 0.09 0.06 0.06 0.06 0.12
0.19 0.27 0.19 0.19 0.29 0.19 0.19 0.19 0.37 0.32 0.37 0.19 0.27 0.19 0.32 0.22 0.28 0.28 0.37 0.19 0.27 0.37 0.37 0.19 0.20
a
0.34 0.28 0.25 0.20 0.20 0.32 0.40 0.32 0.22 0.35 0.22 0.40 0.28 0.29 0.35 0.20 0.20 0.20 0.20 0.35 0.28 0.20 0.20 0.35 0.40
Dried cranberries 0.00a 0.12 0.05 0.12 0.07 0.12 0.07 0.12 0.02 0.00 0.02 0.07 0.00 0.12 0.00 0.12 0.12 0.00 0.00 0.07 0.03 0.00 0.09 0.00 0.00
a Proportion of ingredient in total formulation where the sum of all ingredients is approximately equal to 1.
The d-optimal mixture design incorporated all of the five ingredients in various proportions using Design–Expert software Version 6.0.1 (Stat-Ease, Inc., Minneapolis, MN). The 25-run design was not blocked and contained 15 runs to form the model, five runs to test for lack of fit and five replicates runs (Table 4). The formulation for each of the 25 runs was determined using the proportion for each ingredient multiplied by the batch size of 650 g. Granola was made for all of the 25 batches on the same day, divided into five lots and stored at 20 °C for 1 week. Baking was completed during 1 week with the five replicate runs baked on different days.
An expert panel consisting of seven members with extensive experience in the measurement of flavor intensity and texture attributes of various food products evaluated the 25 lentil bar formulations. A lentil bar sample containing 221 g low fat sweetened condensed milk, 58.5 g liquid honey, 182 g MFL, 195 g granola and 39 g cut dried cranberries was presented as a reference lentil bar sample. The flavor and texture attributes of the lentil bar was evaluated on a15-cm line scale (low = 0; high = 15). The intensity for sweetness, oat flavor, toasted, stickiness to touch, initial bite (soft = 0; hard = 15), cohesiveness, moistness (dry = 0; wet = 15) chewiness and tooth pack of the reference sample was marked at 7.5 on the line scale. Fruity was marked at 2.0. The 25 formulations were rated for intensity of all attributes in relation to the marked reference. The average value from the seven panelists for each attributes was entered into the Design–Expert program for all formulations. Numerical optimization was performed on this data set based on desired outcomes for the six lentil bars. Constraints were set within the mixture design to address the following criteria: lentil flavor was less acceptable than oat/grain flavor; sweetness decreased with the addition of dried cranberries; and liquid to dry ingredient ratios affected the hardness and chewiness properties of the bars. The criteria for each of the six bars were as follows: Bar 1 – low sweet, high fruit flavor; Bar 2 – high sweet, no fruit flavor; Bar 3 – high oat, low toasted, no fruit flavor; Bar 4 – low oat, high toasted, high fruit flavor; Bar 5 – soft, chewy, moist, moderate fruit flavor; Bar 6 – low softness, low chewiness, low moisture, no fruit flavor. Criteria entered into the Design–Expert numerical optimization procedure are shown in Table 5. In most cases the formulation with the highest desirability was chosen unless otherwise noted as the rationale for selection (Table 6).
