Overall Difference Tests
Zaki Utama, STP., MP. Program Studi Teknologi Pangan dan Hasil Pertanian Fakultas Teknologi Pertanian Universitas Gadjah Mada
Klasifikasi Uji Sensoris berdasarkan tujuan pengujian • Uji berorientasi konsumen Consumer-oriented (affective) tests • Uji berorientasi produk Product-oriented (analytical) tests
Consumer-oriented tests • Preference tests allow consumers to express a choice between samples; one sample is preferred and chosen over another or there is no preference
• Acceptance tests to determine the degree of consumer acceptance for a product
• Hedonic tests to measure degree of liking for a product
Product-oriented tests • Difference tests to determine whether two samples can be distinguished from each other by sensory analysis
• Ranking for intensity tests
to obtain preliminary information on product differences, or to screen panelists
• Scoring for intensity tests
measure the amount of difference between samples, and allow samples to be ordered by increasing or decreasing intensity of characteristic
• Descriptive tests
trained panelists provide a total sensory description of the sample, including appearance, odour, flavour, texture and aftertaste
Methods Selection Consumer Acceptability Question?
yes
no Choose from: Preference/choice Ranking Rated Acceptability
no
go to Panel Setup
Sensory Analytical Question? yes
Simple Same/different Question?
re-open discussion of objectives
no
no
yes
Choose from: Overall difference tests n-alternative forced choice Rated difference from control
go to Panel Setup Nature of Difference Question? yes
Choose from: descriptive analysis techniques or modifications
go to Panel Setup
Fig. 1.4 A sensory evaluation department may interact with many other departments in a food or consumer products com-
company’s marketing efforts. However, they may also interH.T. Lawless, H. Heymann, Sensory Evaluation of Food, 2010. act with quality control, marketing research, packaging and
Difference Tests Overall difference tests
Attribute difference tests
Does a sensory difference exist between samples?
How Does Attribute X Differ Between Samples?
These are tests, which are designated to show subjects can detect any difference at all between samples.
Subjects are asked to concentrate on a single (or a few) attribute(s).
Pemilihan metoda yang tepat • Berdasarkan jumlah sampel yang diuji Ø Dua (2): duo-trio test, triangle test, pair comparison, two-outof-five test, “A” – “Not A” test Ø Lebih dari dua: perbedaan terhadap kontrol, ranking, scoring, dan/atau multiple comparison test
• Sifat atribut yang dinilai Ø Sulit dideskripsikan seperti karakter flavor: duo-trio test, triangle test, pair comparison, two-out-of-five test, “A” – “Not A” test Ø Jika perbedaan sudah dapat diprediksi perbedaan intensitasnya: menggunakan attribute difference test seperti ranking, scoring, dan multiple comparison test
Overall Difference Tests • Triangle test • Duo-trio test • Difference-from-control test • Same/Different Test (Simple difference test) • “A” – “Not A” test • Two-out-of-five test
Overall Difference Test • Efektif pada situasi: − Untuk menentukan apakah perbedaan produk dihasilkan dari perubahan bahan, pengolahan, pengemasan, atau penyimpanan − Untuk menentukan apakah terdapat perbedaan menyeluruh, dimana tidak ada atribut spesifik teridentifikasi yang terpengaruh − Untuk menyeleksi dan memonitor kemampuan panelis dalam membedakan (triangle dan two-out-offive)
Menilai warna • Pastikan panelis tidak buta warna dan gangguan pada mata • Manusia sangat baik mengevaluasi perbedaan warna ketika sampel berdampingan atau ada standard, namun tidak baik ketika mengevaluasi berdasarkan memori. • Makanan yang kenampakannya berubah dengan cepat dapat menimbulkan permasalahan. • Gunakan warna-warna yang bersifat putih, cream atau abuabu pada area evaluasi (booth panel, ruang pelatihan/ pengujian) untuk memberikan suasana yang netral dan nonreflektif. • Perhatikan sudut pandang panelis. • Perhatikan jenis dan letak lampu penerangan.
