See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/325649866

Quality Control of Biscuit Products by Applying Methods of Sensory Analysis—

DSKAS and CATA

Article in Current Nutrition & Food Science · June 2018

DOI: 10.2174/1573401314666180606090306

CITATION READS
1 13,687

2 authors:

Marieta Georgieva Stefanova Denka Zlateva

Nikola Vaptsarov Naval Academy University of Economics - Varna
27 PUBLICATIONS 18 CITATIONS 16 PUBLICATIONS 24 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Marieta Georgieva Stefanova on 26 October 2018.

The user has requested enhancement of the downloaded file.

 Send Orders for Reprints to reprints@benthamscience.ae
Current Nutrition & Food Science, 2018, 14, 391-399 391

RESEARCH ARTICLE

Quality Control of Biscuit Products by Applying Methods of Sensory

Analysis—DSKAS and CATA

Marieta Stefanova* and Denka Zlateva

Department of Commodity Science, Faculty of Economics, Varna University of Economics, bul. Knyaz Boris I № 77,
Varna 9002, Bulgaria

Abstract: Background: Modern sensory evaluation methods have two potential applications in bis-
cuit production: to establish product conformity to regulatory and company requirements and to con-
firm a sustainable product quality or minimum shelf life. To ensure product quality control and effec-
tive management, sensory evaluation procedures that can be implemented during the biscuit produc-
tion process are needed.
Objective: This study aimed to develop and validate a procedure for biscuit product quality manage-
ment based on two sensory analysis methods: Difference Scoring with Key Attribute Scales (DSKAS)
using trained evaluators and Check-All-That-Apply (CATA) using untrained evaluators, such as pro-
ARTICLE HISTORY duction workers.
Received: February 10, 2018 Methods: The potential of two sensory analysis methods, DSKAS and CATA, to serve as real-time
Revised: May 11, 2018
Accepted: May 31, 2018 quality measures during biscuit production was assessed. Various factors that may contribute to relia-
DOI:
ble assessments using these methods were analyzed. A process for sensory evaluation was established
10.2174/1573401314666180606090306 and sensory indicators for biscuit product quality were benchmarked.
Results: Comparable results were obtained using the two methods. Results obtained for the two meth-
ods (despite the stated study limitations) showed no significant difference in the scores of product
quality between the panels of trained and untrained experts. The results indicated that a quality control
system based on these sensory methods has practical applications in biscuit production and permits the
rapid collation of information regarding the causes of deviations in quality indicators.
Conclusion: Thus, the DSKAS and CATA methods could be successfully applied during an uninter-
rupted production process to analyze the quality conformity of biscuit products based on sensory indi-
cators.

Keywords: Biscuit, CATA, DSKAS, quality, sensory analysis, sensory indicators.

1. INTRODUCTION food manufacturing industry [4]. The application of these

methods makes it possible to measure consumer preferences
Biscuit products have been widely popular among con-
and to establish the identity of a certain brand of food prod-
sumers of different age groups for several centuries [1]. The
ucts [5-7].
analysis of biscuit sensory indicators has been subject to a
number of studies focusing on the improvement of their taste According to Munoz et al. [8], discriminative analysis is
properties by taking into account the impact of different fac- very widely applied for quality control based on sensory
tors [2, 3]. However, the application of sensory analysis indicators, particularly the “deviation from reference” meth-
methods to biscuit product quality control in production od, which is used to detect and quantify deviations between
plants during the technological process has not been ana- the properties of a reference sample and a studied sample.
lyzed in sufficient detail in the scientific literature. Some authors discourage the application of this method and
recommend the scoring scales method [9] or benchmarking
Sensory analysis methods for quality control have ad- methods to evaluate not only the degree of deviation, but
vanced dramatically during the second half of the 20th centu-
also to quantitatively evaluate individual sensory indicators
ry, together with the rapid development and expansion of the
defining the desired product characteristics.
Beckley and Kroll [10] applied a method that combines
*Address correspondence to this author at the Department of Commodity
Science, Faculty of Economics, Varna University of Economics, bul. Knyaz
the benchmarking method with descriptive methods for anal-
Boris I № 77, Varna 9002, Bulgaria; Tel: 0359 887 419101; ysis. This approach for quality assessment has been subse-
Fax: 0359 52 605371; E-mail: mstefanova@ue-varna.bg quently revised and improved by Pecore et al. [11] and