2.5. Presentation of bars for preference mapping analysis The six lentil bar formulations and the two commercial bars were rated for acceptability on a 9-point hedonic scale and for purchase intent on a 7-point scale. Seven important attributes determined by the focus groups including stickiness to touch with
Table 5 Criteria entered into the Design–Expert numerical optimization procedure for selection of six lentil bar samples. Sensory attribute
Bar 1 – low sweet high fruit flavor
Bar 2 – high sweet; no fruit flavor
Bar 3 – high oat low toasted no fruit flavor
Bar 4 – low oat high toasted high fruit flavor
Bar 5 – soft chewy moist moderate fruit flavor
Bar 6 – low softness low chewiness moist low moisture no fruit flavor
Sweetness Oat flavor Fruity Toasted Sticky to touch Initial bite Cohesiveness Moistness Chewiness Toothpack
Minimum Target = 7.2 Maximum Target = 7.8 Minimum Target = 7.4 Target = 7.6 Target = 7.4 Target = 8.5 Minimum
Maximum Target = 7.2 Minimum Target = 7.8 Minimum Target = 7.4 Target = 7.6 Target = 7.4 Target = 8.5 Minimum
Target = 7.6 Maximum Minimum Minimum Minimum Target = 7.4 Target = 7.6 Target = 7.4 Target = 8.5 Minimum
Target = 7.6 Minimum Maximum Maximum Minimum Target = 7.4 Target = 7.6 Target = 7.4 Target = 8.5 Minimum
Target = 7.6 Target = 7.2 Target = 1.4 Target = 7.8 Minimum Minimum Minimum Maximum Maximum Minimum
Target = 7.6 Target = 7.2 Minimum Target = 7.8 Minimum Maximum Maximum Minimum Minimum Minimum
Table 6 Bar formulations selected from the Design–Expert numerical optimization procedure for presenting to consumer and descriptive panels.
a
Sample
Description
Rationale
SCM
Honey
MFL
Granola
Dried cranberries
LB1 LB2 LB3 LB4 LB5 LB6
Low sweet high fruit flavor High sweet no fruit flavor Low toasted no fruit flavor High toasted high fruit flavor Soft chewy moist moderate fruit flavor Low softness low chewiness low moisture no fruit flavor
Only solution provided Maximize lentils Highest desirability Highest desirability Highest desirability Lowest SCM, lowest sticky
0.29 0.40 0.28 0.32 0.34 0.33
0.09 0.06 0.11 0.07 0.09 0.06
0.26 0.27 0.19 0.29 0.32 0.19
0.30 0.27 0.40 0.20 0.20 0.39
0.07 0.01a 0.02a 0.12 0.04 0.03a
Add to lentil proportion.
645
D. Ryland et al. / Food Research International 43 (2010) 642–649
fingers, sweetness, cranberry flavor (for those bars containing cranberries), cereal/grain flavor, firmness, chewiness, and moisture were measured for appropriateness on a 7-point ‘just right’ scale. Consumers consisted of 62 women from 18 to 50 years old who had eaten snack bars at least once. Samples were coded with 3-digit random numbers and presented to minimize carry over effects (MacFie, Bratchell, Greenhoff, & Vallis, 1989). Consumers evaluated the samples in individual panel booths. The same eight samples were also evaluated by a panel of eleven members trained according to the method of Stone and Sidel (1993) for quantitative descriptive analysis. Definitions, techniques for measuring attributes, line scale endpoint standards and reference sample attribute intensities were determined during seven 45-min sessions. Flavor attributes and their corresponding definitions included: ‘sweet’ taste associated with sucrose; ‘sour/fruity’ flavor associated with dried cranberry; ‘grainy’ flavor associated with an oat based product; ‘lentil’ flavor associated with MFL’s. Texture attributes and their corresponding definitions included: ‘stickiness to touch’ the degree to which the surface of the sample adheres to the finger after pressing gently on the sample for 2 s and releasing; ‘hardness’ the force required to bite completely through the sample placed between the incisors at a rate of one bite per second; ‘cohesiveness’ deformation undergone by the material before rupture when biting completely through the sample using the incisors; ‘cohesiveness of mass’ the degree to which the chewed sample holds together in a mass after 10 chews with the molars; ‘moisture absorption’ the amount of saliva absorbed by the sample after 10 chews; ‘chewiness’ time required to masticate a sample at a constant rate of force application (force required to penetrate a gum drop in one-half a second) to reduce it to a consistency suitable for swallowing; ‘adhesiveness to teeth’ the amount of sample adhering on/in teeth after the sample is swallowed. All attribute intensities were measured on a 15-cm line scale from 0 = low to 15 = high. The eight samples were given 3-digit random codes and presented in a Latin square design (Bower & Whitten, 2000). The eight sample evaluations were replicated (evaluated 1 day later) to determine panelist repeatability. Panelists sat in individual booths and evaluated the samples under red light. Purified water and saltines with unsalted tops were available for cleansing the palate before each sample for both the consumer and trained panels. Consumer and trained panelists were volunteers recruited according to the ethical protocols approved by the University of Manitoba Research Ethics Board. 