I have been thinking about this for a couple days - you will find my response below!
I can find no resource indicating any data collected in sea pens. If it's being done at all I would imagine it's conducted during a rehab and release project under one of the anti captivity organizations.
Two weeks ago I asked the question regarding application of collected data to two "behaviorist/trainers" in The Living Seas at EPCOT. Answers I received were "They can identify a shape of a shoe". "They respond to our directions". "If they don't want to work with us we don't make them." "You can stay for the next show at 2:00 and see what we do." My jaw dropped but I smiled and remained non-confrontational. Even my husband (who humors me with my cause) walked away and said "they don't have a clue."
During our vacations in 2011 and 2012, my husband and I watched dolphin research at The Seas at Epcot and were impressed with the type of research being presented. The particular researchers we saw seemed to know their stuff. We did attend two presentations (one in 2011 and one in 2012).
The concept that I saw the dolphins working on in 2011 was same/different. They would show the dolphin a pair of objects - either two of the same thing or a "pair" of different things. The dolphin would then go over to another researcher who would show them two or three? different sets of objects (could contain objects from the first pairing or completely different objects) and the dolphin would pick out the pair in the group that either contained a pair of the same objects or a pair of different objects depending on if she saw a set or same or different from the first researcher. The dolphin was not seeing a pair of objects and finding its match in another spot but instead learning the concept of same and different and then applying that knowledge. For example, let's say the dolphin was first shown two apples. The correct answer out of the set she was given in the answers section was a pair of pineapples (2 of the same object). Or let's say the dolphin was first shown a pineapple and a banana. The correct answer would be an apple and a pear (not pineapple/pineapple or even shoe/shoe). I hope I did an ok job explaining it, it would be a lot easier with visual aids.
There were three researchers there, one to show the first set of objects, the second to show the "answers" and the third to videotape the whole thing. It seemed well constructed so that one of the trainers didn't have the ability to "tip off" the dolphin as the right answer. There was also a written table of what order and how the objects should be shown so there couldn't be any researcher introduced bias either.
The second research we saw a year later was on object rotation. This work was with a younger dolphin who was just learning and he did not always pick the right answers like the dolphin in 2011 (who got almost 100% right). Basically, the researchers would show the dolphin an object and then take it away. The second researcher would then show the dolphin a set (I think it was three, but could have been two) of objects, one of which was the first object rotated. The dolphin was then supposed to pick the object he first saw which was now in a rotated position.
The ones I'm thinking of specifically don't require pools. I should actually say that the place I did my internship for that does a lot of research actually does use sea pens, but a sea pen is still captivity, so I don't think it's a pool vs. sea pen issue necessarily. The size of a sea pen and a pool are comparable. You CAN make sea pens larger for orcas, but not every sea pen is giant. The pools we used are comparable to the size of dolphin pools at other facilities, just in the ocean.
I do think there is important work being done on dolphin intelligence in captivity. Am I positive that, on balance, that it is worth it? No, I am not positive, but I lean more towards yes it is a good thing than no it should be stopped.
To start with two examples from other species so that I can best illustrate my point - first, Irene Pepperberg's work with Alex (the African Grey parrot). Over time, Pepperberg was able to teach (or communicate with) Alex in order to prove that he knew concepts like counting, shapes and colors. In the end, she also did work with Alex that appeared to show that he understood the very abstract concept of zero. But in order to get to the point where she could find this out about Alex, a lot of teaching and learning to communicate had to be done. The bridge between species to communicate had to be built over time and develop with a relationship. I am doubtful that such a relationship would ever develop in the wild. Thanks to Pepperberg's work (and others too), we now have a completely new understanding of parrot intelligence. Our best estimates now are that parrots are at least as smart as a 4-5 year-old human child.
Another example, chimps (I believe) were taught to use "money" (actually tokens) to exchange for food. Tests were run by behavioral economists to see how changing the price of food items would matter to the chimps. If I remember correctly, the researchers made a mediocre item go up in value (to mimic the human model/belief that very expensive things must be worth their high price - why else would it be so expensive unless it is worth it?) which had no effect on making the chimps demand/want more just because the price went up and was now a luxury item. Instead, the chimps continued to want the things they individually valued most and could not be influenced by someone else making a value judgement for them, proving perhaps, that with regard to economic choices chimps may be more rational than humans.
This research has a real world effect because I think understanding (eventually) leads to more respect and care for a species - both those in the wild and those in captivity. I am not saying that only "intelligent" species should be protected, just that when we have a better understanding of a species it is harder to see an animal as a "them" versus an "us." Humans have been on a constant quest to find the obvious dividing line as to what makes "us" human and what makes animals "them." At one time, it was widely thought, that only humans use tools. Now, of course, we know that chimps, elephants, and ravens can use tools too - so that one gets knocked off the list. What about the ability to understand the concept of self - with the famous mirror test animals such as dolphins and elephants have shown that they have the concept of self as well. What about the ability remember and mourn the dead - elephants in the wild have been shown to stop at the bones of dead members of their herd when they journey past them (what exactly they are doing, we don't know, but the behavior is notable). What about understanding abstract concepts - dogs who know the names of their toys seem to be able to figure out to retrieve an unnamed toy if it is left in a pile of toys that the dog already has labels for - so figuring out the requested toy must be the one they don't know as it is the only one in the pile they don't have a mental label for, Alex the African Gray understands zero, and the dolphins at Epcot are proving that the concept of same and different (not just matching up pairs) is something they can understand and process.
So to bring the discussion around to dolphins in captivity in general. In addition to the work at Epcot which I believes does require a relationship across species to make communication possible, I also saw an experiment with captive dolphins on an episode of Nova Science Now on PBS. Basically, the dolphins had been working with the trainers for a long time and the dolphins and trainers had a vocabulary/communication system. Through this communication system, the trainers were able to tell a pair of dolphins to make up a new trick on together, just the two dolphins. The dolphins then exchanged clicks underwater and were able to do something for the trainers that the trainers hadn't taught them and the dolphins hadn't done before together. This adds to our body of knowledge of dolphin intelligence and in the long run can be of benefit to both wild and captive dolphins.
I am very excited about all of the intelligent life that we have here on earth that hasn't fully been researched and is far, far from being understood. Therefore, I think dolphins in captivity for research and education purposes probably passes my gut check test for now. Perhaps 10, 15, 20 years from now when we have a lot more research and perhaps when humans become less concerned with the clear line dividing "us" and "them" and more concerned with the planet's ecosystem and inhabitants as a whole, it will no longer be an acceptable practice.
That being said, I don't think just any organization should have dolphins. The two main goals should be research and education. Last year I was at a zoo in the mid-west which has a dolphin show. I was dismayed that the entire show was just the trainer dancing around to blaring music while the dolphin "danced" around in the water. No new information was gained for research purposes and no one in the audience was taught anything either. This is also an area where I think SeaWorld fails. As mentioned in this thread before, the main requirements to be a trainer at SW are high school graduate, physically fit, and looking good in front of an audience, so I am not sure how much actual research is going on there (my guess is little to none) and based on the shows I have seen at SW, the emphasis is heavily on entertainment and light on education (to me).