Avians With Rivalry

I'm posting more about the birds and the research we are undertaking into Operant Conditioning and Model/Rival strategies in training. Irene Pepperberg has found the M/R much more effective than classical OC, and I definitely saw that it was effective with at least one of the birds. I originally started the conditioning, hoping to train on recall so that I would feel safer taking the birds outside, that if he flew off by mistake, he would come back. While we're not progressing on that aspect of the training, their enthusiasm has inspired me to see how much more they can learn. We'll be somewhat limited in communication since the birds don't speak English, and my Cockatiel is not exactly fluent. I'm improving, though -- I think!

An unfortunate side effect from the Model/Rival technique and my inexperience with all this:
Since B has been jumping down to compete with C during C's sessions, I have tried to be consistent, and if B touches the indicated target during C's session, I reward him accordingly before bringing him back to the perch. Well, I guess the Model/Rival now works both ways, because now C sometimes jumps down during B's sessions to compete with B back! Chaos. Without thinking, I inadvertently taught both birds to interrupt each other's sessions for rewards! It's cute that they're so excited about it, though, and I'm gratified to know that they're so smart. (The alternative -- that I am so dumb -- is considerably less gratifying.) So now I just need to pick up the interloper as soon as he starts clambering down, without giving him a chance to get the right target, in which case I will be committed to supplying the reward as I have been trained.

[By the way, the above isn't an image of their rivalry but rather the opposite. I took them outside to sit in a tree, which is something that birds like to do. It was such a nice day that C started singing to B. You can see in B's expression how excited and pleased he is to be serenaded...]

I'm still doing flight training every day with Bird B, to practice several things: i) flying from high to low; ii) angling and flying around corners; and iii) flying back to me and Bird C, especially when called (although he pretty much does it whenever he wants, which I figure is also good, as long as he's safe). The flight training seems to be helping. We were practicing outside the other day, where Bird B typically gets "blown upwards" whenever he tries to fly down to us. B was on the roof and I was on the ground with C, urging B to come to us. Suddenly, B launched himself into the air, flapping so hard and fast, fighting frantically to fly down without getting lifted up, and then in seconds, he had made a perfectly abrupt landing right onto my hand! I was so proud of him for that flight! I piled on him much praise and admiration as his little body was heaving from the effort.

Observations: 1 target added since previous session.
7 targets used: pink, clear, green, shiny, rust, purple and orange rocks.
2007/06/30
B1: 65% correct. 17/9 successful repetitions over errors. 5 errors on rust rock. 3 errors on shiny. Target to stick on clear. While target to stick is recorded as an error, I think that strictly speaking, it's the researcher's fault rather than the subject. Researcher must withdraw stick in time so that it is not in the way. 1039-1044h.

C1: 89% correct. 25/3. One error each on purple and clear. Target to stick on rust. 1044-1051h.

B2: 58% correct. 18/13. No errors on green. Extremely poor performance. B's performance deteriorating today with repeated sessions. 1052-1059h.

C2: 76% correct. 22/7. Many errors are target to stick. No errors on rust or orange. B jumps down to join C toward end of session. 1100-1108h.

C3: 79% correct. 19/5. Three errors on orange. One error each on clear and purple. No errors on pink, rust, shiny, green. 1948-1953h.

B3: 82% correct. 9/2. One error each on orange and shiny. Session interrupted by telecommunications. B eager to continue his training session and insisted on finishing training before researcher should finish telecommunications session. 1954h.

B4: 73% correct. 16/6. Three errors on orange. Two errors on purple. One error on clear. No errors on pink, rust, shiny, green. 2014-2019h.

C4: 93% correct. 14/1. Single error was target to stick on green. B jumped down during this session. 2019h.

B5: 76% correct. 13/4. Two errors on pink. One error each on purple, orange.

C5: 86% correct. 19/3. One error each on rust, orange, clear. One error due to researcher impatience as subject still masticating. 2030h.

Trends in errors: I have noticed a trend that I will call the Ultimate Indolence Effect. The birds often have errors with any targets that are at the end of the array. They almost always go for the penultimate target, too lazy to go all the way to the end to get the correct target. This appears to be ineffective, since it only delays when the subject will get the reward. Nonetheless, the UIE seems to be a factor with both birds. I have since tried to rearrange the targets so that they will be roughly equidistant to the subject in a surrounding arc, instead of in a straight line where the ends are much farther away. I hope this will negate the Ultimate Indolence Effect.

I should probably feed data into a spreadsheet for analysis, as typing up notes on each session and the errors isn't getting anywhere.