Table 12.2 Light sources, color temperatures, and color rendering indicesa Color rendering Light source
Color temperature (K)
Ambiance description
Index (Ra )
Quality
Candle High-pressure sodium lamp 40 watt incandescent light bulb 100 watt incandescent light bulb CIE source A
1,800 2,100 2,770 2,870 2,856
Very warm Very warm Warm Warm Warm
22 Close to 100 Close to 100 Close to 100
Poor Excellent Excellent Excellent
Warm white fluorescent light Sylvania T5-warm
3,000
Warm
82
Very good
Metal halide lamp Sylvania MetalArc ProTech GroLux Wide Spectrum lamp
3,000 3400
Warm Neutral
85+ 89
Very good Excellent
Neutral fluorescent light PureLite
3,500
Neutral
85
Very good
Cool white fluorescent light Sylvania T5-cool
4,100
Cool
82
Very good
Tungsten/halogen light SoLux CIE source B (direct sunlight)
4,700 4,870
Cool Cool
99
Excellent
Full spectrum fluorescent light DuroTest Vitalite
5,500
Cool
90
Excellent
Daylight fluorescent light Sylvania F40D CIE source D65
6,300 6,500
Cool-blue Cool-blue
76 100
Good Excellent
92
Excellent
Daylight fluorescent light DuroTest DayLite 65 6,500 Cool-blue CIE source C (overcast daylight) 6,774 Cool-blue CIE source D (daylight) 7,500 Cool-blue a Values collated from commercial literature and Hutchings (1999)
H.T. Lawless, H. Heymann, Sensory Evaluation of Food, 2010.
Same/Different Test (Simple Difference Test) • Aplikasi − Membandingkan sampel dengan flavor yang kuat dan tak mudah hilang − Sampel perlu diaplikasikan ke kulit pada half-face test − Sampel memiliki stimuli yang kompleks dan membingungkan panelis
Physiology & Behavior, Vol. 32, pp. 995-998. Copyright©Pergamon Press Ltd., 1984. Printed in the U.S.A.
0031-9384/84 $3.00 + .00
Effects of Oral Chemical Irritation on Taste HARRY LAWLESS
Monell Chemical Senses Center, 3500 M a r k e t Street, Philadelphia, P A 19104 AND DAVID A. STEVENS
Psychology Department, Clark University, Worcester, M A 01610 R e c e i v e d 12 S e p t e m b e r 1983 LAWLESS, H. AND D. A. STEVENS. Effects of oral chemical irritation on taste. PHYSIOL BEI-IAV 32(6) 995-998, 1984.--Oral irritation was induced by rinses with capsicum oleoresin and with piperine, constituents of red and black pepper, respectively. The perceived intensities of two concentrations of each of four tastants representing the four classical taste qualities were evaluated after rinsing with these irritants. Comparing taste intensity after rinses with capsicum and after control rinses with emulsifying agents or water, there were significant decrements in taste intensity of citric acid and quinine, and on one concentration of sucrose, but no effect on salt. The effects of piperine were more broad, with significant decrements in perceived intensity relative to emulsion controls for all substances. Taste
Trigeminal nerve
Oral sensations
Sensory inhibition
IN humans, gustatory sensation is mediated in part by branches of the facial nerve innervating the taste buds of the fungiform papillae on the front of the tongue via the chorda tympani and the taste buds on the soft palate via the greater
called mixture suppression [2, 11, 12]. In olfaction, two odorants presented simultaneously often undergo decrements in perceived intensity relative to their intensity when presented alone, commonly referred to as odor counterac-
Physiology& Behavior,Vol. 57, No. 3, pp. 421-429, 1995 Copyright © 1995ElsevierScienceLtd Printedin the USA. All rights reserved 0031-9384/95 $9.50 + .00
Pergamon 0031-9384(94)E0076-G
Effects of Capsaicin Desensitization on Taste in Humans TARCEY
KARRER I AND LINDA BARTOSHUK 2
Otolaryngology, Yale University School of Medicine, New Haven, CT 06510 R e c e i v e d 17 J u n e 1991 KARRER, T. AND L. M. BARTOSHUK. Effectsof capsaicindesensitizationon taste in humans. PHYSIOL BEHAV 57(3) 421-429, 1995.--Tbe desensitization effects on taste resulting from application of 100 or 10 ppm capsaicin, accompanied by daily testing of a capsaicin series (1-1000 ppm, in log steps), or the desensitization resulting from application of 100 ppm capsaicin without the daily capsaicin testing, were investigated. The taste stimuli were three concentrations each of NaCI, sucrose, citric acid, quinine and 6-npropylthiouracil. Following Other type of 100 ppm desensitization, the magnitude estimates of the two bitter tastes, in particular, and citric acid showed significant decrements. Following 10 ppm capsaicin desensitization or an ethanol coua'oi procedure, there were no such effects. Recovery was complete in 1-3 days. It seems possible that the taste decrements are due to effects on both the taste and tactile components of taste, though there is a stronger case for effects on the tactile component. Capsaicin