1573-4013/18 $58.00+.00 © 2018 Bentham Science Publishers

392 Current Nutrition & Food Science, 2018, Vol. 14, No. 5 Stefanova and Zlateva

Young et al. [12]. The Difference Scoring with Key Attrib- Of course, when using panels composed of non-sensory
ute Scales (DSKAS) method is a version of the method de- analysis experts, many recommendations for the minimum
veloped by Beckley and Kroll; it replaces the scale for prod- number of participants required for alternative profiling in
uct rejection and acceptance with a scale for deviation from CATA are known in the literature to ensure the reliability of
the ideal product. The procedure includes the evaluation of the analysis performed. We believe that an approach in
product indicators that could vary during production and which the sensory evaluation is performed by workers who
those that could be used to precisely characterize defective are directly involved in product production has the potential
products. The DSKAS methodology has been studied and to address these limitations. Specifically, it may eliminate
described in detail by Lawless and Heymann in “Sensory the need to form sensory panels of trained experts, minimize
evaluation of food: principles and practices” [13]. costs, and improve the speed of decision-making for the dis-
During the past few years, the Check-All-That-Apply posal of the product, thus reducing waste.
(CATA) method has been increasingly applied because it can The objective of this study was, therefore, to develop and
be performed by evaluators without preliminary preparation validate a procedure for biscuit product quality management
or training. The CATA method was first applied in 2006 [14] based on two sensory analysis methods: DSKAS (using
and was further improved. Its main flaw is that it uses the trained evaluators) and CATA (using untrained evaluators,
number of repetitions, rather than the intensity of the studied such as production workers). In this study, the sensory
indicator for the analysis [15]. The CATA and DSKAS characteristics were not assessed by a user panel, but by a
methods have been used for scientific and applied studies panel of food manufacturing workers who were then
related to the sensory evaluation of product quality [11, 16]. unknown to the authors. The CATA method coincides with
However, the application of these methods to biscuit product the already established DSKAS method that involves
quality control has not been examined. assessment by an expert panel; the DSKAS method can,
Modern sensory evaluation methods have two potential therefore, be used to validate the results obtained by CATA.
For this purpose, the same batches of test products were used
applications to biscuit production: 1) to establish product
by both panels (expert-DSKAS and not expert-CATA).The
conformity to regulatory and company requirements [8] and
proposed procedure would be applicable to biscuit product
2) to confirm a sustainable product quality or minimum shelf
quality control and management during the uninterrupted
life [17]. In this case, sensory evaluation needs to be per-
production process, with the goal of minimizing time and
formed over a very short period of time; i.e., immediately
after the end of the production process and before product costs.
packaging and distribution. This is a major limitation for the
implementation of most sensory evaluation procedures [18], 2. MATERIALS AND METHODS
which typically require substantial time resources [19]. 2.1 Study Design
At present, the sensory control of biscuit product quality This study was carried out in a Bulgarian factory that
is evaluated after the product has already been packaged. produces biscuit products distributed in the national and in-
This can lead to the disposal of non-conforming products, ternational markets. The products used for the analysis were
resulting in substantial economic losses. Accordingly, an representative of the batch and the product range of the fac-
approach that enables sensory control and quality manage- tory.
ment at an earlier stage, in combination with the timely pro-
vision of information about any deviations from standards The proposed procedure for the control and management
for quality indicators, will be important for minimizing costs of the quality of biscuit products by applying the DSKAS
and maximizing consumer satisfaction. However, this re- and CATA methods comprised of six stages, as shown in
quires experts to conduct the sensory analysis as well as the Fig. (1).
statistical analysis of the results, which is very difficult with Stage One. The precise definition of the target “ideal”
a 24-hour production working regime [20, 21]. This marked- product and scope of acceptable limits of deviation, deter-
ly hinders the timely and effective implementation of most mined by experts from the marketing, technological, research
sensory evaluation procedures. [22, 23]. and development, or quality assurance department.
Preliminary studies were designed to analyze the possi- Stage Two. Identification of participants in the sensory
bilities of replacing panels for analytical testing by expert panel. The panel for both studies should be selected from the
appraisers with testing using non-expert methods by manu- available human resources of the production factory based
facturing workers to control the quality of the final biscuit on their common interests.
product. Many researchers have found that the results of
evaluations conducted with untrained professionals are relia- Stage Three. Training for the panel participants. The pan-
ble and comparable to those conducted with panels of expert elists should be familiar with the main guidelines of the
assessors [24-28]. An argument to support the use of non- methodology. They should be familiar with the product spec-
expert methods of sensory analysis in product control is the ification, the main ingredients, and the acceptable limits of
fact that some of the manufacturers typically cannot afford to deviation for the indicators. The main purpose of the training
use sensory assessment experts for a 24-hour production and for panelists was to illustrate the effects of certain production
control cycle. In this case, the use of available human re- errors and the scope of deviations in the indicators that could
sources (manufacturing workers) may be the only alternative result from such errors.
for conducting objective sensory control of the product [29- Stage Four. Data collection and filling in the question-
32]. naires. After the training period, the panelists fill in the ques-
Quality Control of Biscuits by Applying Sensory Analysis Methods Current Nutrition & Food Science, 2018, Vol. 14, No. 5 393

