I think Mark has very astutely brought into question some of the theoretical underpinnings of PBL (which by the way are post-hoc as Barrows had no theory in mind, this was a technological approach in the purist sense of the term). I think the ties to Dewey and Freire work (although I've seen Dewey's name invoked for quite a range of things) in the sense that this is, essentially about an attempt to embed instruction in authentic practice (or as authentic as you can get safely and with a reasonable cost)--I'm not sure that Barrows is as interested in empowerment per se in the sense that Freire discusses it but I don't think Mark is making that claim either.
Mark essentially proposes an alternative or enhancement to the lecture based approach that is deeply rooted in cognition:
To be able to retrieve something from memory, information must be encoded from short- to long-term memory. If the problem is merely one of inappropriate learning objectives we could easily divide out the tasks into a stepwise process, create the learning objectives, and proceed. The following procedures might suffice: 1) Requires students to memorize neuralgic information; 2) provide sufficient testing to ensure facts are encoded in long-term memory; and 3) encourage students to review information until retrieval is necessary, such as clinical experience. . . . According to Barrows students memorized the information the freshman year and then were not expected to use it until their junior year. But isn’t the problem with this approach the lack of additional timely experience requiring students to demonstrate their ability to understand, apply, analyze, evaluate, and create? So what would PBL add to this scenario that more solid instructional design would not?
Mark makes a great point here. It is perhaps a bit of a straw man to compare PBL to straight lecture, developed and delivered by faculty of all people--the majority of whom receive absolutely no training about how to teach much less engage in instructional design. I say this as an equal offender by the way, if I had a dollar for every behavioral objective I haven't written for my own classes or audience analysis I've skipped I think I could make my next house payment. Ok, maybe not a house payment but I could get an iPod. But I digress, so in a head-to-head between a principled instructional design approach lecture format class and PBL what is the difference? I would say ostensibly not a whole lot. An effort to supply opportunities to apply, analyze, evaluate, and create in relation to new knowledge might lead you to something fairly proxmiate to PBL (certainly in comparison to traditional lecture--whatever that means). I think the biggest points of divergence are 1) forcing students to take a leadership role in identifying their own learning issues (through meta-cognitive prompts) and then resolve them (by seeking out and evaluating information) with respect to a given problem, 2) a happy consequence of forcing leadership; increased propensity to engage in lifelong learning (if you go through medical school having everything handed to you, you wind up with the typical graduate school dilemma--now that I'm responsible for figuring out what it is that I want to learn, and how to go about doing it I'm not quite sure how to get started, 3) a practitioner focus, which deserves a more elaborate explanation.
Mark infers the purpose of PBL as "[encouraging] learners to become independent problem solvers." Which I think is close but not quite on. I think if you asked medical school faculty who lecture about their teaching responsibilities they'd say "I teach medical students" or perhaps "I teach (insert discipline here) to medical students." Put differently they'd say "I'm in the business of teaching students about medicine." PBL is not attempting to teach students about medicine, it's in the business of making doctors. This is embodied perhaps best in a quote from a PBL student found in Schwartz et al. (1992):
“We would dive into the case, and soon it would be students teaching other students about the information we had found. I also found these group sessions encouraged questions to be asked about cost of treatments, benefits of a treatment or patient options.” (emphasis added)
In short, students in a PBL approach start to focus on the practice of being a doctor, some of the less tangible questions (costs, dealing with HMOs, etc . . . ) that are often the focus only after medical school is over. Now you can certainly apply Mark's argument here that this is something ISD would pick up as well--it's a content issue, you aren't giving the entire picture because you having gone through a principled input/output analysis, gap analysis, etc . . . I would say that's quite true, but that Seels and Richey would still put you on a path to learning about what it's like to be a doctor, and not necessarily about being a doctor.
Mark invoked cognition, which I think is important to address. I'll go back to some of the basic (and quite famous) research done by Hyde & Jenkins (1969) on enhancing rehearsal/recall. They had three comparison groups, one "intentional" group that memorized a list of words and attempted to recall as many as possible. If I remember correctly the time spacing for all three groups was such that it involved long-term memory. I'll equate that with a great deal of overgeneralization and partial hubris to "traditional lecture", the other two groups were not told they would be asked to recall the words at all (thus no conscious attempt was made on their part to encode them into long-term memory). One group was asked to identify the words that contained the letter "e", the other was asked to rate the "pleasantness" of the words. The unintentional "pleasantness" group did as well as the group that used rote rehearsal. I would (again with much in the way of overgeneralization) equate the pleasantness group to PBL in that students are making evaluative judgements about information and it's application and relevance to a given problem (as they seek out and encounter that information). In essence, PBL affords additional retrieval cues, when (as Mark describes it well) they are contextualized in patient cases. Looking back, I think Mark discussed this point a bit better than me (certainly more concisely).
I do question the claim about transfer, however. PBL affords highly contextualized knowledge. I'm not convinced that highly contextualized knowledge will transfer more easily to novel situations. Certainly, many PBL advocates claim this--but if you read between the lines they are typically talking about novel patient cases. When I think about transfer I'm thinking about a discipline switch--which I think you will get with your weak (broad) problem solving strategies, but I'm not convinced you'll get good transfer--or more to the point, better transfer of your knowledge over a more traditional lecture approach. To take this out of the abstract I've encountere Maslow's heirarchy of needs in the context of a parent-child relationships class in the field of Human Development, a management course in the context of Business, and motivation (learning theory) in the context of Instructional Design. All of them through traditional lecture--I'm not sure that learning Maslow's Heirarchy through PBL would help me transfer what I know to these different disciplines.
Anyway, that's my long-winded response. Thanks for the post Mark.