Desensitization
Taste
Irritation
THERE have been several investigations of the interaction of concurrent capsaicin b u m and taste stimuli, in which low concentrations of capsaicin [ 1 - 2 ppm (rag/l)] reduced the intensities of either sour alone or all four tastes (5,20), or in which stronger
Potential effects of capsaicin desensitization on suprathreshold taste can be divided into two categories. First, capsaicin desensitization may simply remove any irritation sensation produced by taste solutions, particularly at the higher con-
Menthol: A refreshing look at this ancient compound Tejesh Patel, MD,a Yozo Ishiuji, MD,a and Gil Yosipovitch, MDa,b Winston-Salem, North Carolina Menthol is a naturally occurring cyclic terpene alcohol of plant origin, which has been used since antiquity for medicinal purposes. Its use in dermatology is ubiquitous, where it is frequently part of topical antipruritic, antiseptic, analgesic, and cooling formulations. Despite its widespread use, it was only recently that the mechanism by which menthol elicits the same cool sensation as low temperature was elucidated upon, with the discovery of the TRPM8 receptor. Although almost 5 years have passed since the discovery of this receptor, many dermatologists are still unaware of menthol’s underlying target. The purpose of this review is to highlight the recent advances in the mechanism of action of menthol and to provide an overview of its dermatologic applications. ( J Am Acad Dermatol 2007;57:873-8.)
• The sensory properties of menthol itself are complex, inducing a menthol’s use since antiquity, the mechanism by enthol (C10H20O; molecular weight, 156) is which it is able to impart a cooling sensationeffects when a naturallyof occurring cyclic terpene alco- aromatic, number cooling, warming, and other sensory
M
hol of plant origin, which gives plants of the Mentha species their distinctive smell and flavor. It was first isolated as a crystalline principle in 1771 by the Dutch botanist Gambius. However, Shimoyama asserted that the peppermint plant, the main source of menthol, has been cultivated for medicinal purposes in Japan for more than 2000 years.1 In the present day, menthol consumption is staggering, and exceeds 7000 tons annually, with a raw product value approaching $300 million.2 Its use is multifold and includes oral hygiene products, confectionary, pharmaceuticals, cosmetics, pesticides, and as a flavoring agent, to name but a few.2,3 With regards to its medicinal purposes, menthol is currently available in both prescribed and over-the-counter
applied topically to the skin or mucous membrane remained a mystery until recently. Although almost 5 years have passed since the discovery of a common receptor for menthol and low temperature, many dermatologists are still unaware of menthol’s underlying target. The purpose of this review is to highlight the recent advances in the mechanism of action of menthol and to provide an overview of its dermatologic applications.
Astringency: “A complex sensation combining three distinct aspects: drying of the mouth, roughing of oral tissues, and puckery or drawing sensations felt in the cheeks and muscles of the face.” MECHANISM OF ACTION Over the past 5 years, understanding of the mechanism by which menthol elicits the same cool sensation as low temperature has advanced a great
Prinsip • Sajikan ke tiap panelis dua sampel, dengan ditanya apakah kedua sampel tersebut sama atau berbeda. • Separuh dari pasangan dengan dua sampel yang berbeda; separuh lagi dengan dua sampel yang sama. • Analisis hasil dengan membandingkan jumlah respon “berbeda” untuk pasangan sampel yang sama dengan jumlah respon “berbeda” untuk pasangan dengan dua sampel berbeda, menggunakan χ2-test.
Sama atau berbeda?
A
B
XXX
XXX
Test Subjects • Umumnya, 20 hingga 50 penyajian untuk tiap empat kombinasi sampel (A/A, B/B, A/B, B/A) diperlukan untuk menentukan perbedaan. • Hingga 200 panelis berbeda dapat digunakan atau 100 panelis yang menerima dua pasang sampel. • Jika Same/Difference test dipilih karena kompleksitas stimuli, maka tidak lebih dari satu pasang yang disajikan pada tiap orang pada satu waktu. • Panelis dapat terlatih atau tidak terlatih, namun tidak boleh campuran antara keduanya.