2.6. Statistical analysis PROC GLM statistical analysis was performed to determine differences in acceptability, purchase intent ratings, and attribute intensities of the eight bars. Main fixed effects were panelist and sample for the consumer panel. For the trained panel, main fixed effects were panelist, replicate and sample with the interactions of panelist and sample and panelist and replicate performed in
order to check panelist consistency. As well the sample by replication interaction was used to determine consistency of samples from time to time. A level of probability less than 0.05 was declared significant. Tukey’s test was used for multiple comparisons of mean values when significant differences between samples were found. Multivariate analysis (PROC FACTOR) produced loadings which were plotted to determine the relationship between the samples and the attribute intensities. Internal preference mapping (relationship among consumers based on sample liking) (PROC PRINQUAL) and external preference mapping (the relationship between the consumer acceptability and attribute intensities) was determined using partial least squares regression (PROC PLS). All analyses were performed using the SAS statistical software (SAS Institute Inc., 1999).
3. Results and discussion 3.1. Selection and quality of six lentil bar samples A large range of sensory characteristics for the snack bar sample set is desirable to predict which ones have the greatest impact on consumer acceptability. Six lentil bar formulations (Table 6) were selected to cover this range using the flavor and texture sensory responses from the 25 samples presented to the expert panel. Water activity, an indicator of storage stability, was measured on lentil bars after approximately 3 weeks storage at room temperature using an AW Spring–Novasina Swiss-made TH-500 instrument. Water activity for all six bars was 60.62 the upper limit set by Troller and Christian (1978) for yeast and mold tolerance. Nutrient profiles were generated with Owl Software (Tech WizardTM, 2003) for the six lentil bar samples and compared to bars containing only oats (Table 7). Energy and total carbohydrate levels were similar between the lentil/oat bars and the all oat bars. Lentil bars had less total fat and more dietary fibre, protein and iron compared to the corresponding all oat bars. The most striking difference was the folate content where the lentil bars contained about 6–9 times more of the nutrient. The amount of folate in one 30 g snack bar would provide 8–10% of the recommended daily amount of 400 lg for adults. 3.2. Sensory attribute intensities by descriptive panel Investigation of significant interactions determined that no particular panelist was consistently scoring the samples or replicates differently than the rest of the group. A significant sample by replication interaction for stickiness to touch could be due to ambient humidity variation between days of sensory testing. All attributes showed a significant sample effect (Table 8). Sour/ fruity flavor, lentil flavor, grainy flavor, stickiness to touch, sweetness and hardness attributes showed the greatest range in mean intensity values among all bars.
Table 7 Major nutrients in snack bars containing lentils and oats (L/O) compared to bars with only oats (O) 30 g serving size). Sample
LB1 LB2 LB3 LB4 LB5 LB6 a b
Folate (lg)
Energy (kcal)
Total fat (g)
Total carbohydrate (g)
Dietary fibre (g)
Protein (g)
Iron (mg)
L/Oa
Ob
L/O
O
L/O
O
L/O
O
L/O
O
L/O
O
L/O
O
87 84 98 77 82 94
91 88 101 80 86 97
2.3 2.1 3.1 1.6 1.6 3.0
2.7 2.5 3.4 2.1 2.1 3.3
13.2 12.4 14.3 12.0 12.9 13.3
13.9 13.2 14.9 12.8 13.8 13.9
3.1 3.0 2.8 3.2 3.4 2.9
1.6 1.5 1.6 1.5 1.5 1.6
4.1 4.4 4.0 4.1 4.5 4.2
3.2 3.5 3.2 3.1 3.3 3.4
1.0 1.0 1.0 1.0 1.0 1.0
0.6 0.6 0.7 0.6 0.6 0.7
34.8 36.2 29.3 38.6 42.8 30.4
4.6 4.5 4.8 4.3 4.6 4.8
Nutrient content of bars containing lentils and oats. Nutrient content of bars replacing lentils with oats.
