Sensory Computer Systems Forums Index Sensory Computer Systems Forums
Welcome to the Sensory Computer Systems Online Forums!
 
 FAQFAQ   SearchSearch   RegisterRegister 
 ProfileProfile   Log inLog in   
   Sensory Computer Systems Home   

3-AFC (Alternative Forced Choice)

 
Post new topic   Reply to topic    Sensory Computer Systems Forums Index -> Test Design
View previous topic :: View next topic  
Author Message
Jane



Joined: Dec 21 2005
Posts: 5

PostPosted: Fri Jun 09, 2006 11:00 am    Post subject: 3-AFC (Alternative Forced Choice) Reply with quote


How do I set up a 3-AFC (Alternative Forced Choice) Test in SIMS ?

Thanks,
Mary

Back to top
View user's profile
John



Joined: Jun 13 2005
Posts: 23

PostPosted: Fri Jun 09, 2006 11:14 am    Post subject: 3-AFC (Alternative Forced Choice) Reply with quote



For the THRESHOLD Testing discussion please skip to next message posting below.




3-AFC (Alternative Forced Choice)

2 sample test, 3 presented

Yes, this design is possible in your SIMS software. From our experience and recent Internet research on the AFC topic, it is evident that this type of Sensory test is not commonly used. And when used, AFC tests seem to be in a variety of strategies and tests methods and question designs. Thorough understanding of desired test objectives is important for selecting proper AFC design. Terms like stimuli, individual panelist thresholds, panelist calibration, series, concentrations, steps, ratios, etc are unique to AFC studies. It is complex.

The best reference document we found regarding AFC tests is ASTM E1432-04(2004) titled Standard Practice for Defining and Calculating Individual and Group Sensory Thresholds from Forced-Choice Data Sets of Intermediate Size. Notably Paragraph 4.1.

Whatever AFC method is being considered, individual panelists are a focus of AFC, resulting in this key sentence from ASTM document: "There is a large degree of random error associated with estimating the probability of detection from less than approximately 500 3-AFC presentations."

Sensory Triangle Tests A A B B are much more common as are Duo-Trio Tests.


3-AFC alternative forced choice (AFC) discrimination test where panelists are presented with 3 products and asked to select one product based on a pre-specified attribute. The3-AFC test is used to establish if there is a discernible difference between 2 products in respect of a specified attribute. The panelist is required to select 1 product from the set of 3 that differences in the specified attribute.

The guessing probability (probability of getting a correct answer by guessing only) is ⅓.


SIMS Basic Design Notes for 3-AFC

For setup in your SIMS software, 3-AFC are simply a "2 sample" Discrimination Test, where Panelists see 3 presented, similar to a typical triangle test A A B B.

Sample Presentation Strategy:
Typical Triangle: A A B B , where A or B is shown Once/Twice, and either can be the Unique Sample.

3-AFC (Alternative Forced Choice): A B B , where A is always shown once and is the Unique Sample.
(the presentation order can change, such as A B B, B A B, B B A)

One terminology saw referred to AFC as a Triangle Constant, vs a typical Triangle Rotated.

SIMS Program Steps:

1. Create your new Discrimination Ballot. Add Triangle Attribute question.

2. Create your new Experimental Design.
2a. Specify TWICE the number of Panelists you'll actually need.
2b. Select Experimental Plan: Discrimination Triangle (DTRI)

3. Create your SIMS software Test Definition.
3a. View Rotation Plan for acceptability.
Note that it is OK that Rotation Plan includes all Sample Sets of a typical Triangle Test Design A A B B. For your AFC Test, you're only going to use the Sample Sets sets needed for AFC Design A B B. This is why we specified TWICE the Panelists in 2a. above.

4. Create your new Result Master Test Execution.
4a. Select Sample Set Selection: Force Panelist to Register Sample Set as Ballot.

5. Execute your test. Only use the Rotation Plan's Sample Set #s which meet you AFC Design A B B. You'll need to indicate on your panelists trays, instructions, laminated cards, etc. what the Sample Set # is. The Panelist is going to Register Sample Set as Ballot at start of test.

This process is closely identical to a typical Triangle test, except for the actual intentional presentation of Rotation Plan Sample Sets A B B.


3-AFC Test Result Interpretation

Sensory Computer Systems can not assist in your Interpretation of the data. Interpretation is really about your objectives, panelists, methods, risk tolerance, culture. All of the SIMS standard Discrimination reports are available in your SIMS software, and most importantly in this 3-AFC case, full Test Data Export support to Microsoft Excel, SAS, JMP, SPSS, etc. Individual Panelist ratios are important to 3-AFC data results interpretation, each and every Threshold Concentration Level for Each Panelist.


Data tabulation/reporting:
SIMS Raw Data Export Report, use report column option [X] If Caption is BC then display the Absolute Sample # selected.
- This makes it quick-n-easy to see if your panelist choose the blank (S1) or the concentrate (any S#>1).
And use report option (*) Sort When Multiple Rep Tests Sort by Panelist, Rep.
- This shows you for each panelist when the test sample is correctly identified, and when they continue to choose correctly in higher concentration test sample sets.