Kombinasi sampel A : 273 & 892
B : 451 & 635
A
A
A
B
273
892
273
451
B
B
B
A
451
635
635
892
Test Procedure • Siapkan jumlah yang sama untuk tiap empat kombinasi pasangan dan sajikan dengan acak ke panelis, jika setiap orang hanya mengevaluasi satu pasang saja. • Jika pengujian dirancang sehingga setiap panelis mengevaluasi lebih dari satu pasang (satu sama dan satu berbeda atau semua/ keempat kombinasi), maka catatan dari setiap nilai tes panelis harus disimpan.
Example: Replacing a filter for fruit juice • Problem/situation – In an attempt to modernize a beverage manufacturer must replace an old filter used to process fruit juice. The plant manager would like to know if the product produced in the new filter appear the same as that made in the old cooker. • Project objective – To determine if the new filter can be put into service in the plant in place of the old filter. • Test objective – To determine if the two fruit juice products, produced in different filter, can be distinguished by its turbidity (cloudiness).
Example: Fruit juice
(continued)
• Test design – A total of 60 responses, 30 matched and 30 unmatched pairs, are collected from 60 panelists. Each panelist evaluates either a matched pair (A/A or B/B) or an unmatched pair (A/B or B/A) in a single session. The worksheet and the score-sheet for the test are shown in Figures 11 and 12. The test is conducted in the booth area under the light source (D65 at 1,000 lux). • Conduct test – Just before each panelist is to evaluate, add the fruit juice to the transparent glass. Place samples on labeled plates in the order indicated on the worksheet for each panelist.
Date 2-28-91
WORKSHEET
Test code 84-46F09
Post this sheet in the area where trays are prepared. Code scoresheets ahead of time. Label serving containers ahead of time. Type of sample :
Barbecue Fruit juicesauce on white bread pieces
Type of test :
Simple difference test
Sample identification 5-117-36 (old (old cooker) filter) (new cooker) filter) 5-117-39 (new
Code 36 39
Code serving containers with 3-digit random numbers and divide into two lots, one lot to receive sample 36, the other sample 39. When preparing panelists’ trays, place samples from left to right in the following order: Panelist Code
Sample Order
1 – 15
36 – 36
16 – 30
36 – 39
31 – 45
39 – 36
46 – 60
39 – 39 Figure 11. Worksheet for Simple Difference test
Test No. 84-4639
Simple Difference Test Taster No. _________
Name ________________
Type of Sample
Fruit juicesauce on white bread pieces Barbecue
Date _______
Instructions of the samples from left to right. 1. Evaluate Taste the the twoturbidity samples(cloudiness) from left to right. 2. Determine if samples are the same/identical of different. 3. Mark your response below. Note that some of the sets consist of two identical samples.
_____________
Products are the same
_____________
Products are different
Comments: _____________________________________________________ _______________________________________________________________ Figure 12. Scoresheet for Simple Difference test
Example: Fruit juice •
(continued)
Analyze results – In the table below, the columns indicate the samples which were tested, the rows indicate how they were identified by the panelists:
Panelists said:
Panelists received Total
Matched pair AA or BB
Unmatched pair AB or BA
Same
17
9
26
Different
13
21
34
Total
30
30
60
Example: Fruit juice
(continued)
• Analyze results (continued) – The χ2-analysis is used to compare the placebo effect (17/13) with the treatment effect (9/21). The χ2statistic is calculated as:
(O – E)2 χ2 = Σ E where O is the observed number and E is the expected number, in each of the four boxes same/matched, same/unmatched, different/ matched, and different/unmatched. E = (26 x 30) / 60 = 13 E = (34 x 30) / 60 = 17
χ2 =
(17 – 13)2 13
+
(9 – 13)2 13
+
(13 – 17)2 17
+
(21 – 17)2 17
= 4.34
Example: Fruit juice
(continued)
• Interpret results – 4,34 which greater than the value in Table T5* or Table 17.5** (df = 1, probability = 0.05, χ2 = 3.84), i.e., a significant difference exists. The results show a significant difference between the fruit juice prepared in the two different filters. Ø The sensory analyst informs the plant manager that the equipment supplier’s claim is not true. A difference has been detected between the two products. Ø The analyst suggests that if the substitution of the new filter remains an important cost/efficiency item in the plant, the two fruit juice should be tested for preference among users. A consumer test resulting in parity for the two juices or in preference for the juices from the new filter would permit the plant to implement the process.
