[MD] Imaginings

[Arlo Bensinger]

[Arlo previously]
By this same logic, schools in America in regions of "less diverse 
culture" should do better than both schools in America and schools in 
F/J in areas of "largely diverse populations".

[Craig]
Bad sampling.  You need to compare schools in cosmopolitan Tokyo to a 
cosmopolitan American city & in rural Oklahoma to rural Japan.

[Arlo]
Well the variable in question was only that schools in areas with 
"less diversity" perform better than schools in areas with "more 
diversity". You are adding a new variable (urban-rural) to the 
equation, saying that population-size also has an effect on 
educational performance (hence the need to hold it constant).

But if we compare (as had been the original case in point with 
Finland and Japan) areas there are here we DO see that in both 
urban-urban and rural-rural comparisons, they outperform us most of the time.

The question, I suppose, then becomes. Is there truly less diversity 
in rural Japan than in rural Oklahoma, so much so that this alone 
accounts for the difference in education results? And this takes us 
back to, how are we defining "diversity"? Because if we compare 
Oklahoma City to, say Funabashi, I'd want to know what variables are 
being used to determine one is more diverse than the other.

[Craig]
Again, bad statistics.  In order to be legitimate, you need  to 
compare a diverse school district with bottom -up control with one 
with top-down control & a non-diverse school district with bottom-up 
control with one with top-down control.

[Arlo]
But, again, Craig, the only variable in question was that schools in 
areas with "less diversity" perform better (when education is handled 
via "central planning") than schools in areas with "more diversity". 
"Top-down" and "bottom-up" control simply adds another variable to 
the equation. Indeed, you seem to say here that comparing schools 
with different levels of "diversity" is in err, since both of your 
comparisons pit only "top-down" against "bottom-up". In my example 
(SC/BE), both schools have "top-down" planning, both are "public 
state schools", so those variables are held constant. The difference 
is really only one of "diversity". Well, the SC district is larger, 
but educational results take this into account. Its not that SC has 
MORE high performing students, but a greater percentage of the total 
student population performs better. So, again, I think its valid to 
ask, why does SC outperform BE if, as is being suggested, schools in 
areas with "less diversity" perform better than schools in areas with 
"more diversity".