[quoting Oliver, 24 Oct 96] 1. The paper aims "replicate previous results without apparent confounding factors." So, his aim was not to discover whether there is any effect on learning due to differing media, but to prove that there was an effect...
My purpose, technically, was to disconfirm the negative hypothesis: media do not influence learning. I did that.
In a later reply, he confirms that this was the case. "I had reason to believe Clark's hypothesis was wrong and I wanted to test that belief." So, we are in the belief testing business here, not in any way objective.
Give me a break! I'm not allowed to believe things? The question of objectivity should be addressed to my methodology. Did I prejudice my results in some way?
But in the same reply he claims that he wants "to ascertain TRUTH!!!" These aims seem to be at odds, especially as other responses make claims about this being "science."
The ALL-CAPS and triple exclamation marks should have been a tip-off that I was being ironic. Lighten up.
In another response, he says "What's vague about text/audio? My guess is that everyone reading this knows what those words mean."
I don't think guessing should form part of any work that makes claims about being "truth" and "scientific."
All science involves guessing. That's what hypotheses are.
The final words in the paper state baldly "Media influence learning." Surely a definition of "media" must be included before that can be taken seriously.
I defined media in my original essay.
I discussed in previous postings my concern about the design of the experiment, which a number of people have supported, both in this forum and in private messages. These concerns remain, despite assertive if not downright aggressive cries of "Irrelevant" from Steve.
The criticisms were irrelevant to the hypothesis being tested.
I would like to be constructive (not constructivist) and offer a compromise...In an experiment with 36 subjects of unmeasured reading and writing ability, visual acuity and hearing ability, the scores on a written recall test as interpreted by the researcher, following exposure to either textual information presented on a VDU under unspecified conditions, or spoken information presented under unspecified conditions, appeared to be different.
Sorry...
In an experiment in which subjects were randomly assigned to treatments, thus negating influences of reading, writing, visual acuity, etc., subjects exposed to text on a VDU screen remembered more information than subjects who heard the same content through headphones. The chances that this difference was a fluke are less than two in 10,000. This experiment replicated similar results obtained in 14 separate comparisons, with different content, delivered through different means, and with subjects drawn from different populations.
[quoting Reeves, 24 Oct 96.a] But the actual CBT they were showing boiled down to essentially what we have been doing since the days of Skinner's teaching machine, i.e., show a little content, ask a question about it, judge the learner's response, and provide feedback.
I agree completely. All instructional theory that I know of can be represented at the lowest level of the Chomsky hierarchy. The problem is that normal human discourse (and student-tutor discourse) is almost certainly at (at least) the next higher level (context-free grammar) and maybe higher. If CBT is going to progress it needs a theory at a higher level. It is the job of Professors of IT to devise that theory.
Your criticism of the current state of CBT is on target. Dave Merrill's group is addressing part of this problem but not all of it.