* Table T5, Upper-α Probability Points of χ2-Distribution; from Meilgaard M, Civille GV, Carr BT. 1991. Sensory Evaluation Techniques, 2nd ed. CRC Press, Boca Raton, FL, page 330. ** Table 17.5, Upper-α Probability Points of χ2-Distribution; from Meilgaard M, Civille GV, Carr BT. 2007. Sensory Evaluation Techniques, 4th ed. CRC Press, Boca Raton, FL, page 426.
“A” – “Not A” Test • Aplikasi − Membandingkan sampel dengan flavor yang kuat dan tak mudah hilang − Sampel mempunyai sedikit perbedaan dalam kenampakan − Sampel memiliki stimuli yang kompleks dan membingungkan panelis − Tes ini juga berguna untuk menseleksi panelis, misalnya, menentukan apakah seorang atau kelompok panelis mengenali pemanis tertentu relatif terhadap pemanis lainnya, dan dapat juga digunakan untuk menentukan thresholds dengan Signal Detection method.
“A” – “Not A” Test • Effective in situations − Same as Same/Difference test − Use the “A” – “Not A” test in preference to the Same/Difference test when one of the two product, is familiar to the panelists, or is essential to the project as the current sample againts which all others are measured.
Prinsip • Kenalkan panelis dengan sampel “A” dan “not A”. • Sajikan tiap panelis dengan sampel, sebagian adalah produk “A” dan yang lainnya produk “not A”; untuk tiap sampel panelis menilai apakah merupakan “A” atau “not A”. • Tentukan kemampuan panelis untuk membedakan dengan membandingkan identifikasi yang benar dengan yang salah menggunakan χ2-test.
Test Subjects • Latih 10 hingga 50 panelis untuk mengenali sampel “A” dan yang “not A”. • Gunakan 20 hingga 50 penyajian untuk setiap sampel dalam pengujian. Setiap panelis menerima hanya satu sampel (“A” atau “not A”), dua sampel (satu “A” dan satu “not A”), atau setiap panelis dapat menguji hingga sepuluh sampel dalam satu seri. • Jumlah sampel yang bisa disajikan ditentukan dari derajat kelelahan fisik dan/atau mental yang mungkin dihasilkannya terhadap panelis. • Catatan: Varian dari metode ini, dimana panelis tidak dikenalkan dengan sampel “not A”, tidak direkomendasikan. Hal ini karena panelis kurang memiliki kerangka referensi, dapat menerka dengan luas dan memberikan hasil yang bias.
Test Procedure • Beri kode semua sampel dengan angka acak dan sajikan dengan acak sehingga panelis tidak dapat mendeteksi pola “A” dan “not A” pada tiap seri sampel. Jangan ungkap identitas sampel hingga panelis menyelesaikan semua seri.
Contoh kasus: Pemanis baru dibandingkan dengan sukrosa • Problem/situation – Pengujian sebelumnya menunjukkan 0,1% pemanis sejajar dengan 5% sukrosa, namun pengujian lebih dari 1 sampel bersamaan menunjukkan adanya carryover rasa manis dan faktor mouthfeel. • Project objective – Mengetahui apakah alternatif pemanis 0,1% dapat digunakan menggantikan sukrosa 5%. • Test objective – Membandingkan kedua pemanis secara langsung dengan mengurangi pengaruh carryover dan kelelahan.
Date 1-15-91
WORKSHEET
Test code 612A83
Post this sheet in the area where trays are prepared. Code scoresheets ahead of time. Label serving containers ahead of time. Type of sample : Type of test :
Sweetened beverage “A” – “Not A” test
Sample identification Beverage with 0.1% sweetener (“A”) Beverage with 5% sucrose (“Not A”)
Code A B
Code 200 6-oz cups with random 3-digit numbers and divide into two lots of 100 each. Use sample “A” for the first 100 cups and sample “Not A” for the second 100 cups. When preparing panelists’ trays, place samples from left to right in the following order: Panelist Code
Sample Order
1–5
A A B B A B A B B A
6 – 10
B A B A A B A A B B
11 – 15
A B A B B A B B A A
16 – 20
B B A A B A B A A B Figure 13. Worksheet for “A” – “Not A” test
“A” – “Not A” Test Name ___________________________
Date ________________
Type of Sample _________________________________________ Instructions 1. Before taking this test, familiarize yourself with the flavor of the sample A” and “Not A” which are available from the attendant. 2. Taste the test samples from left to right. After each sample, record your response below, rinse your palate with water, and wait one full minute between samples. Note: You have received approximately equal numbers of “A” and “Not A” samples.