646
D. Ryland et al. / Food Research International 43 (2010) 642–649
Table 8 Sensory attribute intensities (0 = low; 15 = high) for eight snack bars measured by descriptive panelA. Sensory attribute
LB1
Stickiness to touch Sweetness Sour/fruity flavor Grainy flavor Lentil flavor Hardness Cohesiveness Cohesiveness of mass Moistness Chewiness Adhesiveness to teeth
5.6 7.9 5.9 9.5 7.1 6.3 6.5 8.2 7.7 7.2 4.1
LB2 (2.5) (1.7) (2.2) (1.3) (2.4) (1.2) (1.8) (1.8) (1.6) (1.8) (1.7)
b bc b ab a b a a a a abc
5.0 6.6 2.7 8.5 5.8 4.2 6.0 7.9 7.4 5.8 3.9
LB3 (1.8) (2.0) (2.0) (2.1) (2.4) (1.5) (2.4) (2.3) (1.6) (1.7) (2.1)
bc d d c bc e a a a c abc
8.4 7.1 3.4 8.7 5.4 5.5 6.4 7.1 7.5 6.8 4.4
LB4 (3.0) (2.1) (3.0) (1.8) (2.5) (1.3) (2.2) (2.0) (1.6) (1.4) (2.3)
a cd cd abc c bc a a a abc ab
8.4 8.6 8.6 8.4 6.5 6.3 6.3 8.4 6.7 7.3 4.9
LB5 (3.0) (1.7) (3.5) (2.2) (2.8) (1.9) (2.2) (2.1) (2.0) (1.9) (2.4)
a b a c ab b a a ab a a
3.8 6.9 4.3 8.5 6.5 4.5 6.1 7.3 7.4 6.7 4.8
LB6 (1.5) (2.4) (2.4) (1.5) (2.8) (1.3) (2.1) (1.5) (1.4) (2.2) (2.6)
de d c bc ab de a a a abc a
7.8 6.7 2.6 9.5 5.3 6.1 6.2 7.4 7.3 6.9 4.5
(3.0) (1.9) (2.4) (1.5) (2.5) (1.8) (1.5) (2.2) (2.1) (1.9) (2.5)
a d d a c bc a a a ab ab
CB1
CB2
4.5 (2.2) cd 10.5 (2.2) a 6.7 (4.3) b 4.6 (3.0) e 1.7 (1.9) d 7.7 (1.9) a 6.4 (2.6) a 7.3 (1.9) a 6.4 (2.0) b 6.0 (2.1) bc 3.1 (1.4) c
7.7 (2.8) a 10.0 (2.4) a 2.4 (3.3) d 5.7 (3.6) d 2.4 (3.6) d 5.3 (3.5) cd 3.9 (2.8) b 6.5 (3.4) a 4.9 (2.1) c 7.2 (2.7) a 3.6 (2.1) bc
A Results are mean values of 11 panelists 1 replicate (n = 22) ± (standard deviation). Values with the same letter within the same row are not significantly different (p < 0.05).
