Do individual differences in visual attention to CATA questions affect sensory product characterization? A case study with plain crackers