Fig. (1). The procedure for the control and management of the quality of biscuit products by applying the DSKAS and CATA methods com-
prised six stages.

Table 1. Questionnaire used in the application of the difference scoring with key attribute scales method.

Panelist Date:

Please, note the characteristics of the studied product as compared to the ideal product:
Form No. 1 2 3 4 5
Completely Different Very Different Slightly Different Subtly Different No Difference

Please, evaluate the indicators by writing your score:

Appearance and shape 1 2 3 4 5

Surface 1 2 3 4 5

Color 1 2 3 4 5

Odor 1 2 3 4 5

Taste 1 2 3 4 5

Surface in cases of
1 2 3 4 5
breaking

tionnaires presented in Tables 1 and 2 and evaluate the prod- ples 1 and 2 (with slight deviations from the ideal product),
ucts. samples 3 and 4 (with minor deviations), and samples 5
(without deviations from the ideal product) were analyzed.
Stage Five. Data analysis and interpretation of the results
Each of the samples was representative of a particular prod-
using XLSTAT-Sensory®.
uct class and its quantity allowed for more than one tasting
Stage Six. Analysis of results and decision-making for evaluation in order to better understand the studied product.
product disposal based on evidence. The causes for devia- The products were in a package designed for distribution.
tions from the ideal product are also determined. The analyzed samples were subjected to the typical storage
According to the study goals, 360 (60 DSKAS and 60 and distribution conditions (light, temperature, and relative
CATA) products were selected based on their similarity to or air humidity) with inherent variation immediately prior to
difference from a reference sample (which fully met product and during the test. The scoring procedure was performed in
specifications) and were classified as products with “minor”, the laboratory of the factory for biscuit products under the
“medium”, or “major” deviations from the reference. Sam- necessary conditions for conducting a sensory analysis with
394 Current Nutrition & Food Science, 2018, Vol. 14, No. 5 Stefanova and Zlateva

Table 2. Questionnaire used in the application of the Check All That Apply method.

Panelist: Date:

Please, note the characteristics of the studied product as compared to the ideal one:
Form No. 1 2 3 4 5
Completely different Very different Slightly different Subtly different No difference

Please, check the condition that applies to the evaluated product:

□ The shape is like the shape of the ideal product □ The taste is like the taste of the ideal product

□ The surface is like the surface of the ideal product □ The surface after breaking is like the one of the ideal product

□ The color is like the color of the ideal product □ The odor is like the odor of the ideal product