Table 9 Variance, eigenvalues and loadings for sensory attributes measured by the descriptive panel determined by principal component analysis. Parameter
PCa1
PC2
PC3
Percentage variance Cumulative percentage variance Eigenvalue (k)
45.7 45.7 5.0
26.0 71.7 2.9
14.6 86.3 1.6
Sensory attribute Stickiness to touch Sweetness Sour/fruity flavor Grainy flavor Lentil flavor Hardness Cohesiveness Cohesiveness of mass Moistness Chewiness Adhesiveness to teeth
Loadings 0.25 0.83 0.15 0.92 0.96 0.33 0.68 0.70 0.88 0.14 0.79
0.71b 0.49 0.94 0.30 0.06 0.72 0.50 0.58 0.05 0.09 0.15
0.10 0.24 0.27 0.14 0.17 0.13 0.36 0.02 0.37 0.97 0.44
a
Principal component. Loadings over 0.40 make an important contribution to the principal component (Stevens, 1992). b
Multivariate analysis showed that three principal components explained 86% of the variability in the sensory attribute data (Table 9). The first principal component explained 46% of the variability and contained the majority of the flavor attributes – sweetness, grainy flavor, lentil flavor as well as textural components dealing with moistness, cohesiveness on initial bite and while chewing, and adhesiveness to teeth. Generally as sweetness increased grainy and lentil flavors decreased (Fig. 1). The second principal component explained 26% of the variability and contained sour/fruity flavor, stickiness to touch and hardness. The harder the snack bars the lower the stickiness to touch. Interestingly chewiness loaded exclusively on the third principal component (15% of the variability). Relating the samples to these sensory characteristics resulted in lentil bar (LB) samples 2, 3, 5 and 6 having the highest lentil and grainy flavors (in close proximity on the bi-plot – Fig. 1) and LB1 and LB4 having the highest sour/fruity flavor (Fig. 1). The commercial samples were sweeter overall (shown on the same side of the bi-plot) with CB1 exhibiting a high degree of hardness (sharing the same quadrant on the bi-plot) and CB2 having low cohesive properties (found in opposite quadrants).
3.3. Acceptability, purchase intent and ‘just right’ responses by consumer panel Consumer acceptability also showed a significant sample effect (Fig. 2). All samples had mean scores that corresponded to
descriptors from ‘neither like nor dislike’ to ‘like moderately’. No significant differences in acceptability were found between LB1, LB3, LB6 and CB2 with the three lentil bars receiving scores corresponding to ‘like slightly’. The average acceptability and purchase intent values for all lentil bars showed no significant differences although LB2, LB4 and LB5 had significantly lower values than both commercial bars. Among the lentil bar samples the percentage of responses in the ‘like’ categories (those responses P5) was highest for LB3 at 76% followed by LB1 at 66% compared to CB2 which received 79% of the responses in these same categories. According to data from the ‘just right’ scales LB6 received the same percentage of attributes with the majority of responses in the ‘just right’ category as CB1 the most liked of the eight samples (Table 10). In both samples sweetness was the attribute that needed improvement. For the commercial sample less sweetness was desired while for the lentil sample more sweetness was desired. LB1 and LB4 had the next highest percentage of attributes in the ‘just right’ category. Improvements for both these samples as with the remaining lentil bar samples included decreasing the cereal/grain flavor. Increase in cranberry flavor was noted as an improvement for LB1 with an increase in moistness noted for LB4. 3.4. Preference mapping Internal preference mapping showed that the 62 consumers divided into four groups (Fig. 3). The largest group, containing 29 members, found the commercial samples most acceptable. The next largest groups were equally divided between those that liked cranberry lentil bars (LB4, LB1 and LB5) and those that liked plain lentil bars (LB2, LB3 and LB6) with 15 and 14 members, respectively. Four consumers found none of the bars acceptable. Implications from these results from a product development perspective indicated that bars formulated as close as possible to the commercial samples in sensory quality would be most successful. The inclusion of fruit did not influence acceptability. External preference mapping using partial least squares regression determined that two factors represented 97% of the variability with 65% from the sensory attributes (x variables) and 32% from the consumer aspects (y variables) (Table 11). Of the 65% variability in the attribute data set, factor 1 represented 45% of the variability with the highest loadings for grainy and lentil flavors. Factor 2 received the highest loadings for stickiness to touch, sweetness, hardness, cohesiveness of mass, and chewiness. Within the consumer data, acceptability was more heavily loaded than purchase intent and contributed to factor 2. Regression coefficients, required for the prediction equations, are a combination of the correlations of all attributes (loadings) with the amount of liking associated with the attributes (weights) and were calculated as the product of the loadings and the weights
647
D. Ryland et al. / Food Research International 43 (2010) 642–649 2
PC2 = 26% 1.5
CB1 LB4
1
sour/fruity flavor hardness cohesiveness of mass sweetness
0.5
LB1
Loadings
cohesiveness
PC1 = 46%
0
moistness lentil flavor adhesiveness to teeth grainy flavor
-0.5 stickiness
LB3
to touch
-1
LB5 LB2 LB6
CB2 -1.5
-2 -2
-1.5
-1
-0.5
0
0.5
1
1.5
2
Loadings
Mean Acceptability/Purchase Intent Value
Fig. 1. Loadings from the first and second principal components for attributes and snack bars evaluated by the descriptive panela. aResults are mean values of 11 panelists with replicated measurements.
