**B e h a v i o r a l D at a **Customers leave traces of their purchasing behavior in store scanning data, catalog purchases, and customer databases. Marketers learn much by a able these data. Actual purchases can are more rcli- reflect consumers' preferences and often than statements they offer to market researchers. For example, grocery shopping data show that people don't necessarily buy the more expensive brands, co to what they might state in interviews; nd many low-income people buy some expensive brands. Clearly, ANA can learn many useful things about its passengers by analyzing ticket purchase records and online behavior.
Ex p e r i m ent a l R e s e arc h The most scientifically valid research is experi mental research, designed to capture cause-and-effect relationships by eliminating coinpet- ing explanations of the observed findings. If the experiment is well designed and executed research and marketing managers can have confidence in the conclusions. Experiments call for selecting matched groups of subjects, subjecting them to different treatments, controlling extraneous variables, and checking whether observed response dif- ferences are statistically significant. If we can eliminate or control extraneous can relate the observed effects to the variations in the treatments or stimuli. ANA might intro duce in-flight Internet service on one of its regular flights from Chicago to Tokyo and charge one week and $15 the next week. If the plane carried approximately the same number of first-class passengers each week and the particular weeks made no difference, the airline could relate any significant difference in the number of passengers using the service to the different prices charged