Material
- T-shaped arena
- 4 colored boxes of different colors. Blue, yellow, black and white
- 2 short walls, used to block the view in phase 2
Method
9 dogs were trained to associate colors with a reward in 3 phases before being assessed in the 4th phase. The rewarded color was randomised so half the dogs started with blue and the other half started with yellow. The sides of the boxes were also psudorandomised, each session consisted of 10 trials where 5 trials had the rewarded box on the right and 5 trials had the box on the left.
Phase 1
The dog learned to associate either the blue- or yellow box with a reward at the end of a T-shaped arena. This was done in order for the dog to learn that if it approched this box it gained a reward. See figure 1.
Phase 2
Two short walls were placed on each side of the dividing wall in the arena, and the black and white boxes were positioned in front of them. Behind the walls, the blue and yellow boxes were hidden from view. The idea was that the dog would first see the black or white box, and only after passing it would the blue or yellow box be revealed. This setup helped the dog learn a two-step association: black leads to blue, and white leads to yellow. See figure 1.
Phase 3
The dog went through a reversal learning phase where the rewards linked to blue and yellow were swapped. The color that was previously rewarded no longer gave a treat, and the unrewarded color became the new rewarded one. During this phase, the black and white boxes were not shown. The goal was to change how the dog valued black and white by changing the reward linked to blue and yellow, without showing black or white during this learning phase. See figure 2.
Phase 4
In the final phase the short walls were added back and the dog was shown only the black and white boxes. When the dog approached one, it would find a gray box behind it with no reward. This happened for 30 trials and gra+no reward was done in oder for the dog not to learn anything in this phase. The goal was to see if the dog understood the connection between the boxes based on what it had learned earlier. In Phase 2, white was always placed in front of the rewarded yellow box, and black in front of the unrewarded blue. In Phase 3, the rewards were switched, yellow no longer gave a treat, but blue did. If the dog understood this change, it might realize that white now leads to no reward, and black leads to a reward (in theory). By observing whether the dog stuck with its association from phase 2 or changed based on the information from phase 3, I could assess if the dog was using causal reasoning. See figure 2.
