CATA penalty/reward

A method of analyzing check-all-that-apply (CATA) data similar to penalty analysis is given. The approach gives a mean drop/increase relative to a “reference” variable such as overall liking. The method is demonstrated on “emotion” responses for five citrus flavored sodas. Six significance testing methods are compared. The effects of sample size, a CATA proportion threshold and covariance thresholds on significance are investigated in an appendix via simulations.

► CATA responses can be analyzed similar to JAR variables in penalty analysis. ► Six significance testing approaches for CATA penalty/rewards were compared. ► Simulations show increasing test sensitivity with increasing sample size. ► CATA proportions can be unreliable in determining “significance thresholds”. ► Negative emotions can have notable impact on liking despite infrequent use.