9 8 7 6
C 5.6
C 5.2
5 4
c
3
2.7
BC
BC 5.9
C 5.6 c
c
2.9
2.8
bc 3
AB
BC
6
A 7.1
6.5
6
bc
bc
3
3
ab
a 4.1
3.6
2 1
LB5
LB4
LB2
LB6
Acceptability
LB3
LB1
CB2
CB1
Purchase Intent
Fig. 2. Acceptability and purchase intent of eight snack bars evaluated by the consumer panela. aResults are mean values of 62 consumers. Mean values with the same uppercase letter within the attribute ‘‘acceptability” are not significantly different (p < 0.05). Mean values with the same lower case letter within the attribute ‘‘purchase intent” are not significantly different (p < 0.05). bAcceptability: 1 = dislike extremely; 5 = neither like nor dislike; 9 = like extremely. cPurchase Intent: 1 = never; 4 = sometimes; 7 = every opportunity.
(Table 11). The higher the regression coefficient either in a positive or negative direction the more it contributed to the acceptability/ purchase intent. Regression coefficients for factor 1 were highest for sweetness, grainy and lentil flavors. This finding agrees with the report of Bower and Whitten (2000) that aromas and flavors have the most influence on consumer acceptability. Regression coefficients for factor 2 were highest for hardness, cohesiveness, cohesiveness of mass and moistness. The following proposed prediction equation for the ‘flavor factor’ (factor 1) was Y1 Y2 (Consumer acceptability Purchase intent) = 0.1974 (sweetness) + 0.2352 (grainy flavor) + 0.2835 (lentil flavor). A proposed prediction equation for the ‘texture factor’
(factor 2) was Y1 Y2 (Consumer acceptability Purchase intent) = +0.3468 (hardness) + 0.2368 (cohesiveness) + 0.1128 (cohesiveness of mass) + 0.1592 (moistness). To apply this to further product development work with MFL bars the descriptive panel would evaluate sweetness, grainy and lentil flavors, hardness, cohesiveness, cohesiveness of mass and moistness in new formulations and the mean scores entered into the equation. The higher the result the more likely the product formulation would be acceptable to a consumer group comparable to the one in this study. These new formulations could be generated in the numerical option of Design–Expert using data from the original mixture experiment. For example, setting levels for ingredients and attributes at a high level for sweetness and hardness and a low level for attributes related to grainy and lentil flavors, cohesiveness and moistness resulted in a solution selected with desirability of 0.59 and attribute intensities provided (Table 12). 4. Conclusions The development of a nutritious acceptable snack bar made from MFL followed the process from concept to final formulation. A d-optimal mixture design produced a 25 run experiment from which six lentil bar samples were selected using numerical optimization procedures based on criteria determined from focus group results. Regression coefficients were calculated from partial least squares analysis using consumer and quantitative descriptive panel data to predict consumer acceptance/purchase intent. Sweet, grainy and lentil flavors had the highest regression coefficients in factor 1 indicating their relatively high degree of importance to consumers in this snack bar study. Texture attributes of importance included hardness, cohesiveness, cohesiveness of mass and moistness. Compared to snack bars made from oats alone those formulated with lentils and oats contained 8–10% more folate providing about 10% of the daily required amount in a 30 g serving.