Sample No. Code
The sample is: “A” “Not A”
Sample No. Code
The sample is: “A” “Not A”
1
_______
o
o
6
_______
o
o
2
_______
o
o
7
_______
o
o
3
_______
o
o
8
_______
o
o
4
_______
o
o
9
_______
o
o
5
_______
o
o
10
_______
o
o
Comments: ____________________________________________________________
Example of scoresheet for “A” – “Not A” test
Example: Sweetened beverage •
(continued)
Analyze results – In the table below, the columns indicate the samples which were tested, the rows indicate how they were identified by the panelists:
Panelists said:
Panelists received Total ‘A’
‘Not A’
‘A’
60
50
110
‘Not A’
40
50
90
Total
100
100
200
Example: Sweetened beverage (continued) • Analyze results (continued) – The χ2-analysis is used to compare the placebo effect (60/40) with the treatment effect (35/65). The χ2statistic is calculated as:
(O – E)2 χ2 = Σ E where O is the observed number and E is the expected number, in each of the four boxes same/matched, same/unmatched, different/ matched, and different/unmatched. E = (110 x 100) / 200 = 55 E = (90 x 100) / 200 = 45
χ2 =
(60 – 55)2 55
+
(50 – 55)2 55
+
(40 – 45)2 45
+
(50 – 45)2 45
= 2.02
Example: Sweetened beverage (continued)
• Interpret results – 2.02 which smaller than the value in Table T5* or Table 17.5** (df = 1, probability = 0.05, χ2 = 3.84), i.e., a significant difference not exists. The results indicate that the 0.1% sweetener solution is not significantly different from the 5% sucrose solution. The sensory analyst informs the development chemist that the particular sweetener is likely not to cause a detectable change in flavor of the beverage.
* Table T5, Upper-α Probability Points of χ2-Distribution; from Meilgaard M, Civille GV, Carr BT. 1991. Sensory Evaluation Techniques, 2nd ed. CRC Press, Boca Raton, FL, page 330. ** Table 17.5, Upper-α Probability Points of χ2-Distribution; from Meilgaard M, Civille GV, Carr BT. 2007. Sensory Evaluation Techniques, 4th ed. CRC Press, Boca Raton, FL, page 426.
Two-Out-Of-Five Test • Scope and application − Untuk menentukan apakah perbedaan produk dihasilkan dari perubahan bahan, pengolahan, pengemasan, atau penyimpanan − Untuk menentukan apakah terdapat perbedaan menyeluruh, dimana tidak ada atribut spesifik teridentifikasi yang terpengaruh − Untuk memilih dan memonitor panelis terhadap kemampuannya untuk membedakan
Prinsip • Sajikan ke tiap panelis lima sampel berkode. • Beritahukan ke panelis bahwa dua sampel sama dan tiga lainnya dalam kelompok yang berbeda. • Perintahkan kepada panelis untuk mencecap sampel dari kiri ke kanan dan pilih sampel yang berbeda. • Hitung jumlah jawaban yang benar dan lihat Tabel T10* atau 17.14** untuk interpretasi.
* Table T10, Two-out-of-Five Test for Difference: Critical Number (Minimum) of Correct Answer; from Meilgaard M, Civille GV, Carr BT. 1991. Sensory Evaluation Techniques, 2nd ed. CRC Press, Boca Raton, FL, page 341. ** Table 17.14, Critical Number of Correct Responses in Two-out-of-Five Test; from Meilgaard M, Civille GV, Carr BT. 2007. Sensory Evaluation Techniques, 4th ed. CRC Press, Boca Raton, FL, page 439.
Test Subjects • Umumnya 10 hingga 20 panelis digunakan untuk Twoout-of-Five test, walaupun 5 sampai 6 juga dapat digunakan ketika perbedaan itu cukup besar dan mudah diketahui. • Minimal panelis sudah terbiasa dengan Triangle test (format, yang harus dilakukan, prosedur evaluasi), dan dengan produk yang diujikan. • Direkomendasikan untuk melakukan sesi orientasi sebelum pelaksanaan uji yang sebenarnya guna mengenalkan panelis dengan prosedur tes dan karakteristik produk. • Gunakan hanya panelis terlatih.