respect to light intensity, temperature, and air flow according The technologists who participated in the panel had more
to ISO 6658. All samples were marked with randomly as- than 600 hours of collective experience in discriminative and
signed three-digit codes and were presented simultaneously descriptive tests for studying the quality of biscuit products.
for evaluation in equal quantities and in identical containers. The analysis focused on the following indicators: surface,
flavor, color, taste, appearance, and shape.
The evaluators were given questionnaires (Tables 1 and
2), preliminarily drafted on paper. For the purpose of this
study, an “ideal product” was defined as a product with no 2.3. Testing Based on the CATA Method
deviations from the specifications indicated in the manufac- During the application of the CATA method, the sensory
turer’s technological documentation. panel was comprised of 50 workers who had no expertise in
Intensity scales were used in the questionnaire, but a rat- sensory evaluation and were directly involved in the produc-
ing for the ideal product was included along with the actual tion of the analyzed products (25 women and 25 men aged
product rating to estimate deviations from the ideal [33]. In between 20 and 48 years). To compensate for the lack of
the first part of the questionnaire, the scale was divided into expertise, we increased the number of panelists in this exper-
five ranges, allowing for faster disposal of the product based iment to 50. The panelists were selected based on their moti-
on the score. These results provided a basis for decision- vation and availability, and all regularly consumed biscuit
making when products deviated from the specification. The products. The analysis focused on the following indicators:
second part of the questionnaire allowed for the collection of texture, flavor, color, taste, and shape. Each evaluator re-
information about the studied site, which complemented the ceived a questionnaire with descriptors that should be noted
evaluation of the extent of difference from the ideal product. for conformity with the ideal product. After recording the
If a sample was evaluated as different from the reference, attributes, each respondent was required to rank the products
panelists indicated the reason for this difference in the se- according to their preferences.
cond part of the questionnaire. The scores (ranging from 1,
“completely different” to 5, “no difference; Tables 1 and 2) 2.4. Statistical Analysis of the Results
were interpreted based on the possibilities for product dis- The data obtained following the DSKAS and CATA
posal at the end of the production process. Products with a methods were analyzed using XLSTAT-Sensory® [34]. This
score of 1-2 were discarded, products with scores of 3 could tool allows the determination of the indicators that best char-
be used after processing or further treatment, products with acterize the quality of biscuit products and that are definitive
scores of 4 were standard products with minor quality devia- of the product subject to evaluation. Statistical differences
tions and could be sold, and products with a score 5 fully met were assessed using the analysis of variance (ANOVA),
the requirements of the specification and, therefore, of the Cochran’s Q test and dispersion model analyses. The model
ideal product. used was a function of the interaction between product
The results were reported immediately. Testing under the scores and the impact of the evaluators on the objectivity of
two methods (DSKAS and CATA) was carried out in two the score. XLSTAT-Sensory® allows for very rapid automat-
separate sessions (in November 2017). The panelists who ed data processing with a high level of accuracy, which
participated in the studies provided informed consent and could not be achieved by manual processing. The data ob-
received compensation for their participation. tained in the CATA and DSKAS analysis were recorded in
binary format (1 for matches to the ideal product and 0 for
2.2. Testing Based on the DSKAS Method deviations from the ideal product).
The DSKAS method was used to evaluate the quality of 3. RESULTS AND DISCUSSION
biscuit products by a panel of ten expert technologists direct-
ly involved in the control of the analyzed products (three The main purpose of the sensory analysis using the two
men and two women aged between 26 and 49 years). They methods (DSKAS and CATA) was to establish the conformi-
were selected based on their sensitivity to differences and ty of the manufactured products with the established specifi-
their previous experience in the evaluation of biscuit quality. cation, which also includes the customer requirements. The
Quality Control of Biscuits by Applying Sensory Analysis Methods Current Nutrition & Food Science, 2018, Vol. 14, No. 5 395

Table 3. Discriminatory power for various descriptors.

DSKAS CATA
Descriptor
Test value P-value Test value P-value

Color 7.838 0.000 12.022 0.000

Surface in cases of breakage 7.176 0.000 11.526 0.000

Appearance and shape 6.875 0.000 11.445 0.000

Taste 6.410 0.000 11.104 0.000

Odor 4.846 0.000 10.655 0.000

Surface 4.419 0.000 8.023 0.000

Fig. (2). DSKAS sensory profiles. Results of the dispersion analysis are shown. Location of the samples confidence ellipses indicate that the
quality of the analysis is good, with 86.11% of the total inertia explained by the five (F1 to F5) dimensions of the DSKAS method. Location
of the samples confidence ellipses were computed at 5% of error.

