1、October 2016 English price group 14No part of this translation may be reproduced without prior permission ofDIN Deutsches Institut fr Normung e. V., Berlin. Beuth Verlag GmbH, 10772 Berlin, Germany,has the exclusive right of sale for German Standards (DIN-Normen).ICS 67.240!%n 1/2. 3.5 two-sided tes
2、t test in which the test supervisor does not have any a priori knowledge concerning the direction of difference NOTE The null hypothesis is H0, the products are not different; the proportion of responses observed for one of the samples, p, is equal to 1/2. The alternative hypothesis is H1, p 1/2. 3.
3、6 correct responses expected responses number of assessors, in the case of a one-sided test, having selected the sample expected by the test supervisor to be the most intense for the sensory attribute under study 3.8 product material to be evaluated 3.9 sample unit of product prepared, presented and
4、 evaluated during the course of the test 3.7 consensual responses highest value, in the case of a two-sided test, of the number of assessors having selected sample A and those having selected sample B NOTE This is calculated as above since there are not any correct responses. DIN EN ISO 5495:2016-10
5、 EN ISO 5495:2007 + A1:2016 (E) 7!“3.10 sensitivity general term employed to summarize the performance characteristics of the test NOTE In statistical terms, the sensitivity of the test is defined by the values of , and pd. 3.11 similarity situation in which any perceptible differences between the s
6、amples are so small that the products can be used interchangeably 4 Principle The number of assessors is chosen on the basis of the sensitivity desired for the test (See 6.2 and the footnote that accompanies Tables A.4 and A.5). The assessors receive a set of two samples (i.e. a pair). They designat
7、e the sample which they consider to be the most intense regarding the sensory attribute under consideration, even if this choice is based only on a guess. NOTE One of the samples may be a control. The number of times that each sample is selected is counted and the significance is determined by refer
8、ence to a statistical table, taking into consideration the results obtained for the expected sample (one-sided test) or the highest number of responses obtained for either of the samples (two-sided test). 5 General test conditions 5.1 Define the objective of the test in a clear way to determine if t
9、he attempt is to be a one-sided or a two-sided test, if it is a difference or similarity test, and which is the most appropriate sensitivity. 5.2 Carry out the test under conditions that prevent all communication among assessors until the evaluations have been completed, using facilities and booths
10、complying with ISO 8589. 5.3 Prepare the samples out of sight of the assessors and in an identical manner for each one of them; i.e. same apparatus, same vessels. 5.4 Assessors shall not be able to draw any conclusions regarding the intensity of the attribute from the manner in which the samples are
11、 presented to them. For example, for a tactile test, any differences in appearance shall be avoided. Mask all colour differences if the test objective does not concern the colour by using light filters and/or subdued lighting. The samples may also be presented successively and non-simultaneously in
12、the case of slight differences in appearance. 5.5 Code the samples or the vessels containing the samples in a uniform manner, preferably using 3-digit numbers chosen at random for each test. Each pair is composed of two samples, each with a different code. Preferably, different codes should be used
13、for each assessor during a session. However, the two same codes may be used for all assessors within a test, provided that each code is used only once per assessor during a test session (e.g. if several paired tests on different products are being conducted during the same session). 5.6 The quantity
14、 or volume served shall be identical for the two samples constituting each pair, just as that of all the other samples in a series of tests on a given type of product. The quantity or volume to be assessed can be imposed. If it is not, it should however be specified to the assessors to take quantiti
15、es or volumes that are always similar whatever the sample. 5.7 The temperature of the samples constituting each pair shall be identical just as that of all the other samples in a series of tests on a given type of product. It is preferable to present the samples at the temperature at which the produ
16、ct is generally consumed. DIN EN ISO 5495:2016-10 EN ISO 5495:2007 + A1:2016 (E) 8 5.8 The assessors shall be told whether or not they have to follow a special protocol in order to assess the products (e.g. whether or not to swallow the samples for a taste test, or carry out a specific gesture for a
