6533b860fe1ef96bd12c2f4c
RESEARCH PRODUCT
Risk tables for discrimination tests
Pascal Schlichsubject
0303 health sciencesNutrition and Dietetics030309 nutrition & dieteticsBinomial test04 agricultural and veterinary sciences[SDV.IDA] Life Sciences [q-bio]/Food engineering040401 food scienceDiscrimination testingTest (assessment)03 medical and health sciences0404 agricultural biotechnology[SDV.IDA]Life Sciences [q-bio]/Food engineeringStatisticsEconometricsTable (database)MacroComputingMilieux_MISCELLANEOUSFood ScienceTriangle testMathematicsType I and type II errorsdescription
Abstract Duo-trio and triangle test are often used in the food industry for the purpose of declaring two products non-distinguishable. In that situation, it is much more important to control the power of the test rather than the type 1 error risk. This paper makes available by e-mail a SAS ® macro, called BINRISKS, for computing type 1 and type 2 risks for any one-tailed binomial test and for any level of the percentage above chance to be detected. Using this macro, two sets of tables have been compiled. The first table includes for any total number of responses below 50, for any number of correct responses and for three levels of the percentage above chance to be detected, the corresponding type 1 and 2 risk values. The second table indicates the minimum total number of responses and the associated number of correct responses required to achieve several combinations of risks of types 1 and 2 for six different percentages above chance.
| year | journal | country | edition | language |
|---|---|---|---|---|
| 1993-01-01 | Food Quality and Preference |