I understand that direct experimentation is not the only way to determine whether something has an effect or not. I said that you would want to have matched samples of students/teachers, not that they would be an absolute necessity. However, this school's setup contains a number of unusual features, all of which will be operating simultaneously.
They have vastly higher paid teachers who were rigorously selected from a national pool of applicants with an odd work and class schedule and novel administrative structure all being applied simultaneously. The existence of a private school with small classes and skilled teachers with lower pay does not necessarily provide a good basis for comparison, largely because the student body would likely be vastly different in composition as compared to this group of kids chosen because they were economically and educationally disadvantaged.
Supposing students at this school did achieve unusually highly, we still wouldn't know if it was the teachers, the class size, the general structure of the school day and class environment (it sounds like there will often be multiple teachers in the same room, as in their observation period), or the administrative structure (not that that seems likely to have a great deal of influence on student performance, but it would be interesting to see how it impacts the teachers' opinions of their working environment and the parents' opinions of the school and its responsiveness.) It could be any or all of those factors acting together or against each other, or something else entirely that we haven't considered at this point.
In order to pick apart the various effects, you'd need some level of comparison among the various conditions, no amount of statistics can pull apart the effect of class size and teacher selection/skill/pay (whatever you want to call this or these conditions) if both are covaried in all of the cases we have to look at (one, it would seem...) I guess the solution might be what you refer to as 'careful research design,' i.e. not always covarying the two conditions.
Yes.
However, it's also the case that if this school manages to consistently produce good results, it will show that some combination of factors which does not include smaller class size is, in fact, workable, which in turn will falsify the notion that smaller class size is a necessary condition for producing better results--a position often tossed about politically by people who are not knowlegeable in the field.
If that's the case, I'm not sure what your problem with my statement is.
Only that it included an unnecessary and inaccurate veiled insult, to which I responded in kind.
My apologies for missing the single phrase, "I don't doubt that class-size is relevant," in which you suggest that small class size is beneficial. I would suggest that you improve your skills in arguing. I set up no strawmen here, I simply missed a small portion of a single sentence. Additionally, the ad hominem attacks that are your response do little to further your position.
You'll need to look up the definition of
ad hominem, and use the term correctly if you expect to gain any respect for your own argumentative skills by tossing it about. And since my initial conjecture was that, "It will be interesting to see if they can put the lie to the 'smaller class size' theory," that small sentence you claim to have missed was not of central importance anyway. Whether you
intended to set up a strawman or not, it's exactly what you did.
For the record, however, I do not suggest that smaller class size
is beneficial, only that it appears to be a relevant factor. Certainly it reduces teacher workload, as drkitten points out, and it's certainly plausible, then, that it might be expected to allow teachers to perform better,
all other things being equal. However, as you've pointed out here, all other factors are rarely, if ever, equal.
I made it quite clear that my two comments were anecdotal and NEVER claimed that they were proof of anything, thank you. There is nothing wrong with anecdotal evidence as long as it is clearly viewed as such. If we had 'boring old data' to look at, we probably wouldn't be having this little chat.
Yes you did make it clear that your comments were anecdotal, and that's fine--except that you used them inappropriately in support of a sweeping generalization, which you've conceded was poorly stated. I never would have jumped on this at all, had you not got my dander up from the beginning with an unnecessary insult.
I apologize.
Moving forward, I hope I've at least made it clear that I'm not advocating this model, I'm just interested in what may or may not be learned from whatever their results turn out to be. If they have come up with some set of conditions which renders it possible to produce good results even with larger class sizes, then it will be of further interest to determine just which of those conditions are sufficient to produce that effect, and whether it might be feasible to attempt to replicate them elsewhere. As I mentioned above, I am particularly interested in the idea of having teachers do double-duty as administrators/support staff, rather than having 3 different classes of employee who do not always communicate/work well with each other. My guess is that this will turn out to be more important than either the higher salaries or the difference in average class size (both of which seem only to be pragmatic concessions made in order to allow for the change in teachers' job description, anyway).
