The A–not-A test, is a method of discrimination testing comprised of at least two samples; at least one sample is a previously identified sample (“A”) and at least one is a test sample. All samples are presented blindly, and the assessor’s task is to assign the label “A” or “not-A” to each of the samples. The following answer choices are available within the A-not-A Test:

  • Different, I'm sure of it
  • Different, but I'm not sure of it
  • I don't know, but guess it's different
  • I don't know, but guess it's the same
  • The same, but I'm not sure of it
  • The same, I'm sure of it


NOTES:

  • A "Same" response will be "Correct" if the Tracking Code of the current sample matches the "Reference" sample. 
  • A "Different" response will be "Correct" if the Tracking Code of the current sample does not match the Reference" samples.
  • The A-not-A question will be automatically hidden if the current sample being evaluated is the "Reference" sample.
  • Tracking Codes can be updated/added any time to (re)calculate results.
  • The "Reference" sample in the question's configuration options can be updated any time to (re)calculate results.
  • The "Reference" sample does not need to be included in the questionnaire serving or design block. If Reference is a serving in the block, it should come before all the other samples in the A-not-A test. Currently there is no design block validation for this.
  • Correct/Incorrect Logic Triggers are not currently supported.
  • Analysis>View/Edit data is not currently supported. This is consistent with all discrim methods.




Test Setup


Create Samples:

  1. The first sample will be your Reference sample. Input a unique tracking code - e.g. Current.



  2. Create another sample. This should be the same as the Reference Sample. Match the tracking code for the first sample - Current.



  3. Create another unique sample. This should be a Test Sample. Give the sample a unique sample code. E.g., Test.



Create Questionnaire:

  1. Create a Single Sample Questionnaire.



  2. Attach samples to the questionnaire.



  3. Add questions to the questionnaire. Be sure to include the A Not A Discrimination question from the Basic Library > Discrimination Questions area.



  4. Under the question options area, select the Reference Sample. Here we choose "Control". This question will NOT show for the Reference Sample selected.



  5. Add any additional questions. E.g., Overall Opinion, JAR Scales, etc. Data will be analyzed as in any other Single Sample questionnaire setup. Generally, inserting a new page first will streamline your questionnaire.



Finalize Test and Launch:

  1. Create your Design Block. The general recommendation is to drag the Reference Sample identified in Step 4 into the "First" area in Design Block.
     

  2. Launch your test and collect data.



View Results

  1. A Not A results will update continuously throughout testing. Percent Correct, Confidence Level, P Value / Alpha, Beta, Power, D', Standard Error, Upper and Lower Confidence Levels, P(c) and P(A) (R-Index).



  2. Additional questions will update continuously throughout testing as in all Single Sample questionnaires.