re: Seeking advice

[glen e. p. ropella]

 > don't understand the interplay of all of these things.  While I don't 
 > know if the components of educational systems are so fuzzy and 
 > qualitative that modeling won't be of great value as a predictive 
 > tool, I have hopes it will allow us to make decsions about some of the 
 > things we are doing where we have real numbers of at least ranges.

 > the period of our award.  We require awardees to reform their 
 > standards, curriculum, instructional methods, student assessment, 
 > professional development, convergence of resources, policy, parent and 
 > community support and report on the impact of these changes yearly.  
 > While this essentially describes our theory of what it will take to 
 > enable high achievement, in my mind, there is no really good 
 > conception of what a school district is or how these factors interact 
 > to drive what we want.  

Dan,

The basic questions you'll have to ask yourself is whether or not
these things (standards, curriculum, ...) and their effects (student
achievement, ... [anything else?]) are modellable.  Even if you 
are restricted to some primative modelling (like using abstracted
distributions and correlations), putting those models into an 
integrative environment will help point out the interactions between
the various pieces.  This type of thing helps the modeller to understand
the system better.

So, what I'm saying is that the "districts" may well be too fuzzy
and qualitative to model *now*.  But, after having tried to model
them, you will be more aware of what you should study to make the 
system less fuzzy and more quantitative.

Just to get started.... what might be a correlation between, say,
curriculum and student achievement?  Could you say that choosing
curriculum X out of a set of curricula (X, Y, Z) increases or 
decreases student achievement?  Or is this sort of question still
too broad?  For instance, it could be the case that curriculum X
works in some districts but not in others to increase student 
performance....  Can we ask these types of questions and get 
reasonable responses?

If we can't, then it may be reasonable to start at the bottom and
assume we know what motivates students, teachers, and policy decision
makers.  Then we could model the three types of agents based on what
motivates them and how they behave in certain environments.

glen