648
D. Ryland et al. / Food Research International 43 (2010) 642–649
Table 10 Percentage of attributes with majority of responses in ‘just right’ (JR) category and suggested improvements for snack bars.
a b
Sample
% attributes with majority of responses in JR category (total # of attributes measured)
Suggested improvements
CB1 LB6 LB1 LB4 CB2 LB3 LB5 LB2
83 83 71 71 50 50 43 17
;aSweetness "bSweetness ;Cereal/grain; "cranberry flavor ;Cereal/grain; "moistness ;Stickiness to touch; "firmness; ;chewiness ;Cereal/grain; "sweetness; ;stickiness to touch ;Cereal/grain; "sweetness; "cranberry flavor; "moistness ;Cereal/grain; "sweetness2, ;stickiness to touch; "firmness; ;moistness2
(6) (6) (7) (7) (6) (6) (7) (6)
Decrease (related to high frequency in ‘too much’ category). Increase (related to high frequency in ‘not enough’ category).
3
PC2 = 20%
Group 2 n=15
2
LB4 Loadings
1
Group 1 n=29
LB1 LB5
PC1=30%
CB2
CB1
0
LB6 LB3
Group 4 n=4
-1
-2
LB2 Group 3 n=14
-3 -3
-2
-1
0
1
2
3
Loadings Fig. 3. Groupings of consumers based on acceptability of snack bars using loadings from the first two principal componentsa. aResults are mean values of 62 consumers.
Table 11 Variance, loadings, weights and regression coefficients for attributes (x variables) and consumer acceptability data (y variables) for the first two factors from the partial least squares analysis. Parameter
Factor 1
Factor 2
Percentage variance (x variables – attributes) Cumulative percentage variance (x variables) Percentage variance (y variables – acceptability and purchase intent) Cumulative percentage variance (y variables)
45.2 45.2 88.1 88.1
20.2 65.4 8.9 97.0
Attribute
Stickiness to touch Sweetness Sour/fruity flavor Grainy flavor Lentil flavor Hardness Cohesiveness Cohesiveness of mass Moistness Chewiness Adhesiveness to teeth Acceptability Purchase intent a b
Factor 1
Factor 2 a
Loadings
Weights
RC
0.04 0.47 0.17 0.56 0.63 0.26 0.10 0.10 0.23 0.03 0.17 0.19 0.16
0.11 0.42 0.05 0.42 0.45 0.35 0.14 0.24 0.31 0.09 0.41 0.69 0.73
0.0044 0.1974 0.0085 0.2352 0.2835 0.0910 0.0140 0.0240 0.0713 0.0027 0.0697 0.1311 0.1168
Regression coefficient = loadings * weights. Loadings over 0.40 make an important contribution to the principal component (Stevens, 1992).
Loadings
Weights
RC
0.41b 0.75 0.39 0.29 0.07 0.51 0.37 0.47 0.37 0.44 0.21 0.49 0.28
0.22 0.01 0.25 0.13 0.01 0.68 0.64 0.24 0.43 0.09 0.14 0.80 0.60
0.0902 0.0075 0.0975 0.0377 0.0007 0.3468 0.2368 0.1128 0.1591 0.0396 0.0294 0.3920 0.1680
649
D. Ryland et al. / Food Research International 43 (2010) 642–649 Table 12 Formulation development, new snack bar formulation and sensory characteristics. Ingredient (I) Attribute (A)
Goal
Lower limit
Upper limit
Lower weight
Upper weight
Importance
Ingredient and attribute criteria SCM (I) Honey (I) MFL (I) Granola (I) Sweetness (A) Oat flavor (A) Toasted (A) Initial bite (A) Cohesiveness (A) Moistness (A) New formulation
Maximize Maximize Minimize Maximize Maximize Minimize Minimize Maximize Minimize Minimize
0.28 0.06 0.19 0.20 5.40 5.20 5.80 5.10 5.50 4.90
0.40 0.12 0.37 0.40 9.80 9.10 9.80 9.70 9.70 9.90 Snack bar sensory characteristics
1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1
3 3 3 3 5 3 3 5 3 3
Ingredient
Proportion
Attribute
Mean value from 0 to 15
SCM Honey MFL Granola
0.40 0.09 0.22 0.29