results are presented in Tables 3-9. The analyses confirmed For both methods, product 1 should be discarded; prod-
that the panelists correctly identified deviations of indicator ucts 2 and 3 may be used if processed or refined. In both
values from the limits defined in stage one. This suggested panels, products 4 and 5 can be retained. These results
that the method could be used for quality control in the pro- demonstrate that there is no statistically significant differ-
duction of biscuits. ence between the values of product quality assessments
Table 3 presents the significance of each of the listed in- awarded by the two panels.
dicators for the evaluation of biscuit product quality. Surface Table 4 (DSKAS) and Table 5 (CATA) show all possible
in cases of breakage (of the product) had the most discrimi- combinations between the analyzed products and the tested
natory power, whereas the odor condition was the least sig- indicators and can be used for making objective decisions
nificant. The P-values for all indicators were <0.05, indicat- regarding the disposal of the product. The green color corre-
ing statistical significance. sponds to values above the average, whereas the red color
represents values below the average score for each indicator.
The order reflects characteristics with the highest to the
The results marked in green are related to positive-valued
lowest discriminating power on the biscuits, test values with
coefficients, whereas the ones marked in red represent coef-
associated P-values
ficients with negative values.
Fig. (1) (DSKAS) and Fig. (2), Fig. (3) (CATA) summa-
rize the results of the dispersion analysis, including the test Prior to the application of the DSKAS and CATA meth-
ods, a test for independence between rows and columns was
values and confidence intervals. The results for sample 5,
performed (Table 6). The null hypothesis H0 was that the
which did not differ from the ideal product and received the
rows and columns in the table are independent, and Ha was
highest score from the panelists, are presented. Confidence
that there is a relationship between the rows and columns.
ellipses for DSKAS and CATA can be added on the symmet-
Based on the observed P-value of <0.05, the null hypothesis
ric plots, as shown on the Sensory profiles. In our example,
the ellipses confirm that samples 1, 2, and 3 do not contrib- H0 could be rejected. This analysis confirmed that the panel-
ists established actual deviations between studied objects
ute whereas samples 4 and 5 contribute to the dependency
with respect to their sensory attributes. The risk to reject the
between variables.
396 Current Nutrition & Food Science, 2018, Vol. 14, No. 5 Stefanova and Zlateva

Fig. (3). CATA sensory profiles. Results of the dispersion analysis are shown. Location of the samples confidence ellipses indicate that the
quality of the analysis is good, with 70.93% of the total inertia explained by the five (F1 to F5) dimensions of the CATA method. Location of
the samples confidence ellipses were computed at 5% of error.

Table 4. Adjusted means by-product DSKAS.

Product
Indicator
Ideal 5 4 3 2 1

Surface 1.00 1.00 0.60 0.00 0.10 0.10

Odor 1.00 1.00 0.30 0.00 0.10 0.00

Appearance and shape 1.00 1.00 1.00 0.80 0.00 030

Surface in cases of breakage 1.00 1.00 1.00 0.80 0.50 0.00

Taste 1.00 0.90 0.50 0.80 0.60 0.00

Color 100 0.90 1.00 0.30 0.20 0.80

Values show the adjusted mean for each product/indicator combination

Table 5. Adjusted means by product CATA.

Product
Indicator
Ideal 5 4 3 2 1

Surface 1.00 0.72 0.60 0.36 0.12 0.12

Odor 1.00 0.76 0.56 0.56 0.20 0.12

Appearance and shape 1.00 0.80 0.68 0.50 0.16 0.16

Surface in cases of breakage 1.00 0.50 0.38 0.46 0.04 0.12

Taste 1.00 0.70 0.46 0.32 0.58 0.28

Color 1.00 0.54 0.36 0.04 0.44 0.24

Quality Control of Biscuits by Applying Sensory Analysis Methods Current Nutrition & Food Science, 2018, Vol. 14, No. 5 397

Table 6. Test of independence between the rows and the columns.

DSKAS CATA

Chi-square (Observed value) 51.303 77.122

Chi-square (Critical value) 37.652 37.652

Degrees of freedom 25 25

P-value 0.0015 < 0.0001

Alpha 0.05 0.05

Table 7. DSKAS method with expert technologists.

