Some insightful analysis (particularly about the experimental method): http://www.neverworkintheory.org/?p=432
I suspect that the effects of the different ideologies would become more pronounced with a larger project, or (as the authors suggest) with team projects. There’s that tricky old saying: “any project small enough to be completed within a software engineering course does not actually require software engineering.”
[Comment removed by author]
This is one of the problems that I’ve had with papers on software engineering; they never seem have tests that I’d consider significant. I admit to not being incredibly familiar with the state of the art and the current literature, but I’ve seen far too many papers that try to extrapolate these kinds of grandiose statements from seemingly tiny sample sizes. Or making statements that appear to be about industry, but they did the experiments on CS undergrads.