Answers to Test Yourself
(1) d. (2) c. (3) b. (4) c. (5) Manipulation of an independent variable and control of extraneous variables. (6) The Latin square allows fewer orders to be used than a full counterbalancing--useful for designs with more than three conditions. The number of orders in a Latin square matches the number of conditions in the experiment. (7a) The independent variable is the “treatment” for separation anxiety with levels of chew toy, our used objects, and confinement. (7b) Within subjects. (7c) Each treatment was used five times to control for length of time away and other possible variables that might change with our times away from home; the treatment orders were counterbalanced to control for order effects. (7d) This experiment would likely be done by choosing a sample of dogs that represent the population (see Chapter 6 for information about sampling techniques) and randomly assigning a portion of the sample to each of the treatments (e.g., 25 dogs to each treatment group). Confounding variables as described for the scenarios in the chapter are controlled for each treatment group (e.g., test all dogs for same length of visit away in same environment and at same time of day). This experiment uses a between-subjects design. The dependent variable is measured in the same manner as the chapter scenario with the operational definition specified. (8a) The independent variable is smile group: smile or no smile--it is manipulated between-subjects because the subjects only complete the experiment in one of the two groups. (8b) The main ethical issue in this study is that deception is used (the subjects are misled about what the study is actually about and what behaviors are being observed). The researchers need to justify why the deception is necessary and thoroughly debrief the subjects at the end of the experiment so they understand why they were deceived and what the real purpose of the study was.