PS>Yet, the project revealed no statisical significance in learning growth
PS>between girls and boys.
 
RM>How is the "learning growth" being assessed, and may we
RM>assume that this is not the same as LEVEL of learning?
 
PS>Okay, what was measured was a two year growth of math computation
PS>abilities - as measured on the Iowa Basic Skills Test.  Each student was
PS>his own control, and the tool of measurement was the improved
PS>Standardize Score.
 
Ok... How did you allow for random fluctuations in score?
We are all very conscious of the fact that there are 
considerable variations in an individual's IOWA score.
Assuming for a second that we're going to assume that the
sample will "average out" the fluctuations, how large are
the various sample groups?
PS>This data is really weird,
RM>So far I haven't seen anything weird?
 
PS>girls should have done better than boys... 
Doesn't this depend upon age, and is this expectation on
LEVEL of learning and not GROWTH?  Also, this may be
impacted upon by other factors; financial, even religious.
PS>minorities should not have done as well...
Isn't there a hidden financial factor here as well?  Most
minorities live in urban areas and are poor.  Do YOUR
minorities really fit into that group, or are they, in fact,
sufficiently different to skew the results?
 
PS>remedial students should have not done as well....
These have exhibited problems which place them in a specific
group... Of all the groups listed, this is the one most
likely to be accurately defined...
PS>lower socioeconomic class should not have done as well...
Again, lower socioeconomic is also associated these days 
with urban areas, and this, in turn, is often associated 
with other factors such as one-parent homes, violence,
drugs, multi-generational histories of poor education, etc.
I think that one would see differences in a poor rural
community which had access to decent schools, for example,
or where there might be immigrants who embrace literacy
but are poor (asians in the early 1900's perhaps, jews in
the mid 1900's, etc)...
PS>With those groups there was no significance in growth.
Then I would look to these, and other, variables...
PS>What I find remarkable is to be wrong four out of five cases,
PS>with the only case to be correct is behavior.  This is highly
PS>inconsistant with the data.
I'm surprised only by the remedial students, but even here
there is an extenuating factor; the nature of the remedial
program itself.  Many remedial programs are inadequate and
fail the students; perhaps yours does not?  This plays into
the discussion I've been having with Chuck (that I'm trying
to drag you into )...
PS>The literature is based on other studies - usually more that one
PS>study.  So, basically, many studies would indicate the above to be true.
But there ARE "hidden" variables that can make a significant
difference in the outcomes!  One of the biggest may be the
apparently random fluctuations in IOWA scores!
___
 * MR/2 2.26 * He who laughs last uses OS/2 WARP.
--- Silver Xpress Mail System 5.4P1a
---------------