Test Procedure • Jika jumlah panelis bukan 20, pilih secara acak kombinasi yang mungkin seperti dibawah, dengan mempertimbangkan jumlah yang sama antara 3 As dan 3 Bs: AAABB AABAB ABAAB BAAAB AABBA
ABABA BAABA ABBAA BABAA BBAAA
BBBAA BBABA BABBA ABBBA BBAAB
BABAB ABBAB BAABB ABABB AABBB
Two-Out-Of-Five Test Name ___________________________
Date ________________
Type of Sample _________________________________________ Instructions 1. Examine the samples in the order listed below. Two are of one type and the other three of another. 2. Identity the two samples that feel the same by placing an X in the corresponding boxes. Codes
X
Comments
_____________
o
__________________________________
_____________
o
__________________________________
_____________
o
__________________________________
_____________
o
__________________________________
_____________
o
__________________________________
Example of scoresheet for Two-out-of-five test
Example: Liquid foundation makeup • Samples – Portions of the samples were dispensed into 50-mm “wa-
ter white” plastic Petri dishes, which were completely filled and sealed around the edge with cling film. This gave a sample depth of 8 mm. Five Petri dishes were filled from each sample to give a pair for the triangle test and a triplet for the two-out-of-five test. Colour measurement made on the sealed samples (using a Labex spectrophotometer) indicated that the colour difference of the sample pairs was unchanged by enclosure in the Petri dish. • Panelists – To obtain data relevant to the general population, panels
were chosen to be untrained in colour matching. It should be noted that a number of the previous studies have used observers with experience in the field of colour matching. Seventy-four panelists, who were recruited from staff and students at Auckland University of Technology, completed the sensory tests. Ages ranged from 16 to 54 with equal numbers of males and females. Whiting R, Murray S, Caintic Z, Ellison K. The use of sensory difference tests to investigate perceptible colour-difference in a cosmetic product. Color Res Appl. 2004, 29(4): 299−304.
Example: Liquid foundation makeup
(continued)
• Viewing conditions – Samples were viewed in a sensory booth with
gray (L* = 78.1, a*= 2.8, b* = -2.8) walls and base. Illumination was D65 at 1000 lux. The viewing distance was 60 cm and each sample subtended an angle of 6°. Samples were presented to the panelists on white plastic trays with gray (L* = 62.4, a* = 3.4, b* = 3.1) nonwoven fabric over the bottom of the tray. The samples were placed in a line touching each other (Fig. 1) and attached to the tray using Velcro. The arrangement of the samples on the trays is shown in Fig. 1. Panelists were given a response form that listed each set of samples in the same sequence as they appeared on the tray.
and CIE1931 colourlight-tight container anslucence. Ten samE(CMC1:1)] between pair (pair X) with e samples ranged for , and for b* from 16
sed into 50-mm “wawere completely filled ng film. This gave a hes were filled from ngle test and a triplet easurement made on ectrophotometer) inhe sample pairs was ish.
ruited from staff and chnology, completed 16 to 54 with equal FIG. 1. The arrangement of samples for sensory tests.
Example: Liquid foundation makeup
(continued)
TABLE 1. Sensory results for sample pairs.
•
mum in the correlation tried in the varied param The number of correct was found. choices of 74 For each colour– diff Two-out-of-five ters that gave the best c Triangle test test Pair !E(CMC1:1) then investigated furthe 1 0.2680 34 4 sure that there were no 2 0.3772 33 12 regression plot and the 3 0.4379 26 7 spected to find the poin 4 0.5674 42 25 5 0.7033 41 14 that there was a 95% co 6 0.9077 55 15 effect being measured. 7 0.7972 48 21 Statistically, with 74 8 1.0424 61 29 9 0.9765 40 33 effect is being observed 10 1.6165 53 35 correct choices for the tr the two-out-of-five test. plots, these points (34 Conclusion The was triangle and the two-out-of-five cancolour be used three –dishes showntest as the individual numbers on the testvisual difference a for testing perception colour difference. two-out-of-five test could pe dishes making upof that set. The order (left toThe right) of the some observers dishes on thecorrelation tray was thethan samethe as triangle that shown on The the speed at which gave much better test. response The panelists circled, on the response form, experiment, 22 the test can be form. performed would recommend (the present RESULT the number that appeared on the dish that was the single in colour pairs were assessed in 15 min with no reported cases of panelist fatigue) that set. On completing one set of three dishes, the panelist Sensory Results was presented with another set until each sample pair had