Attribute P-value* Sample 1 Sample 2 Sample 3 Sample 4 Sample 5

Appearance and
0.000 0.100 (ab) 0.100 (ab) 0 (a) 0.600 (ab) 1 (b)
shape

Surface 0.000 0.300 (ab) 0 (a) 0.800 (ab) 1 (b) 1 (b)

Color 0.000 0.800 (a) 0,200 (a) 0.300 (a) 1 (a) 0.900 (a)

Odor 0.001 0 (a) 0.600 (a) 0.800 (a) 0.500 (a) 0.900 (a)

Taste 0.000 0 (a) 0.500 (ab) 0.800 (ab) 1 (b) 1 (b)

Surface in cases of
0.000 0 (a) 0.100 (ab) 0 (a) 0.300 (ab) 1 (b)
breakage
*lowercase letters indicate significant differences, P < 0.05 (from Cochran’s Q test, which compares products independently for each attribute).

Table 8. CATA method with untrained workers.

Attribute P-value* Sample 1 Sample 2 Sample 3 Sample 4 Sample 5

Appearance and shape 0.000 0.120 (a) 0.200 (a) 0.560 (b) 0.560 (b) 0.760 (b)

Surface 0.000 0.160 (a) 0.160 (a) 0.500 (b) 0.680 (b) 0.800 (b)

Color 0.000 0.240 (ab) 0.,440 (bc) 0.040 (a) 0.360 (bc) 0.540 (c)

Odor 0.000 0.280 (a) 0.580 (bc) 0.320 (ab) 0.460 (abc) 0.700 (c)

Taste 0.000 0.120 (ab) 0.040 (a) 0.460 (c) 0.380 (bc) 0.500 (c)

Surface in cases of breakage 0.000 0.120 (a) 0.120 (a) 0.360 (ab) 0.600 (bc) 0.720 (c)
*different lowercase letters indicate significant differences, P < 0.05 (from Cochran’s Q test, which compares products independently for each attribute).

Table 9. Analysis match between ideal and tested products from results of DSKAS (expert technologists) and CATA (untrained
workers) methods.

DSKAS CATA
Attribute
P(No)/(Yes) P(Yes)/(Yes) P(No)/(Yes) P(Yes)/(Yes)

Appearance and shape 32 18 140 110

Surface 19 31 135 115

Color 18 32 169 81

Odor 17 33 133 117

Taste 22 28 175 75

Surface in cases of breakage 36 14 154 96

398 Current Nutrition & Food Science, 2018, Vol. 14, No. 5 Stefanova and Zlateva

null hypothesis H0 while it is true is lower than 0.15% for CONCLUSION

DSKAS, and lower than 0.01% for CATA.
The introduction of a system for the continuous sensory
Table 7 (DSKAS) and Table 8 (CATA) summarize the control of products using CATA and DSKAS sensory analy-
results of the Cochran's Q test and show the relationships and sis methods provides the opportunity for the management
the impact on the variation of each indicator among the test- and disposal of products during production. Our results indi-
ed products. The low P-value (below the significance thresh- cated that a sensory control system based on these sensory
old) confirmed that the panelists found differences in the methods for quality control is applicable in real-life condi-
sensory indicators among the evaluated samples. The great- tions and would allow for the rapid gathering of information
est difference between the tested products was found regard- regarding the causes of any deviations in quality indicators.
ing a lower odor. This system is not innovative in itself, but was adapted
A tabular representation of the frequency of the responses based on the established models for sensory analysis. The
of the panelists for the ideal product and for the product sub- novel contribution of the protocol is the adaptation and ap-
ject to evaluation is provided in Table 9. The analysis allows plication to quality control based on sensory indicators in a
production factory for biscuit (or related) products. The pro-
the interpretation of results for scores that are not marked in
cedure could allow for effective decision-making for prod-
the questionnaire but are present in the evaluation of the ide-
ucts that do not conform to the specification and for the dis-
al product. Table 9 also presents information regarding the
posal of those products. To fulfill this objective, certain con-
frequencies of panelists who gave a (No)/(Yes) and ditions must be met for the sensory evaluation, including the
(Yes)/(Yes) score for each attribute. The representation con- accurate identification of acceptable and unacceptable limits
firms that odor had the highest match rate among the indica- for a conforming product and the clear identification of crite-
tors, with 33 DSKAS and 117 CATA matches among the ria for decision-making and product disposal.
responses for the ideal and tested products. The match rate
for taste was only 75 CATA, and that for surface in case of The greatest challenges with respect to performing the
breakage was only 14. analyses include determining the acceptable deviation and
recruiting an appropriate panel of evaluators with a clear
The application of the CATA method eliminates the need understanding of the established acceptable and unacceptable
for panelists with special training to facilitate ongoing quali- deviations. A clear definition of the “ideal” product meeting
ty control. This enables rapid and adequate decision-making the specification is a necessary condition for the implementa-
with respect to products that do not conform to the specifica- tion of both methods. It was established that the target ideal
tion and timely product disposal. product could not be fixed; rather, acceptable limits of devia-
tion could be established.
3.1. Study Limitations
LIST OF ABBREVIATIONS
• Requirement for a precise definition of the “ideal” prod-
uct and scope of acceptable limits of deviation; CATA = Check-All-That-Apply
• The panelists for both studies should be selected from the DSKAS = Difference Scoring with Key Attribute Scales
available human resources at the biscuit production facto-
ry based on their common interests, not by compulsion or ETHICS APPROVAL AND CONSENT TO PARTICI-
by order; PATE

