Two-pan and operant lever-press tests of dogs' preference for various foods

The traditional two-pan choice test and an operant lever-press choice test were used to assess dogs' preferences between various foods. In the first two experiments the foods differed in flavor as manipulated by the amount of fat-coating added to a dry-food base. In the third experiment, the foods were commercially-available dry and semi-moist dog foods. The experiments illustrate that preference data obtained with these two test methodologies do not always point to the same conclusion, and that complexities are involved in predicting the outcome of a preference test from performance on previous tests.