17、 tactile test) or whether they are free to do as they please. In this latter case, they should be requested to proceed in the same manner for all the samples. 5.9 During the test sessions, avoid giving information about product identity, expected treatment effects or individual performance until all
18、 tests are completed. 6 Assessors 6.1 Qualification All assessors should possess the same level of qualification, this level being chosen on the basis of the test objective (see ISO 8586-1 and ISO 8586-2). Experience and familiarity with the product can increase the performance of an assessor and ca
19、n consequently increase the likelihood of finding a significant difference. Monitoring the performance of assessors over time may prove to be useful for increased sensitivity. All assessors shall be familiar with the mechanisms of the paired test (the scoresheet, the task and the evaluation procedur
20、e). In addition, assessors shall be capable of recognising the sensory attribute on which the test is based. This attribute shall be defined verbally, by means of a reference substance or by presenting a few samples having different levels of intensity for the attribute under examination. 6.2 Number
21、 of assessors Choose the number of assessors so as to obtain the level of sensitivity required for the test (see Table A.4 for a one-sided test and Table A.5 for a two-sided test). The use of a large number of assessors increases the likelihood of detecting small differences between the products. Ho
22、wever, in practice, the number of assessors is often determined by material conditions (e.g. duration of the experiment, number of available assessors, quantity of product). When conducting a difference test, the number of assessors is typically approximately 24 to 30. When conducting a similarity t
23、est, about twice as many assessors (i.e. approximately 60) are required for equivalent sensitivity. When testing for similarity, evaluations should not be replicated by the same assessors. For a difference test, replications may be considered but should still be avoided whenever possible. However, i
24、f replicate evaluations are required in order to produce a sufficient total number of evaluations, every effort should be made to have each assessor perform the same number of replicate evaluations. For example, if only 10 assessors are available, have each assessor perform three paired tests in ord
25、er to obtain a total of 30 evaluations. NOTE Analysing three evaluations performed by 10 assessors as 30 independent evaluations is not valid when testing for similarity using Table A.3. However, the difference test using Tables A.1 and A.2 is valid even when replicate evaluations are performed 5, 6
26、. Some recent publications 1, 2on replicated discrimination tests suggest alternative approaches for analysing replicated evaluations. 7 Procedure 7.1 Prepare the worksheets and scoresheets (see Figures B.1, B.2 and B.3) prior to conducting the test so as to use an equal number of the two possible p
27、resentation sequences of both products, A and B. 7.2 Present the two samples constituting a pair successively or simultaneously (see 5.4). In the case of simultaneous presentation, arrange the two samples in the same manner for each assessor (in line from left to right, in line from the bottom up, e
28、tc.). The assessors shall examine the two samples constituting the pair in the order indicated in the scoresheet, but assessors are generally authorized to make repeated evaluations of each sample if so wished (if, of course, the nature of the product allows for repeated evaluations). 7.3 Provision
29、should be made for one scoresheet per pair of samples. If an assessor is to perform more than one test during the course of a session, collect the completed scoresheet and the unused samples prior DIN EN ISO 5495:2016-10 EN ISO 5495:2007 + A1:2016 (E) 9 to serving the subsequent pair. The assessor c
30、an neither go back to any of the previous samples, nor modify his/her verdict concerning any of the previous tests. 7.4 Do not ask any questions about preference, acceptance or degree of difference following the selection of the most intense sample. The selection the assessor has just made may bias
31、the response to any additional questions. Responses to such questions may be obtained through separate tests concerning preference, acceptance, degree of difference, etc. (see ISO 6658). A “Comments” section requesting the reasons for the choice may be included for the assessors remarks. 7.5 The pai
32、red test is a “forced choice” procedure; assessors are not allowed to choose the “no difference” option. An assessor who detects no difference between the samples should be instructed to select one of the samples and to indicate that the selection was only a guess in the “Comments” section of the sc
33、oresheet. 8 Analysis and interpretation of results 8.1 When testing for a difference 8.1.1 Case of a one-sided test Use Table A.1 to analyse the data obtained from a paired test. If the number of correct responses is greater than or equal to the number given in Table A.1 (corresponding to the number