• The panel participants should be familiar with the prod- Not applicable.
uct specification, the main ingredients, and the acceptable
limits of deviation for the quality indicators; HUMAN AND ANIMAL RIGHTS
• The evaluators were given questionnaires (presented in No Animals/Humans were used for studies that are the
Tables 1 and 2) that had not been used in previous studies basis of this research.
for such a quality control purpose. In the future, these
questionnaires may need to be processed for better recep- CONSENT FOR PUBLICATION
tivity by untrained workers; Not applicable.
• Repeatability of the results may be improved if there is a
constancy of panel members of untrained workers CONFLICT OF INTEREST
(through which they gain experience). The authors declare no conflict of interest, financial or
The results obtained for the two methods (including the otherwise.
stated study limitations) show that there is no major, signifi-
cant difference in the scores for the evaluation of product ACKNOWLEDGEMENTS
quality between the panels of trained and untrained experts. We would like to express our gratitude to “ZIV” Ltd.
This indicates that both methods could be used for the quali- Bulgaria for providing the opportunity to conduct the senso-
ty assessment and control of biscuit products and that they ry analyses and for providing the products for testing in one
both could provide reliable information regarding the ana- of the company factories for the production of biscuit prod-
lyzed products. ucts.
 Quality Control of Biscuits by Applying Sensory Analysis Methods Current Nutrition & Food Science, 2018, Vol. 14, No. 5 399