Sweetness Oat flavor Toasted flavor Initial bite Cohesiveness Moistness
8.2 7.3 7.7 6.1 6.5 7.4
Acknowledgements We thank the Natural Sciences and Engineering Council of Canada, SaskEnergy and Manitoba Hydro for financial support and InfraReady Products (1998) Ltd. for providing the micronized flaked lentils. Assistance with data analysis from Dr. K. Mount in the Department of Statistics at the University of Manitoba is also greatly appreciated. References Arntfield, S. D., Scanlon, M. G., Malcolmson, L. J., Watts, B. M., Cenkowski, S., Ryland, D., et al. (2001). Reduction in lentil cooking time using micronization: Comparison of 2 micronization temperatures. Journal of Food Science, 66, 500–505. Arntfield, S. D., Scanlon, M. G., Malcolmson, L. J., Watts, B., Ryland, D., & Savoie, V. (1997). Effect of tempering and end moisture content on the quality of micronized lentils. Food Research International, 30(5), 371–380. Bower, J. A., & Whitten, R. (2000). Sensory characteristics and consumer liking for cereal bar snack foods. Journal of Sensory Studies, 15, 327–345. Burn, D. (2007). On the rise. Food in Canada, 67(1), 28–32. Dutcosky, S. D., Grossmann, M. V. E., Silva, R. S. S. F., & Welsch, A. K. (2006). Combined sensory optimization of a prebiotic cereal product using multicomponent mixture experiments. Food Chemistry, 98, 630–638. Greenhoff, K., & MacFie, H. J. H. (1994). Preference mapping in practice. In H. J. H. MacFie & D. M. H. Thomson (Eds.), Measurement of food preferences (pp. 137–166). London: Chapman & Hall.
Jaeger, S. R., Wakeling, I. N., & MacFie, H. J. H. (2000). Behavioural extensions to preference mapping: The role of synthesis. Food Quality and. Preference, 11, 349–359. King, J. (2006). Nutrition bar update. Nutraceuticals World, 9(1), 32–36. MacFie, H. J., Bratchell, N., Greenhoff, K., & Vallis, L. V. (1989). Designs to balance the effect of order of presentation and first-order carry-over effects in hall tests. Journal of Sensory Studies, 4, 129–148. Maurer, G., Fukuda, G., & Nielsen, S. (2005). Development of bean-based granola bars and cereal. Cereal Foods World, 50(1), 27–32. Murphy, P. (1990). Countlines and cereal bars. In E. B. Jackson (Ed.), Sugar confectionery manufacture (pp. 300–310). Glasgow: Blackie and Son Ltd. SAS Institute Inc., (1999). SAS/STATÒ user’s guide. Cary, NC: SAS Institute. Stat-Ease, Inc., (2000). Design–Expert Software, Version 6.0.1, Minneapolis, MN: StatEase, Inc. Stevens, J. (1992). Applied multivariate statistics for the social sciences. Hillsdale, NJ: Lawrence Erlbaum Associates. Stone, H., & Sidel, J. L. (1993). Sensory evaluation practices. San Diego, CA: Academic Press. Tang, C., Heymann, H., & Hsieh, F. (2000). Alternatives to data averaging of consumer preference data. Food Quality and Preference, 11, 99–104. Tech WizardTM, (2003). Owl Software, Version 3. Columbia, MO: Owl Software. Troller, J. A., & Christian, J. H. B. (1978). Water activity and food. New York: Academic Press. Zhao, Y. (2000). The effects of pretreatment and micronization on the quality and cookability of lentils. Masters Thesis, University of Manitoba, Winnipeg, MB, Canada, pp. 78–82. Zheng, G. H., Fasina, O., Sosulski, F. W., & Tyler, R. T. (1998). Nitrogen solubility of cereals and legumes subjected to micronization. Journal of Agricultural and Food Chemistry, 46, 4150–4157.