This study compared generating information about food to viewing and verbally producing that information for learning and transfer. Specifically, during training, a list of food items was shown, and some participants generated either real calories (Experiment 1) or fabricated prices (Experiment 2) associated with those foods with feedback provided (seeding), some viewed and read aloud values provided (viewing), and others read aloud the names of the items with no values provided (control). All participants were tested immediately and 1 week later on both old and new foods. In both experiments, the seeding and viewing conditions showed more learning than the control condition, and only the experimental conditions showed transfer of learning from old to new foods, with advantages found for seeding over viewing only in some cases. The results are discussed in terms of the seeding, testing, generation, and production effects, as well as their applications to health and policy.