34、 of assessors and to the -risk level chosen for the test), conclude that a perceptible difference exists between the samples (see B.1). If desired, calculate a confidence interval on the proportion of the population able to distinguish the samples. This method is described in B.5. 8.1.2 Case of a tw
35、o-sided test Use Table A.2 to analyse the data obtained from a paired test. If the number of consensual responses is greater than or equal to the number given in Table A.2 (corresponding to the number of assessors and to the -risk level chosen for the test), conclude that a perceptible difference ex
36、ists between the samples (see B.3). If desired, calculate a confidence interval on the proportion of the population able to distinguish the samples. This method is described in B.5. 8.2 When testing for similarity 1)8.2.1 Case of a one-sided test Use Table A.3 to analyse the data obtained from a pai
37、red test. If the number of correct responses is less than or equal to the number given in Table A.3 (corresponding to the number of assessors, to the -risk level and to the value of pdchosen for the test), conclude that no meaningful difference exists between the samples (see B.2). If the results ar
38、e to be compared from one test to another, then the same value of pdshould be chosen for all tests. If desired, calculate a confidence interval on the proportion of the population able to distinguish the samples. This method is described in B.5. No conclusion should be drawn for maximum numbers of c
39、orrect responses under n/2. 1)In this International Standard, “similar” does not mean “identical”. This term signifies rather that the two products are sufficiently alike to be used interchangeably. It is impossible to prove that two products are identical. However, it can be demonstrated that any d
40、ifference that does exist between two products is so minor as to have no practical significance.DIN EN ISO 5495:2016-10 EN ISO 5495:2007 + A1:2016 (E) 10 8.2.2 Case of a two-sided test Use Table A.3 to analyse the data obtained from a paired test. If the number of consensual responses is less than o
41、r equal to the number given in Table A.3 (corresponding to the number of assessors, to the -risk level and to the value of pdchosen for the test), conclude that no meaningful difference exists between the samples (see B.4). If the results are to be compared from one test to another, then the same va
42、lue of pdshould be chosen for all tests. If desired, calculate a confidence interval on the proportion of the population able to distinguish the samples. This method is described in B.5. 9 Report Report the test objective, the results and the conclusions. It is recommended to add the following addit
43、ional information: the purpose of the test and the nature of the treatment being studied; the complete identification of the samples: origin, method of preparation, quantity, shape, storage prior to testing, quantity served, temperature (the information concerning the sample should indicate that all
44、 storage, handling and preparation operations have been carried out so as to yield samples that differ only due to the variable of interest, if any); the number of assessors, the number of correct or consensual responses and the result of the statistical evaluation (including the values of , and pdu
45、sed for the test); the assessors: experience (in sensory testing, with the product, with the test samples), age and gender (see ISO 8586-1 and ISO 8586-2); any specific information and recommendations given to the assessors in connection with the test, in particular in the case where a precise defin
46、ition and reference samples illustrating the attribute under test and/or a test protocol have been indicated to the assessors; the test environment: test facility used, simultaneous or sequential presentation, whether the identity of the samples was disclosed after the test and if so, in what manner
47、; the location and date of test, and the name of the panel leader. 10 Precision and bias Because the results of sensory discrimination tests are dependent on individual sensitivities, it is impossible to make a general statement regarding the reproducibility of the results that is applicable to all
48、populations of assessors. The precision regarding a particular population of assessors increases as the size of the panel increases, as well as with the training and exposure to the product. As a “forced-choice” procedure is used, the results obtained by this method are bias-free, provided that the precautions given in Clause 7 are fully observed. DIN EN ISO 5495:2016-10 EN ISO 5495:2007 + A1:2016 (E) 11 Annex A (normative) Tables A.1 Determination of perceptible difference or similarity See Tables A.1 to A.3. Table A.1 Minimum number of