We would like to thank Editage (www.editage.com) for [17] Kilcast D, Subramaniam P. The Stability and Shelf-life of Food. 1st
English language editing. ed. CRC Press: Boca Raton 2000; pp. 79-103.
[18] Reece RN. A quality assurance perspective on sensory evaluation.
Food Technol 1979; 33(9): 37.
REFERENCES [19] Poste LM, Mackie DA, Butler G, et al. Laboratory Methods for
Sensory Analysis of Food. Ottawa: Communication Group 1991.
[1] Manley DB. Cookie, and Cracker Manufacturing, Manual 1: Ingre- [20] Mangan PP. Performance assessment of sensory panelists. J Sens
dients, Vol. 1. Wood head Publishing: London 1998. Stud 1992; 7: 229-52.
[2] Manley D. Manley’s Technology of Biscuits, Crackers and Cook- [21] Chambers IVE, Smith EA. Effects of testing experience on perfor-
ies, 4th ed. Woodhead Publishing: London 2011. mance of trained sensory panelists. J Sens Stud 1993; 8: 155-66.
[3] Zhou W, Hui YH. Bakery Products: Science and Technology, 2nd [22] O'Mahony M. Sensory evaluation of food: Statistical methods and
ed. John Wiley & Sons: New Jersey 2014; pp. 285-98. procedures. Routledge, 2017.
[4] Amerine M, Pangborn P, Roessler E. Principles of Sensory Evalua- [23] Hort J, Kemp SE, Hollowood T. Time-dependent Measures of
tion of Food. Academic Press: New York 2013; pp. 1-27. Perception in Sensory Evaluation. John Wiley & Sons, 2017.
[5] Carpenter RP, Lyon DH, Hasdell TA. Guidelines for Sensory [24] Ares G, Varela P. Trained vs. consumer panels for analytical test-
Analysis in Food Product Development and Quality Control, 2nd ed. ing: Fueling a long lasting debate in the field. Food Quality Pref
Springer US: New York 2012; pp. 169-74. 2017; 61: 79-86.
[6] Meilgaard МC, Carr BТ, Civille GV. Sensory Evaluation Tech- [25] Ares G, Varela P. Authors’ reply to commentaries on Ares and
niques, 4th ed. CRC Press: Boca Raton 2006; pp. 255-81. Varela. Food Quality Pref 2017; 61: 100-2.
[7] Stone H, Bleibaum R, Thomas H. Sensory Evaluation Practices, [26] Ares G, Varela P. Consumer-Based Methodologies for Sensory
4th ed. Academic Press: New York 2012; pp. 327-93. Characterization. In: Methods in Consumer Research, Vol 1. 2018;
[8] Munoz АM, Civille GY, Carr BT. Sensory Evaluation in Quality pp. 187-209.
Control, 2nd ed. Van Nostrand Reinhold, Springer US: New York [27] Jaeger SR, Hort J, Porcherot C, et al. Future directions in sensory
1992; pp. 11-22. and consumer science: Four perspectives and audience voting.
[9] Aust LB, Gacula MC, Beard SA. Washam II RW. Degree of differ- Food Quality Pref 2017; 56: 30-9.
ence test method in sensory evaluation of heterogeneous product [28] Muñoz AM, Kemp SE, Hollowood T, et al. Comparison of De-
types. J Food Sci 1985; 50: 511-3. scriptive Analysis Methods. Descrip Anal Sens Eval 2018, 679-
[10] Beckley JP, Kroll DR. Searching for sensory research excellence. 709.
Food Technol 1996; 50: 61-4. [29] Ares G, Tárrega A, Izquierdo L, et al. Investigation of the number
[11] Pecore S, Stoer N, Hooge S, et al. Degree of difference testing: A of consumers necessary to obtain stable sample and descriptor con-
new approach incorporating control lot variability. Food Quality figurations from check-all-that-apply (CATA) questions. Food
Pref 2006; 17: 552-5. Quality Pref 2014; 31: 135-41.
[12] Young TA, Pecore S, Stoer N, et al. Incorporating test and control [30] Giacalone D. Sensory and Consumer Approaches for Targeted
product variability in degree of difference tests. Food Quality Pref Product Development in the Agro-Food Sector. In: Case Studies in
2008; 19: 734-6. the Traditional Food Sector 2018; pp. 91-128.
[13] Lawless HT, Heymann H. Sensory Evaluation of Food: Principles [31] Vidal L, Cadena RS, Antúnez L, et al. Stability of sample configu-
and Practices, 2nd ed. Springer US: New York 2010; pp. 415-7. rations from projective mapping: How many consumers are neces-
[14] Driesener C, Romaniuk J. Comparing methods of brand image sary? Food Quality Pref 2014; 34: 79-87.
measurement. Int J Market Res 2006; 48: 681-98. [32] Nielsen S. Suzanne. Introduction to Food Analysis. In: Food Anal-
[15] Dooley L, Lee Y, Meullenet JF. The application of check-all-that- ysis. Springer, Cham 2017; pp. 3-16.
apply (CATA) consumer profiling to preference mapping of vanilla [33] Van Trijp HC, Punter PH, Mickartz F, et al. The quest for the ideal
ice cream and its comparison to classical external preference map- product: Comparing different methods and approaches. Food
ping. Food Quality Pref 2010; 21: 394-401. Quality Pref 2007; 18: 729-40.
[16] Lassoued N, Delarue J, Launay B, et al. Baked product texture: [34] XLSTAT-Sensory, XLSTAT for Microsoft Excel Statistical Soft-
Correlations between instrumental and sensory characterization us- ware & Data Analysis Add-on for Excel. In: | XLSTAT-Sensory,
ing Flash Profile. J Cereal Sci 2008; 48: 133-43. 19.2 ed. Microsoft Corp., Redmond 2017.

DISCLAIMER: The above article has been published in Epub (ahead of print) on the basis of the materials provided by the author. The Editorial Department
reserves the right to make minor modifications for further improvement of the manuscript.

View publication stats