A psychotherapist is interested in evaluating the effectiveness of the new therapy she has invented. It is called pet therapy and involves convincing depressed people to keep a dog as a pet, with the hope that caring for a dog will cheer them up. To evaluate the therapy, the therapist gives each of her patients a dog from the Humane Society to care for. She measures depression using a written test(which has been shown to be reliable) a week before they get the pets and then again 2 months later. She discovers that patients are much less depressed when assessed the second time and thus concludes that pet therapy is a success. Discuss several things wrong with this piece of research and the conclusion drawn from it. How are regression artifacts likely to have played a part? How could the research have been done better?