Re: Tampering and grading

["Mowery, R. Neal (RNM) " <RNM@ornl.gov>]

Jim the following post may have been lost, and is being resubmitted as you
suggested.  Thx!

Neal

*************

Del Nelson wrote:

> Anyone who disagrees with that MUST be able to resolve/solve 
>the question/problem:
>
>A  +  BC  +  C  =  D
>
>A is the contribution of the individual.
>C is the contribution of the system.
>BC is the contribution of the interaction of the individual and the system.
>D is the observed output/result/measurement/product/ service/count/etc., the

>"evidence" the teacher/boss/ manager/rater sees/receives/has access to.
>
>Now, IF you can tell me what "A" is (and assign it a rating/grade/etc., you 
>MUST be able to tell me what C and BC are. That's M-U-S-T, SHOW ME THE
DATA!!!
>
>Or? Or your are tampering and doing harm both to the individual and to the 
>system. PERIOD!

The word "must" in Del's message started me looking for counter-examples.
(Perhaps it is the rebel in me.)  The link to Deming is tampering and
feedback in PDSA.

For purposes of prediction, I do not need to understand the independent
influence of each element in the equation above.  For prediction, it is
usually sufficient to know what I can expect given this combination of
individual, teaching method, testing method, etc.  If I keep the same
person, same teaching method, same material, I can expect the same result.
If different methods provide different results, I should take advantage of
that, perhaps on a student-by-student basis.  Without some sort of grading,
that difference can never be determined.  Call it what you will but some
sort of testing is needed as feedback to the system.  Implicit in testing is
some sort of scoring, and ultimately an evaluation of the score relative to
expectation.  Perhaps the teacher is in the best position to do this.
Perhaps a standardized feedback system , whether alpha (A-F), numeric
(0-100%), binary (pass/fail), or narrative "meets expectations") is useful
to communicate performance against expectations.  If so, then we will have
grading in some form or fashion.

I suggest that the problem with grading (and with the use of performance
measures in general) is that people *interpret* the results inappropriately.
A letter on a grade card is not tampering.  Tampering occurs when people
fail to understand what that letter is telling them, and when the grade is
not viewed in a larger context.  If grades are viewed (appropriately) to be
limited in the amount of information that they carry, we don't tamper.

Who is wrong:  The schools who give grades?  Or the
parents/students/businesses who read too much into the grade?

More importantly, what do we control?  The schools who give the grade?  Or
our own ability to filter and interpret the grade?

Thanks!

Neal Mowery                          
Statistical Applications - Lockheed Martin Energy Systems
Oak Ridge, Tennessee
RNM@ornl.gov        423/574-0796
=====================================================================