Quote from ASTM E679: "10.1 Because sensory threshold values are functions of sample presentation variables and of individual sensitivities, interlaboratory tests cannot be interpreted statistically…"
SIMS does not do the log calculations mentioned in E679.


Final note about AFC individual panelist thresholds and panelist calibration.
An alternate way to think about the design in SIMS software is to focus on the above sentence "Individual Panelist ratios are important to 3-AFC data results interpretation". If this is the case with your test objectives, then consider setting up a separate SIMS Test for each panelist exclusively, the experimental design for Anonymous Panelists, have each panelist take all Anonymous Sample Sets. For the responses for the panelist, calculate the % ratio correct, aka the threshold. Repeat the process for each of your desired Panelists(n). This is where the 500+ presentations mentioned in the ASTM document can quickly occur. In the end, this will give you a set of (n) ratios (per concentration), which can then be interpreted thru guidelines in the ASTM Documents.


Supportively yours,

John Ream
Sensory Computer Systems
New Jersey 07922 USA
908-665-6464
Back to top
View user's profile
ForumAdmin
Site Admin


Joined: May 20 2005
Posts: 120
Location: New Jersey

PostPosted: Tue Jun 16, 2015 3:04 pm    Post subject: Threshold Testing Reply with quote


Threshold Testing

Threshold testing in SIMS.

Threshold testing is by a forced-choice series of triangle presentations, such as by ascending formula concentrations. Basically it is a 3-AFC but instead of a randomized order, the samples are seen in order of concentration (min of 10 concentrations). When a panelist identifies the “test” sample as sweeter 2x in a row the test is ended. The data is analyzed using a geometric mean. This is most like the example in the ASTM doc appendix (X1) Differences threshold of Ethyl Acetate Added to Beer

11 C3 R10 (11 samples, 3 presented, 10 reps.) Where Sample 1 is “Blank” and next 10 samples are concentrations in rising order.

A series of Affective Hedonics 3-point (like triangles), 10 reps, 1,1,2 ... 1,1,3 ... 1,1,4 ... ... 1,1,10 ... 1,1,11
2 blanks and 1 concentrate is seen with each triangle presentation.


Ballot: Affective ballot, 1 attribute: Hedonic 3-point, data export label 'Triangle'.
<RELATIVE_SAMPLE_1_BLINDING_CODE> <RELATIVE_SAMPLE_2_BLINDING_CODE> <RELATIVE_SAMPLE_3_BLINDING_CODE>


Experimental Design: 11 samples, 3 presented, 10 reps. Select Randomized Sample Presentation to ensure order balance.
Sample #1 = Control
Sample #2 = Concentrate #1
Sample #3 = Concentrate #2
...
Sample #11 = Concentrate #10


From ASTM E679, the standard procedure for panelists:
7.1 The panelist begins judging with that set which contains the test sample with the lowest concentration (highest dilution)
of the odorous or sapid substance, takes the time needed to make a selection, and proceeds systematically toward the higher concentrations.
7.2 Within each set, the panelist indicates that sample which is different from the two others (detection threshold)
or which exhibits a recognizable odor or taste of the substance (recognition threshold).
If the panelist cannot readily discriminate, a guess must be made so that all data may be utilized.
7.3 The judgments are completed when the panelist either
(1) completes the evaluation of all sets of the scale, or
(2) reaches a set wherein the test sample is correctly identified, then continues to choose correctly in higher concentration test sample sets.


Data tabulation/reporting:
SIMS Raw Data Export Report, use report column option [X] If Caption is BC then display the Absolute Sample # selected.
- This makes it quick-n-easy to see if your panelist choose the blank (S1) or the concentrate (any S#>1).
And use report option (*) Sort When Multiple Rep Tests Sort by Panelist, Rep.
- This shows you for each panelist when the test sample is correctly identified, and when they continue to choose correctly in higher concentration test sample sets.

Quote from ASTM E679: "10.1 Because sensory threshold values are functions of sample presentation variables and of individual sensitivities, interlaboratory tests cannot be interpreted statistically…"
SIMS does not do the log calculations mentioned in E679.


A SIMS Express Package of an example Threshold Test is available. THRESHOLD.SIMS2000ExpressPackage.sql.ZIP

ASTM Reference:
See ASTM E679 "Standard Practice for Determination of Odor and Taste Thresholds By a Forced-Choice Ascending Concentration Series Method of Limits"




Last edited by ForumAdmin on Mon Apr 08, 2019 9:21 am; edited 10 times in total

Back to top
View user's profile
ForumAdmin
Site Admin


Joined: May 20 2005
Posts: 120
Location: New Jersey

PostPosted: Wed Dec 16, 2015 12:18 pm    Post subject: Threshold Testing Reply with quote


Threshold Testing



SIMS software express package of the following example: THRESHOLD.SIMS2000ExpressPackage.sql.ZIP



Questionnaire Example:




Setup Questionnaire Example:




Test Definition Example:







Back to top
View user's profile
Display posts from previous:   
Post new topic   Reply to topic    Sensory Computer Systems Forums Index -> Test Design All times are GMT - 5 Hours
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum


Powered by phpBB © 2001, 2005 phpBB Group