A Tale of Two Curricula: The performance of two thousand students in introductory ele

[ZapperZ]

I'm still reading the paper, but this appears to be a rather extensive research (http://arxiv.org/abs/0906.0022) on the students' response to two different curricula on teaching intro E&M at the undergraduate level. They did this at 4 different well-known universities in the US and arrive at a rather conclusive and clear result.

Zz.

[Moonbear]

This is a LOT better controlled than many educational studies I've seen. I like the multi-institution approach, as it seems a sounder method than limiting the study to just one student population of one university, as too many educational research studies do.

I didn't read all of it, because the specific topic they were teaching isn't of much interest to me, and doesn't make a lot of sense to me for that reason, but I was more interested in the methodology.

I have a general question about this type of study...what would you consider the "N" for this study? Is it really 2000, the number of students tested, or would it be 8 (4 for each teaching method), which is the number of classes taught? Though, unless my quick skim missed it, they didn't actually specify how many classes or lecturers were involved at each institution. There may have been more than one of each at some.

I ask, because one concern I always have in comparing teaching methods is that simply having an engaged instructor who is willing to try new methods might mean they are giving a better lecture regardless of the method used. Likewise, from year to year, I really do see noticeable differences in attitudes in whole classes that affect performance of the class as a whole. So, for statistical purposes, I think one must consider the number of classes within the design. I kind of look at this as a nested factorial design...students, students within classes, classes within institutions. In this case, there are still sufficient degrees of freedom to do a proper statistical analysis. This is my concern with too many other studies that they only look at one institution or one class, and the results aren't really sufficiently generalizable outside the institution to be a valid research study.

[ZapperZ]

I think that's why I was a bit taken aback when I read this study because of the sampling size. They certainly address that this can't simply be due to just the instructor alone, because separate analysis of each institution also reveal the same difference between the two curricula. So even the different location and institutions reveal the same "trend".

I posted this also on my blog, and I had a comment from one person who not only took a course using the M&I text, but also taught it as a TA. He/she basically concurred that the M&I approach is completely more ".. effective, interesting, and fun". So that seems to verify the conclusion of this study.

I haven't looked very closely yet at the methodology that they used since I skimmed that part of the paper. But I find it intriguing that this technique can yield that clear of a difference.

Zz.

[Andy Resnick]

It's definitely intriguing. My major question is what the BEMA is assessing- is this a case of 'prepping for the test'? It doesn't help that the paper's authors (not) coincidentally wrote the M&I textbook and the BEMA.

[Doc Al]

Very interesting. I haven't read the paper in detail, nor have I seen the M&I textbook (but I've heard of it). But I've read several excellent pedagogical papers by R. Chabay and B. Sherwood, the authors of that textbook (as Andy points out), over the years. (A couple of Sherwood's earlier papers straightened me out on a few topics that I didn't even know I was messing up. :uhh:)

[ZapperZ]

It's definitely intriguing. My major question is what the BEMA is assessing- is this a case of 'prepping for the test'? It doesn't help that the paper's authors (not) coincidentally wrote the M&I textbook and the BEMA.

Yeah, that's the only part that I wish can be verified independently. But I think that the courses being taught, even the M&I syllabus, were not all done by them.

I would hope that another school not connected to any of the authors would try and duplicate this, maybe even use another test.

Zz.

[physics girl phd]

One of our faculty members used Matter and Interactions for this past year and really liked it. The class population is small though... at our university the calculus-based physics sequence for scientists is separate from the calculus-based physics sequence for engineering students. This faculty member also didn't take any form of pre- / post- assessment that I know of... so all I have to report is the "liking" the faculty member had for the curriculum.

From what I understand, Matter and Interactions works from a momentum/impulse point of view rather than a position/force point of view... the development supposedly makes the transition to Quantum more smooth.

There's also a lot of coding also, I am 99.99% certain that the authors are also associated with developing VPython. I think the coding, etc. is what really makes a difference. Our faculty member taught the course in a computer lab facility and tried to integrate both coding activities and lab activities with the course better, like a studio physics course.